#: 16311 S1/NonTech Cust. Serv. 01-Nov-92 16:04:17 Sb: WIN32/NT SDK Release 2 Fm: Don Perry 76676,1127 To: Stu Wiley 70473,1351 (X) Thanks Stu for the info. Your prompt response is appreciated!!! Don Perry #: 16340 S1/NonTech Cust. Serv. 02-Nov-92 10:35:17 Sb: Setup, Dev Kit Fm: Bob Soodak 71055,542 To: All Re: Problem installing NT from MS-Win32 SDK CD-ROM Motherboard: IBM PS/2 Model 40SX SCSI Adapters: TCM-850IBM Manufactured for IBM by Future Domain; IBM 3.5 inch rewritable optical drive attached, installed internally. Adaptec AHA-1510 (from Corel systems); Corel (manufactured by Toshiba) CD-ROM attached, installed externally. Problem: Boot Disk (July 1992) Setup program fails to recognize the presence of a valid CD-ROM drive. Note: In all other respects, the CD-ROM drive functions normally. I had no problem installing the MS-Developer Network software from the same drive. #: 16350 S1/NonTech Cust. Serv. 02-Nov-92 11:23:29 Sb: WINDOWS NT BETA Fm: Stu Wiley 70473,1351 To: Sung Y. Kim 70353,2743 (X) Send a letter in to Microsoft, attention Windows NT Beta coordinator and request inclusion to the Beta program. Alternatively you may fax to 206/936-7329. Microsoft One Microsoft Way, redmond, WA 98052 Stu Wiley Developer Service team #: 16362 S1/NonTech Cust. Serv. 02-Nov-92 12:43:48 Sb: MSCDEX Fm: Ralph H. Brenner 71730,2203 To: anyone To use my CD-ROM drive I need the Microsoft Compact Disc Extensions (MSCDEX) version 2.2 or later. How or where do I get this? Ralph Brenner #: 16391 S1/NonTech Cust. Serv. 02-Nov-92 15:57:19 Sb: #16264-MSDN CD Crash Fm: Jerry Drain [Microsoft] 72350,2056 To: Richard George 70640,2614 Hi Richard, You might ask this question in the MSDNLIB forum. We don't handle many of their detailed issues in this forum. I haven't seen this problem before (but again you wouldn't expect to see many like this in this section). This sounds like a problem with their installation tool. Another possibility, do you have empty space on the drive where your swapfile is located? See if the problem is resolved by making a little more room for your swapfile to grow. Thanks, -- Jerry [@Microsoft] #: 16278 S1/NonTech Cust. Serv. 31-Oct-92 10:42:25 Sb: SDK install error msg Fm: Bruce D Franson 71053,3415 To: all Thanks for the new cd (October Beta). Looks great! Had an informational message when I installed the sdk after installing the new windows NT. Right when it was creating the new windows under the NT program manager it gave me a non-fatal dde error - twice. I selected the "continue" option and it then the next thing it did was create the sdk group under the program manager, then install the items one after another. Looked ok at the end, but thought I would let you know, since that's what beta programs are for, right :-). I did the normal install from CD, didn't select custom or anything. Bruce There is 1 Reply. #: 16392 S1/NonTech Cust. Serv. 02-Nov-92 15:57:24 Sb: #16278-SDK install error msg Fm: Jerry Drain [Microsoft] 72350,2056 To: Bruce D Franson 71053,3415 (X) Hi Bruce, Thanks for the information. I've passed this information internally to the guys handling these kinds of issues. For future reference, if you experience setup problems with the SDK, please let the folks in Section 7 know about them. Thanks much, -- Jerry [@Microsoft] #: 16393 S1/NonTech Cust. Serv. 02-Nov-92 15:57:29 Sb: #16214-Remote Login? Fm: Jerry Drain [Microsoft] 72350,2056 To: Andrew Potter 71075,614 (X) Hi Andrew, When I saw your initial posting, I happened to see Bruce and asked him to comment on your question. Bruce doesn't typically visit this section. Unfortunately LAN questions are not my forte. If you'd like to hook up with Bruce again and ask him to readdress this question, please repost in Section 12, "WinNet and RPC". There is also a Network Services section available in the WinNT forum, Section 11. Thanks, -- Jerry [@Microsoft] #: 16289 S1/NonTech Cust. Serv. 01-Nov-92 00:39:33 Sb: robustness vs speed Fm: Jeong Ho Lee 70253,1244 To: sysop (X) Hello, Has Oct. Beta achieved *speed* by sacrificing *robustness* ? With Oct. release, everything is reasonably fast. I can use almost ALL Win16 apps and DOS Apps. TapCIS is working well.....Great! But reliability of Oct. release is not as solid as Jul release. With Jul release, my system never be locked or hanged. But Oct release *randomly* is broken. There has been no changes on my system. Even , I can not report BUG, since the behaviour is not re-produceable. We need robust OS, not fast OS. Because hardwares are always improving. Please do not sacrifice *robustness* to improve *speed*. jLee There is 1 Reply. #: 16394 S1/NonTech Cust. Serv. 02-Nov-92 15:57:35 Sb: #16289-robustness vs speed Fm: Jerry Drain [Microsoft] 72350,2056 To: Jeong Ho Lee 70253,1244 (X) We think much along the same lines as you do; it doesn't make since to sacrifice robustness for speed. A slow piece of code is infinitely better than one that doesn't work at all. This is the first time I've seen a report like you've submitted. If you can at all narrow down the problems or find a pattern, please let us know. Even if it is something only specific to your machine, we'd like to figure out what is going on and fix it. Thanks, -- Jerry [@Microsoft] #: 16310 S1/NonTech Cust. Serv. 01-Nov-92 15:18:06 Sb: NTSD display drvr debug Fm: ray shapiro 71431,1450 To: sysop (X) Hi, When I was out in Anaheim, I asked about using NTSD to debug display drivers. The answer was "yes", but the magic incantaions went by a bit fast. Could someone please post the rules for invoking NTSD to debug display drivers. Thanks, Ray There is 1 Reply. #: 16400 S1/NonTech Cust. Serv. 02-Nov-92 16:08:12 Sb: #16310-NTSD display drvr debug Fm: Stu Wiley 70473,1351 To: ray shapiro 71431,1450 (X) Ray, This message will probably get more specific attention if in the WINNT forum section 10 (Drivers). Stu Wiley Developer Service Team #: 16337 S1/NonTech Cust. Serv. 02-Nov-92 10:08:43 Sb: Where is system guide? Fm: Tim Johnson 71071,1215 To: Sysop (X) There's some discussion over on WINNT about how NT/SDK developers got the October release without the System Guide, even though it's mentioned in the two books we DID get (Release notes and WIN32s notes). Evidently users' updates included this guide. How can us humble developers get this thing? Hate having users know more about installing it than me! Tim Johnson Teknekron Software Systems 530 Lytton Ave. Suite 301 Palo Alto, CA 94301 There is 1 Reply. #: 16388 S1/NonTech Cust. Serv. 02-Nov-92 15:27:03 Sb: #16337-Where is system guide? Fm: Dwight Matheny/Microsoft 70750,2340 To: Tim Johnson 71071,1215 (X) the system guide is on the cd in \doc\enduser\ps or \doc\enduser\write. -Dwight (MS) There is 1 Reply. #: 16401 S1/NonTech Cust. Serv. 02-Nov-92 16:19:10 Sb: #16388-Where is system guide? Fm: Tim Johnson 71071,1215 To: Dwight Matheny/Microsoft 70750,2340 (X) Dwight, Thanks for the System Guide pointer. I'll post this info over on WINNT too. Tim #: 16412 S1/NonTech Cust. Serv. 02-Nov-92 16:47:13 Sb: 2nd SDK Ship Date Fm: Jerry Drain [Microsoft] 72350,2056 To: William H. Beebe, Jr. 70720,375 (X) Hi William, PMJI, but Dwight must be out of town. We don't typically have information like this regarding support of specific hardware in future releases. I understand for this, another driver will have to be written to support the CD-ROM under this card; however, Creative Labs will have to tell you specifically what work is being done towards this. Their is a form that you can fill out and submit to help us identify the need for specific drivers on the market, so that we can take steps to insure they're produced. I've cut and pasted the steps on how to fill out this form (hwfeed.txt) and submit it to us. Thanks, -- Jerry [@Microsoft] =============================================================================== Hwfeed.txtcan be found in MSWIN32, Lib 17; or WINNT, Lib 1. In order to send mail to the address mentioned via CompuServe, you have to: * Leave the forum. * At any "!" prompt, enter GO MAIL. * Mail your request to: ">internet:winnthw@microsoft.com" #: 16273 S1/NonTech Cust. Serv. 31-Oct-92 09:17:25 Sb: Device Driver Fm: GMS 100063,2012 To: Leon G Rollison 70421,1702 No! #: 16413 S1/NonTech Cust. Serv. 02-Nov-92 16:47:20 Sb: Device Driver Fm: Jerry Drain [Microsoft] 72350,2056 To: Leon G Rollison 70421,1702 Hi Leon, Just for future reference, if Dave to see something specifically, he doesn't typically visit this section. He is more often in section 9 or section 12. To answer your questions, no, that card is not yet listed on the hardware compatibility list. Thanks, -- Jerry [@Microsoft] #: 16449 S1/NonTech Cust. Serv. 03-Nov-92 06:14:28 Sb: More DRIVERS congestion Fm: Arthur Knowles 71041,2613 To: Robert Eidson 70214,2115 (X) Bob, FWIW: I use qemm and it's stealth technology to create about 200+ Kb of UMB space. Even with WFW and Netware driver and my CD-ROM, etc I still have about 570+ kb free. I've stuffed the UMB space with these drivers (and a few low) so if EMM386 can't free enough memory you might think about qemm. Art PS: I also use Quick/C for Windows to do my normal compiles and only use MS C for the final optimization phase. It's worked very well since you don't have the DOS memory limits. #: 16402 S1/NonTech Cust. Serv. 02-Nov-92 16:30:50 Sb: Beta release Fm: Frido Garritsen 100041,1671 To: SYSOP (X) I have purchased the July version of the SDK. Is it possible to get the newer October version? I have been a Beta tester for Windows 3.1. Frido Garritsen Himalaya Software There is 1 Reply. #: 16451 S1/NonTech Cust. Serv. 03-Nov-92 06:24:46 Sb: #16402-Beta release Fm: Arthur Knowles 71041,2613 To: Frido Garritsen 100041,1671 (X) Frido, <> The oct beta should be sent to you automatically along with all future updates. Art #: 16453 S1/NonTech Cust. Serv. 03-Nov-92 07:42:14 Sb: #16228-Win NT tradmark instruct Fm: Andrew Potter 71075,614 To: Dwight Matheny/Microsoft 70750,2340 (X) Dwight, Found it! Thanks, Andrew #: 16318 S1/NonTech Cust. Serv. 01-Nov-92 19:05:25 Sb: October/ dual boot Fm: frederick c kirk 70570,2273 To: all After installing the october beta version, (and selecting the NTFS file option) I discovered that I could not boot DOS (and could not run my telecommunications software to access this forum!). I re installed NT with the FAT option. then I could boot DOS. Am I doing something wrong? Do I give up the DOS option when I go to NTFS? When I tried to select "previous operating system on C" I got: BOOT: couldn't find NTLDR please insert another disk" and then things kind of locked up. Also, since July, I have had to boot DOS from a diskette, the "previous operating system on c" option hasn't worked. How do I re establish that? There are 2 Replies. #: 16457 S1/NonTech Cust. Serv. 03-Nov-92 09:26:09 Sb: #16318-October/ dual boot Fm: Stu Wiley 70473,1351 To: frederick c kirk 70570,2273 Frederick, I suggest you place this in the WINNT forum section 3. I have heard similar comments, relative to installation procedures. I think someone there will be able to help you. Stu Wiley Developer Service Team #: 16466 S1/NonTech Cust. Serv. 03-Nov-92 09:50:41 Sb: #16318-October/ dual boot Fm: Stu Wiley 70473,1351 To: frederick c kirk 70570,2273 Frederick, I addition to the previous, you may want to take a look in section 7 of WINNT. Stu Wiley Developer Service Team #: 16279 S1/NonTech Cust. Serv. 31-Oct-92 10:44:16 Sb: INSTALL NON_SCSI PRELIM Fm: Schroeder 74746,11 To: All Please Help!...Here's my problem... Equipment: Generic 386DX-33 system Proprietary SCSI (Sysquest) - not use for NT install Sound Blaster Pro based CD-ROM 3.5(A:) & 5.25(B:) Floppies 500MB Hard Storage (C:80MB D:420MB) History: Installed Pre-Release version using DOS method. Experienced one minor problem with keyboard... ...resolved by changing statis of keyboard in CMOS during one portion of the install. No other significant problems encountered in installation or operation of Pre-Release version Problem encounter with Preliminary (October) Release: First, removed prior copy of Windows NT from system. Using METHOD-2 label for networks or unsupported CD-ROM players. WINNT prepares a floppy diskette, places a directory on the designated drive (D:) and attempts to reboot. Unfortunately, floppy will not boot to NT. System hangs with the message: BOOT: I/O error reading disk Please insert another disk Of course, inserting disks, or pressing keys has no effect, the system is quite well hung. Also, this is the same message that is received from the Microsoft suppied boot disk (for supported SCSI drives). Does the floppy or the BIOS lack some degree of compatibility with the new release? Is the proprietary SCSI controller confusing the boot system? Did NT incorrectly create this "Setup" floppy? Am I left out in the cold with no where to go? Answers to these questions and more in messages to follow (I hope). Thank you in advance for any and all assistance. Schroeder There are 2 Replies. #: 16324 S1/NonTech Cust. Serv. 02-Nov-92 06:22:09 Sb: #16279-INSTALL NON_SCSI PRELIM Fm: Schroeder 74746,11 To: Schroeder 74746,11 OK...so I answered my own question... For those that might experience similiar problems, here's the solution. I removed the SyQuest Proprietary SCSI controller from my system and repeated the DOS installation procedure. This time it work just fine. Apparently, if WINNT finds a SCSI controller, it assumes that device 0 on that contrller will be the CD-ROM drive, Unfortunately, this is not a vaild assumtion on my system. I am now plagued with two remaining problems. The first one existed in the Pre-Release and is still present. The second is new. First. After the initial boot, just as the screen turns to blue background text mode (for file system validation), the NumLock light on my keyboard is switched off and the keyboard is no longer accessible to the system. When the Logon prompt appears, the keyboard lights are still off and no keypress, including the requested Crtl-Alt-Del, has any effect. Even worse, this problem is intermittent...50% of the time it occurs and requires a system cold boot (system reset button will do)...50% of the time the system behaves just fine (no flashing keyboard lights or nuttin'). I have been unable to assertain any other relavant criteria for this failure as it occurs from a cold start with the same intermittent behavior... Any suggestions? Second. The Pre-Release version worked just fine with my ATI VGA Wonder XL mouse. The Preliminary Release will not recognize it at all. The Express install does setup the system for a Microsoft Inport Mouse, so I assume it found some hardware signature somewhere. Help!?! Thanks for your patience and assistance. Schroeder There is 1 Reply. #: 16467 S1/NonTech Cust. Serv. 03-Nov-92 09:50:45 Sb: #16324-INSTALL NON_SCSI PRELIM Fm: Stu Wiley 70473,1351 To: Schroeder 74746,11 Schroeder, Still working... Stu Wiley #: 16342 S1/NonTech Cust. Serv. 02-Nov-92 10:51:57 Sb: #16279-INSTALL NON_SCSI PRELIM Fm: Stu Wiley 70473,1351 To: Schroeder 74746,11 (X) Schroeder, You may find better answers on the WINNT forum section 8, which deals specifically with hardware compatabilities. I'll also check with some folks here to see how we can assist you. Please make sure you post this in the WINNT forum as well. Stu Wiley Developer Service Team #: 16486 S1/NonTech Cust. Serv. 03-Nov-92 12:17:07 Sb: INSTALL NON SCSI PROBLEM Fm: Stu Wiley 70473,1351 To: schroeder 74746,11 Schroeder, The advice I received was to have you repost this message in the WINNT forum section 3. Sounds like a driver problem and that's where the experts are. they can assist and redirect you from there, hopfully it's somthing they've seen before. Stu Wiley Developer Service Team #: 16512 S1/NonTech Cust. Serv. 03-Nov-92 16:30:15 Sb: net start/stop [driver] Fm: Michael Hall 72607,3422 To: Paul Sanders [MS] 72350,2053 (X) Paul Sanders: I am writing an NT device driver and would like to be able to dynamically load and unload it using "net start " and "net stop ". If is a monolithic device driver then the driver successfully loads and unloads. Unfortunately, my device driver is layered. My class driver successfully unloads but I am unable to reload it. This is the error that it gives me the first time: System error 5 has occurred. Access Denied. This is the error that it gives me every other time: System error 2 has occurred. The system cannot find the file specified. The information printed out by my class driver indicates that it was unable to get the port device driver object. I believe that this is because when the class driver is unloaded, I don't undo the effect of the IoGetDeviceObjectPointer() call that my class driver's DriverEntry() uses to establish a connection to my port driver. If I try to unload my port driver then the system seems to go through some timeout period and then claims that it has unloaded my port driver: The BRUTUS service is stopping...... The BRUTUS service was successfully stopped. However, the port driver's unload routine is never called and when I subsequently try to load my port driver I get the following error: The requested service has already been started. More help is available by typing NET HELPMSG 2182. I believe that the port driver is never really unloaded, but the net program incorrectly reports that it was. I also believe that the reason that the port driver is unable to unload is because it thinks that it is still being used because the class driver was unable to undo the effect of its IoGetDeviceObjectPointer() call. Please help me understand what is going wrong and explain how I can fix it. Thanks for your time, Clay Bean. #: 16501 S1/NonTech Cust. Serv. 03-Nov-92 15:13:48 Sb: NT and Networks Fm: Michael L. Giancola 76416,1464 To: All Hello: Well, I finally have the correct hardware up and running to give NT's network capabilities a try. I am new to LANs so please be kind. Here goes... I want to run NT on the server. The work stations are currently running DOS 5.0 on one machine and Windows 3.1 on the other. If I can limit my purchases to one software package, what would it take to start the machines talking? Do I need to run NT on all stations? What about LAN Manager? Any and all help is appreciated. Confused in LANtown... Michael There is 1 Reply. #: 16540 S1/NonTech Cust. Serv. 03-Nov-92 20:13:17 Sb: #16501-NT and Networks Fm: Jerry Drain [Microsoft] 72350,2056 To: Michael L. Giancola 76416,1464 (X) Hi Michael, NT will work on a machine as a server, and will allow DOS clients to connect to it. You don't have to be running NT on the work stations. For more information, and follow up questions, please post to the Networking Services section (WinNT, Section 11). Thanks, -- Jerry [Win32 SDK Developer Support] #: 16500 S1/NonTech Cust. Serv. 03-Nov-92 15:11:11 Sb: Can't "install" SDK Fm: Edward C Llewellyn 76360,3564 To: sysop (X) I successfully installed the new (Oct.) Windows NT release, and the Win32 SDK, but I can't get the icons to appear when I set up a new user account. I tried using Windows NT Setup program with the Options|Setup Applications... command, but this won't make the icons for him, even after logging out and in again. I also tried the Options|Add/Remove Windows components... but this seems to have nothing to do with the SDK. I'm using a 0x486 PC, by the way. --Ed There is 1 Reply. #: 16541 S1/NonTech Cust. Serv. 03-Nov-92 20:13:23 Sb: #16500-Can't "install" SDK Fm: Jerry Drain [Microsoft] 72350,2056 To: Edward C Llewellyn 76360,3564 (X) Hi Ed, Sorry, your question is in the wrong forum and wrong section. This is for non-technical issues for the SDK (disk replacements, that sort of thing). Please repost your question in the WINNT forum. I think the most appropriate section would be the Utilities and Applets section (Section 7). It looks like you're having some trouble with the User Manager. Thanks, -- Jerry [Win32 SDK Developer Support] #: 16426 S1/NonTech Cust. Serv. 02-Nov-92 19:13:58 Sb: Logitech Mouse w/NT Fm: Trey C. Johnson 71213,1416 To: Doug Olson [Microsoft] 72350,2635 (X) Doug, Hi, this is the first message I've sent on this forum. I hope you don't mind, I got your name & address from another mail. Is there a generic destination for NT questions? I've installed the new October release of NT and it doesn't recognize my Logitech Bus Mouse. Any ideas? Thanks, Trey Johnson There are 2 Replies. #: 16452 S1/NonTech Cust. Serv. 03-Nov-92 06:26:57 Sb: #16426-Logitech Mouse w/NT Fm: Arthur Knowles 71041,2613 To: Trey C. Johnson 71213,1416 Trey, <> Last I heard NT only supported 100% MS compatible mice. I do not believe the Logitech mouse fits this description. You might talk to logitech and see what their NT plans are. Art #: 16543 S1/NonTech Cust. Serv. 03-Nov-92 20:13:37 Sb: #16426-Logitech Mouse w/NT Fm: Jerry Drain [Microsoft] 72350,2056 To: Trey C. Johnson 71213,1416 Hi Trey, No, we really don't have a generic place to ask NT questions. We try to devide the question between two forums, with multiple sections. The reason that we do this is so that we can locate our expertise in the appropriate sections; and so you can go to a section and look at postings related to your problem (instead of having to look at hundreds of LAN questions to solve your compiler issues etc.) The MSWIN32 forum is for SDK and application development questions. The WINNT forum is more for general questions about Windows NT. The section you're looking for specifically is the WINNT, Section 3 "Setup". They can address your issues for you. Another thing to note, is that you don't have to address your post to anyone specific at Microsoft. If you address it to Microsoft, or SysOp; someone from Microsoft will generally try to get involved in the thread. Thanks, -- Jerry [Win32 SDK Developer Support] #: 16387 S1/NonTech Cust. Serv. 02-Nov-92 15:26:24 Sb: Can't inst. W/NEC CD Dr. Fm: John Brooks 70214,3533 To: all Having problem installing from CD ROM Drive "NEC CDR35" ? Keep getting message no cdrom present??? /exit There are 2 Replies. #: 16450 S1/NonTech Cust. Serv. 03-Nov-92 06:22:09 Sb: #16387-Can't inst. W/NEC CD Dr. Fm: Arthur Knowles 71041,2613 To: John Brooks 70214,3533 (X) John, <> It appears that you have either an unsupported SCSI adapter or CD-ROM drive. But you can install NT with the manuall (WINNT) method and then the SDK with the manual (MANUAL) setup apps. They are documented in the release notes. Art #: 16544 S1/NonTech Cust. Serv. 03-Nov-92 20:13:43 Sb: #16387-Can't inst. W/NEC CD Dr. Fm: Jerry Drain [Microsoft] 72350,2056 To: John Brooks 70214,3533 Hi John, Art is right, the NEC CDR35 doesn't appear to be on the hardware compatibility list yet. For follow up questions, or more information; please post your setup questions in the WINNT forum, Section 3 "System Setup" Thanks, -- Jerry [Win32 SDK Developer Support] #: 16435 S1/NonTech Cust. Serv. 02-Nov-92 22:13:11 Sb: NT for pen-computers? Fm: David A Hand 70621,3624 To: ALL Does anyone know if Microsoft intends to have a version of NT running on pen-computers? I had assumed the memory requirements would preclude this, but I recently read somewhere that they are experimenting with a ROM verson of NT for portables. Will this include the pen extensions? Thanks, Dave Hand There is 1 Reply. #: 16468 S1/NonTech Cust. Serv. 03-Nov-92 09:50:49 Sb: #16435-NT for pen-computers? Fm: Stu Wiley 70473,1351 To: David A Hand 70621,3624 (X) Dave, It is my understanding that we are working on supporting teh same things with NT as we do with Windows. I have no specific information, but undoubtedly this is being addressed somewhere within MS. You may want to ask these theoretical questions in the WINEXT forum...where the PEN folks reside. They may have a better handle on the specific issues. Stu Wiley Developer Service Team There is 1 Reply. #: 16558 S1/NonTech Cust. Serv. 03-Nov-92 23:33:27 Sb: #16468-NT for pen-computers? Fm: David A Hand 70621,3624 To: Stu Wiley 70473,1351 (X) Stu, Thanks, I'll do that. #: 16592 S1/NonTech Cust. Serv. 04-Nov-92 10:35:34 Sb: CMD white space for BAT Fm: Chuck Atwood 76004,1602 To: sysop (X) The new command interpreter CMD.EXE apparently does not treat horizontal tabs (0x09) as white space. Executing the following batch file creates error message "cannot find the label setnt300". The batch file will run correctly under DOS 5.0, but not in the MSDOS window under NT. The characters follow "goto setnt300" in line 3. echo off if "%1"=="" goto help if "%1"=="nt300" goto setnt300 goto help :setnt300 set CADVER=300 set WINVER=0x0314 set CADSRC=\SRCNT300 goto exit :help echo Copyright (c) 1992 Foresight Resource Corporation echo CADSET [os-ver] [country] :exit Chuck Atwood (Foresight Resources Corporation) #: 16595 S1/NonTech Cust. Serv. 04-Nov-92 10:37:22 Sb: SUBST problem Fm: Chuck Atwood 76004,1602 To: sysop (X) Problem with the SUBST command using the MSDOS windows under NT: The copy command won't work when copying a file from one SUBST drive to another ("cant find device"): SUBST I: \SRCWN210 SUBST J: C:\SRCNT300 copy I:a.b J:a.b Chuck Atwood (Foresight Resources Corporation) #: 16576 S1/NonTech Cust. Serv. 04-Nov-92 07:33:25 Sb: DDK offer, when ? Fm: Jeong Ho Lee 70253,1244 To: SYSOP (X) Hello, I heard from this Forum that MS was going to send SDK users DDK offer ( discount ? ) If it's a discount offer, do we have to wait for it ? I mean, can we just call MS right now and get DDK after verifying ourselves SDK owner ? What's going on ? Thanks jLee There is 1 Reply. #: 16624 S1/NonTech Cust. Serv. 04-Nov-92 13:22:30 Sb: #16576-DDK offer, when ? Fm: Dwight Matheny/Microsoft 70750,2340 To: Jeong Ho Lee 70253,1244 (X) It's not a discount offer. We will mail the info on the DDK before it ships, in plenty of time to place an order. We're asking that you not call to place an order right now because our Developer Services Team is swamped w/ calls for the SDK, MSDN, and the new Visual Basic 2.0. -Dwight (MS) #: 16569 S1/NonTech Cust. Serv. 04-Nov-92 05:40:47 Sb: Oct. release in Europe Fm: Tom Doust 100015,3011 To: MSFT What's the deal about european customers getting october release. I got my July PDK by going to the European developers conference and I kind of assumed that I would get the next release automatically. I'm now reading all these messages from people saying how much better the new version is and I haven't got it! Do I have to do something or will it happen automatically - just a bit slower than in the US. There is 1 Reply. #: 16625 S1/NonTech Cust. Serv. 04-Nov-92 13:22:35 Sb: #16569-Oct. release in Europe Fm: Dwight Matheny/Microsoft 70750,2340 To: Tom Doust 100015,3011 (X) It will happen automatically. Our European subs started receiving their SDKs a few days ago. -Dwight (MS) #: 16292 S1/NonTech Cust. Serv. 01-Nov-92 09:00:15 Sb: DDK Availability Fm: Anthony Wilkinson 100010,2312 To: Sysop (X) Hi, I'm in the UK, and I 'phoned MS UK about getting the CD only DDK, and was told this was not possible. I only need the DDK to write 'personal' drivers for some lab equipment - I don't want to spend 500 pounds ($800) going to a conference! Can you please advise on how I can get hold of this CD? Thanks, Anthony. There are 3 Replies. #: 16336 S1/NonTech Cust. Serv. 02-Nov-92 10:07:15 Sb: #16292-DDK Availability Fm: Dwight Matheny/Microsoft 70750,2340 To: Anthony Wilkinson 100010,2312 (X) I've sent mail to MS UK to see what's going on here. Perhaps they're just not taking orders yet since the product won't ship for several weeks. -Dwight (MS) There is 1 Reply. #: 16361 S1/NonTech Cust. Serv. 02-Nov-92 12:40:48 Sb: #16336-DDK Availability Fm: Anthony Wilkinson 100010,2312 To: Dwight Matheny/Microsoft 70750,2340 (X) DM>I've sent mail to MS UK to see what's going on here. Perhaps they're just not DM>taking orders yet since the product won't ship for several weeks. Cheers Dwight - I look forward to the outcome. Anthony. #: 16389 S1/NonTech Cust. Serv. 02-Nov-92 15:38:10 Sb: #16292-DDK Availability Fm: Graham Welland 70023,1267 To: Anthony Wilkinson 100010,2312 (X) I suspect that the response to your query is to that you should wait until the product goes retail. On the other hand, the NT DD conference should be interesting.... certainly better value than the UK windows developer conference. Graham (in UK) #: 16474 S1/NonTech Cust. Serv. 03-Nov-92 10:49:45 Sb: #16292-DDK Availability Fm: Dwight Matheny/Microsoft 70750,2340 To: Anthony Wilkinson 100010,2312 (X) Anthony, I got a response from MS UK. The situation there is the same as here- the DDK was made available first at the conference, but is also orderable by people not attending the conference. We will be shipping the DDK to non-conference attendees as soon as the docs are ready to go. You should be able to order this from MS UK now. -Dwight (MS) There are 2 Replies. #: 16496 S1/NonTech Cust. Serv. 03-Nov-92 14:55:21 Sb: #16474-DDK Availability Fm: Christian Beaumont 100034,2326 To: Dwight Matheny/Microsoft 70750,2340 (X) Errrm, I'm going to the DDK conference this Thursday (5th Nov ) and I am destined to get a copy of the DDK. However I just read that there is to be a NEW DDK in the next week, how long will that take to filter through to the UK developer (2nd class citizens) ? Any ideas? Christian There is 1 Reply. #: 16632 S1/NonTech Cust. Serv. 04-Nov-92 13:45:12 Sb: #16496-DDK Availability Fm: Dwight Matheny/Microsoft 70750,2340 To: Christian Beaumont 100034,2326 There is no new DDK. We cut special CDs without the docs for the two developers conferences. The CDs that everyone else will be receiving will be exactly the same except for the online docs (which weren't ready in time for the conference.) -Dwight (MS) #: 16513 S1/NonTech Cust. Serv. 03-Nov-92 16:40:32 Sb: #16474-DDK Availability Fm: Anthony Wilkinson 100010,2312 To: Dwight Matheny/Microsoft 70750,2340 (X) Dwight, I can't order it now, 'cos I've tried. Do you have a date for when it will be ready (to within a week or to) so I can try again around then? Thanks, Anthony. There is 1 Reply. #: 16631 S1/NonTech Cust. Serv. 04-Nov-92 13:45:08 Sb: #16513-DDK Availability Fm: Dwight Matheny/Microsoft 70750,2340 To: Anthony Wilkinson 100010,2312 No, I don't know the UK dates. My goal is to ship in the US before the end of the month. My contact at MS UK told me they are now taking orders- I'll send him your message to let him know that's not true. -Dwight (MS) #: 16290 S1/NonTech Cust. Serv. 01-Nov-92 04:12:09 Sb: NT BETA !!!!! Fm: Derek Hellmons 100022,1636 To: ALL Hi I'v been calling Microsoft in munchen (germany) trying to get a lead where I can get a hold of the Windows NT beta I did a lot of work on the win 3.1 beta ,even helped Microsoft get the german OEM version out on time and now ....I don't now sir please call next week....is this the way microsoft wants to become the world leader.... If any one can give me a number any where in the World where I can get a Hold of the NT beta then thanks in advance.. Derek. There is 1 Reply. #: 16343 S1/NonTech Cust. Serv. 02-Nov-92 10:52:02 Sb: #16290-NT BETA !!!!! Fm: Stu Wiley 70473,1351 To: Derek Hellmons 100022,1636 (X) I'll forward your request to the Munchen SUB. they typically have a time lag behind the U.S. so be patient. Stu Wiley Developer Service Team There is 1 Reply. #: 16377 S1/NonTech Cust. Serv. 02-Nov-92 14:36:20 Sb: #16343-NT BETA !!!!! Fm: Derek Hellmons 100022,1636 To: Stu Wiley 70473,1351 (X) Thanks But whats his name,I mean the guy in munchen so I can bug him,Iknow that not everyone at MS looks at there email everday (or was that week) so that I can Bug him.We have many Large customers interesed,Insurance companys, Banks and little old MAN (largest truck manufacture) and all we have to show is a copyed SDK....which crashes they want to Spend Mega DM in Downsizing and all we hear is "oh I'm sorry We don't have anyone here you can help you sir..." Thanks in advance Derek... There is 1 Reply. #: 16641 S1/NonTech Cust. Serv. 04-Nov-92 15:32:49 Sb: #16377-NT BETA !!!!! Fm: Stu Wiley 70473,1351 To: Derek Hellmons 100022,1636 (X) The product management isn't on CIS. I'll ask for a phone call if you want to post the number. Stu Wiley Developer Service Team #: 16679 S1/NonTech Cust. Serv. 04-Nov-92 20:11:09 Sb: Windows NT Hardware Fm: Mark Westerhaus 72427,2000 To: Mark Westerhaus How can I obtain the Hardware Compatibility List for Windows NT? There is 1 Reply. #: 16768 S1/NonTech Cust. Serv. 05-Nov-92 16:10:44 Sb: #16679-Windows NT Hardware Fm: Dwight Matheny/Microsoft 70750,2340 To: Mark Westerhaus 72427,2000 (X) This should be in library 17, 1092hw.txt. -Dwight (MS) #: 16779 S1/NonTech Cust. Serv. 05-Nov-92 18:01:46 Sb: Shareware Editor Fm: Dan Sodhi 71212,2333 To: All Does any one know of a Windows based (NT would be better) Programmer's Editor I can download. #: 16593 S1/NonTech Cust. Serv. 04-Nov-92 10:36:25 Sb: NMAKE problems Fm: Chuck Atwood 76004,1602 To: sysop (X) NMAKE doesn't check dependencies correctly for directories. The following makefile works using NMAKE from C7 with DOS 5.0 but will display "'CHUCK' is up-to-date" in the MSDOS window under NT. I am using the FAT file system. Make a directory CHUCK and type: NMAKE -f NTBAD.MAK. # NTBAD.MAK # author: cla DIRS = CHUCK $(DIRS): cd $(*) copy ..\ntbad.mak . cd.. # [EndOfMake] P.S. NMAKE needs !ELSE which is available in 3.1/C700 SDK. Chuck Atwood (Foresight Resources Corporation) #: 16790 S1/NonTech Cust. Serv. 05-Nov-92 19:10:39 Sb: NMAKE problems Fm: Chuck Atwood 76004,1602 To: Jerry Drain [Microsoft] 72350,2056 (X) Sorry, sending my NMAKE problem to Tools was my intention. This is my first adventure in compuserve land. #: 16351 S1/NonTech Cust. Serv. 02-Nov-92 11:36:54 Sb: Shareware contest Fm: Robert Zeff 70323,1251 To: All I have not received confirmation of my entry into the shareware contest. Has anyone else? Thanks, -Robert There is 1 Reply. #: 16528 S1/NonTech Cust. Serv. 03-Nov-92 17:53:45 Sb: #16351-Shareware contest Fm: Daniel Green 70312,3470 To: Robert Zeff 70323,1251 (X) I also submitted an entry to the shareware contest, and I haven't received any acknowledgement or info from Microsoft about it (not that I expected any, but it would nice to know that they at least received the package). By the way, my app doesn't run on the October release - it is looking for "USERRTL.DLL" which doesn't exist (apparently) in the October release. Looks like I must have linked with the wrong run-time library libs or something. Oh well. I guess 24 hours before Comdex I'll get a call asking me to recompile and resubmit my app... -- Dan There are 2 Replies. #: 16572 S1/NonTech Cust. Serv. 04-Nov-92 05:57:35 Sb: #16528-Shareware contest Fm: Bill Cohagan 74375,313 To: Daniel Green 70312,3470 Dan- | ... By the way, my |app doesn't run on the October release - it is looking for "USERRTL.DLL" which |doesn't exist (apparently) in the October release. Looks like I must have |linked with the wrong run-time library libs or something. I've seen others with this problem and, as I recall, it indicates that you have not recompiled your app under the Oct release. If you *have* then you might double check you makefile to see if one or more modules got left out. I think all it takes is one module not recompiled and you'll get this error. As I understand it this DLL doesn't exist anymore (obviously!) and its components have been distributed among other DLLs that *do* exist. Hope this helps. Bill #: 16587 S1/NonTech Cust. Serv. 04-Nov-92 09:56:22 Sb: #16528-Shareware contest Fm: Len Gray 71630,1703 To: Daniel Green 70312,3470 I contacted Microsoft regarding shareware contest entries and the October release. All apps have the same problem - there were changes made to the library call formats. I was told that all entries would be judged under the July pre-release so that no rebuilds are required. Whether or not MS will ask for rebuilds of apps following the judging, I'm not sure. Len There is 1 Reply. #: 16796 S1/NonTech Cust. Serv. 05-Nov-92 19:53:14 Sb: #16587-Shareware contest Fm: Jerry Drain [Microsoft] 72350,2056 To: Len Gray 71630,1703 (X) Hi all, >>I was told that all entries would be judged under the July pre-release so that no rebuilds are required. Whether or not MS will ask for rebuilds of apps following the judging, I'm not sure. I've done some checking around on this. The submissions are being tested under the July release; and although it says in the form that you may be asked to rebuild your sample with a later release, it doesn't look like that is being the case so far. Earlier in this thread someone asked about confirmation for the submissions. Apparently confirmations started going out only just a day or so ago. If you haven't received a confirmation for your submission, it's nothing to worry about at this point. Thanks, -- Jerry [Win32 SDK Developer Support] #: 16546 S1/NonTech Cust. Serv. 03-Nov-92 21:22:47 Sb: WIN32 PROGRAMMING Fm: Michael Trier 71044,2032 To: ALL I am a Visual Basic programmer and before that in Turbo Pascual. I am interested in writing Windows programs in C (With the big boys). I would like to start out writing 32 Bit programs, to avoid the transition from 16 Bit Windows. I have the Win32 SDK CD-ROM. What else do I need to begin writing 32 bit Windows programs. I will need some C instruction, I've messed around with C in the past but I'm not strong in it. Any help would be appreciated Thanx Michael There is 1 Reply. #: 16758 S1/NonTech Cust. Serv. 05-Nov-92 15:17:57 Sb: #16546-WIN32 PROGRAMMING Fm: Stu Wiley 70473,1351 To: Michael Trier 71044,2032 (X) Michael, I strongly suggest that you reconsider your order of development priority. Although with the WIN32SDK (32S) you can create 32 apps and run them on the 16 bit platforms, WINNT is probably best utilized through production of a WIndows 3.1 16 bit application first. Not only more forgiving as a development platform, the skills gained will be readily adaptable to the closely structured WIN32 API, and speed your development effort through a flatter learning curve. (Sort of like graduating the 101 course before tackling the next level.) Microsoft has taken great efforts to bring visual BASIC programmers into the C world through Quick C for Windows, and then teaching you to leverage the skills learned by moving up to the C/C++ 7.0 and the Windows 3.1 SDK. Arguably, this is where the majority of the Windows ISV's are, necessitating Microsoft's attention to this large group needing smooth upward migration. You should consider attending a Microsoft University Windows programming track or C course, the catalog is available at 206/828-1507. Strengthen your skills, and then put them to use with the SDK's. Don't worry, the WIN32SDK CD-Roms will become that much more valuable to you. Anyway, give it a wack, and welcome aboard. Stu Wiley Developer Service Team There is 1 Reply. #: 16800 S1/NonTech Cust. Serv. 05-Nov-92 20:59:02 Sb: #16758-WIN32 PROGRAMMING Fm: Michael Trier 71044,2032 To: Stu Wiley 70473,1351 (X) Stu, <> Thanks for the info. After alot of reviewing today, I also came to your conclussion that Quick C for Windows would be a good place for me to start and then move on to MS C/C++ 7.0 and the Win 3.1 SDK. I'm looking forward to the world of Windows programming. Sincerely, Michael #: 16804 S1/NonTech Cust. Serv. 05-Nov-92 22:44:04 Sb: July NT Video Drivers... Fm: Steve Gibson 76067,4106 To: ALL Hello Anyone! I'm going NUTS looking at this 640x480 VGA standard res screen. I've got S3 and Tseng4000 based boards ... but NT's setup won't give me these options. I can even see some tasty DLL's named TS_1024 and V7_1024 ... but they're "beyond reach" !!! Has ANYONE got any idea how to get the JULY version of NT (October worked fine, but I **MUST** use JULY for judging the NT Shareware contest for COMDEX) to ask me for more video card options?????????? HELP! -----> - Steve Gibson (InfoWorld Magazine) (SpinRite) There is 1 Reply. #: 16805 S1/NonTech Cust. Serv. 06-Nov-92 01:05:42 Sb: #16804-July NT Video Drivers... Fm: Peter Handsman[Inmark] 70550,2570 To: Steve Gibson 76067,4106 Steve, It isn't as easy as a config dialog box... If you read the July release notes, pages 54-55. It has descriptions of how to use et4000 and video-7 boards. It's basicly a matter of copying the .sys file to vga.sys, and the correct .dll file to vga.dll. (if running nt you must ren the old files so that you don't get sharing violations) Since you mentioned the et4000.... et400_60.sys --> vga.sys 60hz et400_70.sys --> vga.sys 70hz pdii.sys --> vga.sys Orchard ProDesigner IIs And ts_1024.dll --> vga.dll 1024x786 ts_800.dll --> vga.dll 800x600 Good Luck. Peter #: 16314 S1/NonTech Cust. Serv. 01-Nov-92 17:25:45 Sb: WIN32 Shipment == WfW ! Fm: Mark Gibbons 76216,1032 To: Wayne C. Cannon 72330,177 (X) Actually, Wayne, the WFW disks probably are worth more than the CD. Hopefully it's WFW 2.0. A great software product, from what I understand. There is 1 Reply. #: 16806 S1/NonTech Cust. Serv. 06-Nov-92 01:21:10 Sb: #16314-WIN32 Shipment == WfW ! Fm: Wayne C. Cannon 72330,177 To: Mark Gibbons 76216,1032 (X) Unfortunately I 1) Already have WfW 2.0, and 2) am pretty insistant on keeping straight-up in all my dealings with vendors, including Microsoft. Unfortunately, with the very significant discounts being offered to and demanded by large accounts, we little guys are the only ones paying significant bucks for application software products like WfW. Thanks. --Wayne #: 16702 S1/NonTech Cust. Serv. 05-Nov-92 04:25:20 Sb: Getting old messages Fm: Randy Threewits 71005,1373 To: sysop (X) Is there a way to retrieve old messages? I was in a message thread addressing a problem with the last message number of 15455. I went to the DDK conference and on vacation. And it now that I'm back it seems that I have missed any responses to my last message. If these old messages are in an archive somewhere, please point me there. I found in the WINNT forum the archive library, where they have a ZIP file of all the messages in WINNT for each week. If there does not exist such a library for MSWIN32, it would be a good idea to add one. Thanks. There is 1 Reply. #: 16781 S1/NonTech Cust. Serv. 05-Nov-92 18:04:43 Sb: #16702-Getting old messages Fm: Jerry Drain [Microsoft] 72350,2056 To: Randy Threewits 71005,1373 (X) What section were you in Randy? I know that Otto Fung was zipping them and uploading messages to one of the libraries; but I haven't been up there nor have I heard anything lately about this, so I don't know where they are on this. However, the internal tool that I use is able to retrieve some of the messages that have scrolled off. If you like, I can grab them and paste them into a post to you. I do have to know what section they are in though so that this tool can find them. Thanks, -- Jerry [Win32 SDK Developer Support] There is 1 Reply. #: 16807 S1/NonTech Cust. Serv. 06-Nov-92 03:59:19 Sb: #16781-Getting old messages Fm: Randy Threewits 71005,1373 To: Jerry Drain [Microsoft] 72350,2056 (X) The section of the message was #13 (NT DDK), the message number that I needed and any resulting responses was #15455. Thanks for bailing me out, I guess I shouldn't take vacations eh? #: 16672 S1/NonTech Cust. Serv. 04-Nov-92 18:49:32 Sb: NT Speaker Fm: Douglas MacDonald 70760,20 To: sysop (X) I am involved in the /cny UNIX Users Group and would like to have someone knowledgeable of Windows NT speak at our December 10th UNIX User Group meeting. I have spoken with a Bryan Moran at Microsoft and he indicated that there was perhaps someone in Ithaca New York who could do it. It has been a couple weeks since I last spoke with Bryan and have left several voice mails on his phone but he has never returned them. Is there someone available to speak on NT? We are a small group (60 members, 300 mailing list) and are not able to pay for the speaker or the transportation. Bryan seemed to indicate that Microsoft had "Evangelists" just for this type of NT presentation. Thanks Doug MacDonald There is 1 Reply. #: 16821 S1/NonTech Cust. Serv. 06-Nov-92 10:47:00 Sb: #16672-NT Speaker Fm: Stu Wiley 70473,1351 To: Douglas MacDonald 70760,20 Doug, I sent your request on to the various parties who handle speaking engagements relative to NT. I'm waiting for a response and will fill you in as soon as I hear anything. Stu Wiley Developer Service Team #: 16302 S1/NonTech Cust. Serv. 01-Nov-92 12:25:44 Sb: Oct. rel. in Canada Fm: Louis Bouchard 72301,3565 To: all When canadian developers are supposed to receive the october release of Windows NT ? and could you please check if my name is still on your mailing list: Louis Bouchard 179 Des Jesuites Chicoutimi, PQ G7H 3B4, Canada Thanks There is 1 Reply. #: 16399 S1/NonTech Cust. Serv. 02-Nov-92 16:08:07 Sb: #16302-Oct. rel. in Canada Fm: Stu Wiley 70473,1351 To: Louis Bouchard 72301,3565 (X) Checking for you, looks liketwo weeks or so. You are still on our mailing list. Stu Wiley Developer Service Team There are 2 Replies. #: 16802 S1/NonTech Cust. Serv. 05-Nov-92 21:43:44 Sb: #16399-Oct. rel. in Canada Fm: Jonathan Carroll 74017,3242 To: Stu Wiley 70473,1351 (X) I too am on the Canadian list or should be, for the Win NT Beta. My name is Jonathan Carroll, Montreal Canada. Could you check if I'm still listed? We've got a large Novell base (87 servers connected by fiber optic backbone) and would like to make sure the stuff we've already written is still ok. Jon. #: 16853 S1/NonTech Cust. Serv. 06-Nov-92 15:45:18 Sb: #16399-Oct. rel. in Canada Fm: Louis Bouchard 72301,3565 To: Stu Wiley 70473,1351 (X) Thanks, I hope i will have less problem to install this new release than I had to install the last one... anyway we will see in a few weeks. Louis #: 16870 S1/NonTech Cust. Serv. 06-Nov-92 18:00:59 Sb: Win32 SDK installation Fm: Ron Simmons 76615,3452 To: SYSOP I just installed NT on an 80386/40 system which uses and Adaptek 1542 controller to support a Segate 296N (80Mb drive) and a Syquest 555 (40 Mb removable). I used a Toshiba XM3301 to install the system from the October 1992 release. Everything worked fine until I tried to install the SDK portion. I had wanted to set it up on the Syquest device, but the installation will not recognise the drive (the rest of NT works just fine with it). Any suggestions?... Thanks, Ron Simmons (76615,3452) #: 16610 S3/SQL Server SDK 04-Nov-92 11:23:21 Sb: SQL server on NT:TCP/IP Fm: Mike Snowden 100021,3015 To: James McDaniel [MS] 71075,415 We would like to be able to use the TCP/IP link from a SYBASE open client interface directly into an NT SQL server; the cost of including the SQL bridge means that we might as well go SYBASE all the way. Any comments? #: 16298 S4/API-User Interface 01-Nov-92 11:49:48 Sb: FM extension Fm: Mark Gamber 76450,2754 To: 76450,2754 (X) I'm writing a File Manager extension that creates a small window used for drag-drop to delete files. The menu hooks in, the window is created and when I drag selected files over the window, I get the little "document" cursor. The problem is, I never get a WM_DROPFILES message from the shell. I've tried the CreateWindowEx using WS_ACCEPT_FILES and DragAcceptFiles() methods, both to no avail. Any idea if something's wrong internally? I can upload source for examination if you'd like. Mark #: 16303 S4/API-User Interface 01-Nov-92 12:58:15 Sb: FM extension (pt 2) Fm: Mark Gamber 76450,2754 To: 76450,2754 (X) Update to previous Drag-Drop problem: When I run Spy and spy on the FM extension's window, it gets the WM_DROPFILES message and everything works ok. Exit Spy and it stops getting the message. I used Pview before during and after and the extension and SHELL32 is loaded into Winfile's address space, so everything seems to be there....just no working. Mark #: 16332 S4/API-User Interface 02-Nov-92 09:44:38 Sb: Win32s programing in 3.1 Fm: Paul Ligeski 76636,1166 To: Sysop (X) Hi, Is there a way to use the Win32s libraries under Windows 3.1? Currently, I cannot get 3.1 apps to run under Windows NT. --Paul There is 1 Reply. #: 16395 S4/API-User Interface 02-Nov-92 15:59:11 Sb: #16332-Win32s programing in 3.1 Fm: Steve Firebaugh [MS] 75430,412 To: Paul Ligeski 76636,1166 (X) Paul, The software that you are using to access CIS may be preventing you from seeing sections 13 & 14 since they were probably closed when you first joined this forum. Both of these sections ARE open at this point. Section 14 is the correct place to get Win32s help. Look for messages 16209 or 16330. These were both posted in section 14. >Is there a way to use the Win32s libraries under Windows 3.1? Yes, the October beta release provides the technology necessary for you to run applications built for the Win32 API on Windows 3.1. See page 11 of the release notes for more information on win32s. See the \mstools\win32s directory on the compact disk for the relevant files. >Currently, I cannot get 3.1 apps to run under Windows NT. This is a separate issue from Win32s. If you just have the executable files for the 3.1 applications, I'd recommend asking for help in section 4 of the WINNT (end user issues) forum. If you are talking about your own applications that you have source code for, you are welcome to ask the questions here. It would help if you could be more specific. Thanks, Steve Firebaugh #: 16396 S4/API-User Interface 02-Nov-92 15:59:16 Sb: #16190-DLGTEMPLATE Fm: Steve Firebaugh [MS] 75430,412 To: Steve Firebaugh [MS] 75430,412 (X) To: Anyone interested in the win32 dlgtemplate format. I just uploaded a small sample to library 4 which shows how to dynamically create a dialog from a dlgtemplate structure in memory. The file is named DYNDLG.ZIP, and includes the c file and a makefile. Howard Myers was correct (perhaps this was no surprise to anyone). Even CreateDialogIndirectA() expected the strings in the dlgtemplate to be unicode strings. I hope that you find this sample useful. Steve Firebaugh #: 16600 S4/API-User Interface 04-Nov-92 10:52:41 Sb: UDP datagrams Fm: John Dinga 75300,1745 To: sysop (X) I asked this question of the WINNT forum and they sent me here. I would like write a program under NT that sends and receives UDP datagrams. Does anyone have any example code. Is this covered in any of the manuals I received with my NT beta kit. Thanks, Max. There is 1 Reply. #: 16621 S4/API-User Interface 04-Nov-92 13:18:46 Sb: #16600-UDP datagrams Fm: David Taniguchi [MS] 72350,2054 To: John Dinga 75300,1745 (X) Hi John, >I would like write a program under NT that sends and receives UDP >datagrams. Does anyone have any example code. Is this covered in any >of the manuals I received with my NT beta kit. You can use the Windows Sockets DLL to send datagrams. This provides API is based on Berkeley sockets. Here is a small example: { PPROTOENT pprotoent; SOCKET s; int cc; pprotoent = getprotobyname("UDP"); s = socket( AF_INET, SOCK_DGRAM, pprotoent->p_proto); cc = sendto( s, buf..) } In the Windows Sockets implementation they provide Windows Specific Api (WSA prefixed) functions which anables the Windows programmer to receive notice of socket events via the message queue. If you have any follow up questions, please post in section 12 of Forum MSWIN32 (RPC/WinNet). Thanks, Dave #: 16397 S4/API-User Interface 02-Nov-92 15:59:22 Sb: Resource formats Fm: Steve Firebaugh [MS] 75430,412 To: All There have been several inquiries in this section about dialog templates and menu templates in Win32. The binary formats have been widened, and are no longer the same as they were in Windows 3.1. I have just uploaded a file which documents the win32 resource formats. The file is RESFMT.ZIP, in library 4. It includes a WinWord DOC file as well as a plain text TXT file. Notice that these formats are not documented anywhere on the distributed compact disk. This is currently the only place that you can get them. (My understanding is that the printed and online references will be updated by the next release to include the information which we are now distributing here.) If anyone has any problems with this file, please let me know. Thanks, Steve Firebaugh There is 1 Reply. #: 16655 S4/API-User Interface 04-Nov-92 16:11:05 Sb: #16397-Resource formats Fm: Samuel Feldman 70403,432 To: Steve Firebaugh [MS] 75430,412 (X) Steve, Thank you for uploading RESFMT.ZIP. Here is some feedback on its contents. Section 2.1.1.5 PUSHBOX is obsolete Section 2.1.2 EXSTYLE The form EXSTYLE= appearing on the DIALOG statement is not supported (apparently). I'm not sure it made sense anyway. Section 2.1.6 MESSAGETABLE What is the format of the message file? Is this really a supported statement? Section 3.5 Additional Header Information Although the data presented here is all correct, the section refers to "additional information" and "the ResAdditional structure." These references should be removed for clarity, since there is a single structure documented. Section 3.5.2 Memory Flags The resource compiler does not ignore the specified settings. Of course they are ignored in the .EXE file (right?). Section 4.2 Icon Resources There is no "padding" field in either the IconHeader or ResourceDirectory structures. Section 4.3 Menu Resources The text should mention that menu items are not dword-aligned. Section 4.4 Dialog Box Resources PUSHBOX is obsolete Section 4.5 Cursor Resources There is no "padding" field in the CursorHeader structure. ResourceDirectory structures do NOT "starts on a DWORD boundary." Section 4.12 Version Resources The structure definitions for all sub-blocks are missing the wLength and wValueLength fields: StringFileInfo, StringTable, String, VarFileInfo, and Var. The szKey[] field in the StringTable structure has a comment that it is 8 bytes. Actually it is 8 characters (16 bytes). Section 4.13 Messagetable Resources The format of these does not match what RC currently does -- which is to copy the content of the message file byte for byte into the [OzCIS: Continued in next msg] There is 1 Reply. #: 16656 S4/API-User Interface 04-Nov-92 16:11:10 Sb: #16655-Resource formats Fm: Samuel Feldman 70403,432 To: Samuel Feldman 70403,432 (X) [OzCIS: Continued from previous msg] .RES file. Is this supported currently? If it will be supported, the fields in the structures need to be documented, especially those that are not obvious (OffsetToEntries, Flags, whether Text[] is null-terminated). That's it for now! -- Samuel #: 16462 S4/API-User Interface 03-Nov-92 09:39:49 Sb: Graying EDIT control Fm: David Manns 100112,2773 To: ALL Having set the initial value of an edit control in the WM_INITDIALOG message handling, how do I get it drawn grayed out when the dialog is first drawn? Setting the edit control to disabled in the dialog editor does not appear to do this. There is 1 Reply. #: 16674 S4/API-User Interface 04-Nov-92 19:16:01 Sb: #16462-Graying EDIT control Fm: Steve Firebaugh [MS] 75430,412 To: David Manns 100112,2773 David Manns, I have just recreated the problem which you describe. For now, a reasonable work around is to call EnableWindow, FALSE with each edit control that you want disabled. I have observed that such edit controls are painted gray, as we would expect. Apparently the text color is set for the HDC only when the edit control receives a WM_ENABLE message (which is a result of calling EnableWindow, FALSE). The edit control does NOT set the text color at create time, based on the WS_DISABLED bit, as it should. This is a bug, I've reported it, and we will try to get it fixed by the next release. Thanks. Steve Firebaugh #: 16461 S4/API-User Interface 03-Nov-92 09:39:45 Sb: rc fails on DLGINCLUDE Fm: David Manns 100112,2773 To: ALL The dialog editor inserts the line: 1 DLGINCLUDE "id.h" in the .DLG file, which causes the resource compiler to fail. Changing the line to be: #include "id.h" allows RC to work okay. Is this a known problem, or am I doing something wrong? There is 1 Reply. #: 16676 S4/API-User Interface 04-Nov-92 19:34:09 Sb: #16461-rc fails on DLGINCLUDE Fm: Steve Firebaugh [MS] 75430,412 To: David Manns 100112,2773 David Manns, I find that I am able to DLGEDIT a res file, allow the DLGEDIT program to write "1 dlginclude file.h" in the .dlg file, and then successfully resource compile the .rc file. However, I will not doubt that you are able to get it to break. The semantics of DLGINCLUDE have changed somewhat in Win32. I am including a "Knowledge Base article" which describes this change. If you feel that this covers the problem you are seeing, please consider the issue reported. If you think that you are seeing a different problem, please respond again with more information which will allow me to reproduce it here. Thanks, Steve Firebaugh INF: Use of DLGINCLUDE in Resource Files ID: Q91697 CREATED: 2-NOV-1992 MODIFIED: 2-NOV-1992 3.10 WINDOWS Summary: The Windows 3.1 SDK Dialog Editor needs a way to know what include file is associated with a resource file that it opens. Rather than prompt the user for the same of the include file, the name of the include file is embedded in the resource file in most cases. Embedding the name of the include file is done with a resource of type RCDATA with the special name DLGINCLUDE. This resource is placed into the .res file and contains the name of the include file. The dialog editor looks for this resource when it loads a .res file. If it is found, then the include file is opened also. If not, the editor prompts the user for the name of the include file. In some Windows 3.1 build environments, the dialog editor was used to create dialogs which were placed in more than one .dlg file. These different .dlg files were then included in one .rc file that was compiled with the resource compiler. So the resource file gets multiple copies of a RCDATA type resource with the same name, DLGINCLUDE, but the resouce compiler and dialog editor do not complain. In the Win32 Preliminary SDK, changes were made so that this resource has its own resource type. It was changed from an RCDATA type resource with the special name, DLGINCLUDE, to a DLGINCLUDE resource type whose [More] There is 1 Reply. #: 16677 S4/API-User Interface 04-Nov-92 19:34:15 Sb: #16676-rc fails on DLGINCLUDE Fm: Steve Firebaugh [MS] 75430,412 To: Steve Firebaugh [MS] 75430,412 (X) [Continued] name can be specified. The dialog editor would look for resources of the type DLGINCLUDE. For the October beta, there was a change made to CvtRes so that it gives an error if it finds a resource that has the same type, name, and language as another resource in the file. We are being stricter about the resources needing to be unique in the Win32 SDK than the Windows 3.1 SDK. This is good, because there was never any guarantee at run time which of the two or more resources would be returned by LoadResource(). This means that some apps being ported to Windows NT will give an error when their resources are compiled, because they will have duplicate RCDATA type resources with the same name (DLGINCLUDE). This error is by design. The work-around is straightforward: delete all the DLGINCLUDE RCDATA type resource statements from all the .dlg files. Finally, because it does not make much sense to have the DLGINCLUDE type resources in the executable at all, CvtRes will strip them out so that they do not get linked into the exe. #: 16270 S4/API-User Interface 31-Oct-92 08:41:40 Sb: Comm Defaults Fm: Kent Stewart 75146,224 To: All I was caught off guard by the comm side of NT. I'm so used to smart modems and comm programms that I missed the lack of baud rates > 19200. What is worse is the choice provide be NT when a high rate is requested. I could understand setting the connection rate to 19200 in these cases but setting the rate to 300 baud left me shaking my head in disgust. I was using Forsberg's zcomm, with the rate set to 38400 and using a ns16550afn, and I kept getting connections at 300 baud. I though I had a modem/port problem until I set the rate to 19200 and I was able to connect to CIS, Genie, and Bix at my normal rates of 2400 to 9600. You really should consider what the default rate is if you don't support the one requested. Since, you are buffering the connection, setting the rate to the very slowest is very expensive since most services (including the phone company) base charges on connection time more than the speed of the connection. Kent There are 2 Replies. #: 16309 S4/API-User Interface 01-Nov-92 14:54:22 Sb: #16270-Comm Defaults Fm: neil colvin 71650,3517 To: Kent Stewart 75146,224 (X) I have been running NT comm at 38400 since the July release with no problem. This sounds like a program problem, not a NT problem. There is 1 Reply. #: 16317 S4/API-User Interface 01-Nov-92 18:00:20 Sb: #16309-Comm Defaults Fm: Kent Stewart 75146,224 To: neil colvin 71650,3517 (X) It was a dos program running under NT. It thought it was running at 38400. I have identical responses with Cim 2.0x and Zcomm, which makes it appear to be a dos/nt interaction. Kent #: 16429 S4/API-User Interface 02-Nov-92 19:45:59 Sb: #16270-Comm Defaults Fm: Jerry Drain [Microsoft] 72350,2056 To: Kent Stewart 75146,224 (X) Hi Kent, Sorry, this is a little outside my normal realm. This really isn't a programming issue, but a configuration issue for the DOS/WOW layer. Let me ask you to re-ask this question in WinNT, Section 4 "DOS/Win3.x Apps" support. I thought the default setting was 1200. If it's been set to 300 under DOS/WOW, then it might be a "Lowest Common Denominator" issue. However, like I said, this is a question that they would be more familiar with. Thanks, -- Jerry [@Microsoft] There is 1 Reply. #: 16682 S4/API-User Interface 04-Nov-92 20:27:09 Sb: #16429-Comm Defaults Fm: Kent Stewart 75146,224 To: Jerry Drain [Microsoft] 72350,2056 (X) I'll move over to that area, Jerry. I couldn't imagine 300 baud for the default but file downloads of 28cps don't lie. Kent #: 16495 S4/API-User Interface 03-Nov-92 14:53:09 Sb: DDEML connect prob Fm: David Armitage 76702,1002 To: all I am working on a DDEML server app that I ported from Win 3.1 and I get an Access Violation from NTSD. The server app runs without problems in 3.1 and if I run it as a 3.1 app in NT I have no problem. But if my 32 bit client app tries to connect to the 32 bit server app I get the error. When stepping through the assembly code in WinDbg the error occurs after I have received my XTYP_CONNECT and *BEFORE* I receive my XTYP_CONNECT_CONFIRM. In other words outside of my callback code. Here is what NTSD says: Access Violation USER32! __ClientDDEMLInitiateServer + 0x151: Is there something that I am not doing that I should be doing for NT? If you have any ideas on why DdeConnect() would give me this, I would love to here them. Thanks Stephen 76701,1002 There is 1 Reply. #: 16688 S4/API-User Interface 04-Nov-92 21:41:06 Sb: #16495-DDEML connect prob Fm: Jerry Drain [Microsoft] 72350,2056 To: David Armitage 76702,1002 (X) Hi David, I went through our knowledge databases, and didn't see anything that seem related to this. My first suggestion is to look at the DDEML samples provided on the SDK. There are several installed under ..\mstools\samples\ddeml\... See if the client and the server applications can give you any suggestions. If you don't get anywhere with that, let me know. We'll talk about stripping your code down into some very small examples which I can look at and test for a possible system bug. Thanks, -- Jerry [Win32 SDK Developer Support] There is 1 Reply. #: 16713 S4/API-User Interface 05-Nov-92 07:52:06 Sb: #16688-DDEML connect prob Fm: David Armitage 76702,1002 To: Jerry Drain [Microsoft] 72350,2056 (X) Jerry, Thanks for getting back to me. I did check the sample apps and found nothing at first and they all ran fine. I'm going to go through them again and see what I can find. Its weird because my apps ran fine on Win 3.1 and also they run fine on Win32s. It is just both 32 bit server and client running on NT. Thanks Stephen #: 16573 S4/API-User Interface 04-Nov-92 07:00:13 Sb: Printf Outputs? Fm: richard c. allen 71621,2420 To: all I am starting to play with DLL's under NT, I have a simple question. My first dll is a simple console dll which used printf()'s to tell me what was happening. (The order coming out was not sync, but okay). I then called the DLL from a generic NT WIndows Application. WHERE did the PRINTF's go? I didn't remove them from the DLL, did the system ignore them? Is there a Stdout/Stdin for Win32? Curious... Richard C. Allen There is 1 Reply. #: 16680 S4/API-User Interface 04-Nov-92 20:12:16 Sb: #16573-Printf Outputs? Fm: Jerry Drain [Microsoft] 72350,2056 To: richard c. allen 71621,2420 (X) Hi Richard, >>I am starting to play with DLL's under NT, I have a simple question. My first dll is a simple console dll which used printf()'s to tell me what was happening. (The order coming out was not sync, but okay). I then called the DLL from a generic NT WIndows Application. WHERE did the PRINTF's go? I didn't remove them from the DLL, did the system ignore them? Is there a Stdout/Stdin for Win32? Curious... I guess I'm a little confused by what you're doing, and what you're seeing. Does your DLL allocate a console??? To answer your question: yes, Win32 apps have a Stdout and a Stdin. By default though, they don't point to anything. You can do things like redirect Stdin/Stdout to a file from the command line, and then call your window app. This should pipe the printfs to the file. You can also allocate a console using AllocConsole from your window app, and get the printfs that way. To learn more, you might start browsing the GetStdHandle() and SetStdHandle() API and go from there (these calls allow you to get and set the handles). Thanks, -- Jerry [Win32 SDK Developer Support] There is 1 Reply. #: 16708 S4/API-User Interface 05-Nov-92 05:51:39 Sb: #16680-Printf Outputs? Fm: richard c. allen 71621,2420 To: Jerry Drain [Microsoft] 72350,2056 (X) Thanks, I will look into it. My simple DLL just used some Printf's in the entry function and I had the SIMPLE app call to the DLL. I saw the outputs from the DLL appear on the console where I started SIMPLE. I then used GENERIC to call my DLL and did not see any outputs from the DLL's Printf's. I did not use any command line nor AllocConsole. Its a new world.... Richard C. Allen There is 1 Reply. #: 16738 S4/API-User Interface 05-Nov-92 12:47:57 Sb: #16708-Printf Outputs? Fm: Jerry Drain [Microsoft] 72350,2056 To: richard c. allen 71621,2420 (X) Hi Richard, It is: a new concept. It took me a little getting use to too. Have a look at the console stuff. If you have any questions regarding the console API or concepts, please post questions in the BASE section (Section 6). Thanks, -- Jerry [Win32 SDK Developer Support] #: 16498 S4/API-User Interface 03-Nov-92 15:06:30 Sb: AllocConsole question Fm: Alan feuer 75500,3610 To: David Armitage 76702,1002 (X) How can a console be created with an initial state? The routines to modify a console require a handle to the console, but the only way to get a handle to a console is to create one. Unfortunately, AllocConsole causes the console to appear. Is there a way to create a console without causing it to appear? alan feuer There is 1 Reply. #: 16681 S4/API-User Interface 04-Nov-92 20:12:23 Sb: #16498-AllocConsole question Fm: Jerry Drain [Microsoft] 72350,2056 To: Alan feuer 75500,3610 (X) Hi Alan, >>How can a console be created with an initial state? The routines to modify a console require a handle to the console, but the only way to get a handle to a console is to create one. Unfortunately, AllocConsole causes the console to appear. Is there a way to create a console without causing it to appear? I don't believe there is a way to do this. I'll have a look in the source code to see how AllocConsole is working, and perhaps I discover some sort of work around. Before I do however, can you please tell me exactly what it is you're trying to do. Perhaps I can provide you with another work around, or it will help me to identify one. Thanks, -- Jerry [Win32 SDK Developer Support] There are 2 Replies. #: 16699 S4/API-User Interface 04-Nov-92 23:54:01 Sb: #16681-AllocConsole question Fm: Marc Singer 72130,2546 To: Jerry Drain [Microsoft] 72350,2056 (X) Have you the ability to create the console 'off screen' so that it can be configured before being moved to an on screen location? There is 1 Reply. #: 16739 S4/API-User Interface 05-Nov-92 12:48:02 Sb: #16699-AllocConsole question Fm: Jerry Drain [Microsoft] 72350,2056 To: Marc Singer 72130,2546 (X) Hi Marc, >>Have you the ability to create the console 'off screen' so that it can be configured before being moved to an on screen location? Not that I know of. I may find out more once I've had a chance to see how the AllocConsole API is emplemented in the source code. Thanks, -- Jerry [Win32 SDK Developer Support] #: 16710 S4/API-User Interface 05-Nov-92 07:03:16 Sb: #16681-AllocConsole question Fm: Alan feuer 75500,3610 To: Jerry Drain [Microsoft] 72350,2056 (X) I'm simply trying to do something similar to ProgMan. I would like to create a console from a GUI app, but have the console begin with certain attributes. Alan Feuer #: 16554 S4/API-User Interface 03-Nov-92 21:38:03 Sb: Win16 on NT Fm: Jerry Drain [Microsoft] 72350,2056 To: David Risack 75300,3712 (X) Hi David, >>I'm trying to figure out how to debug a win-16 app under NT. Currently the App crashes on exit from it and on certain other occasions. How can I currently debug I win-16 app. Even something that would tell me the last API call made would give me a hint. Options are pretty limited, as we haven't had time to work on this yet. I understand we have some things in the works, but their aren't any details yet available. I know that CodeView will not work. I've heard that QCWin will work as a debugger now, but I haven't had a chance to do this yet myself, so I can't answer anything about its robustness yet. If you have it available to you, please try it. >>Is there anyway to figure out what's going on? (without a way to debug under NT it's hard to be more specific other than the app crashes-- we'll probably have to start throwing message boxes or doing file output to try and figure where if there is no current way to do this.) You say this works under Windows 3.1, but not with Windows NT under the WOW layer. Are you touching hardware directly with this app? Is it doing anything tricky with memory? It's very hard to suggest anything without having the problem narrowed down some. Thanks, -- Jerry [Win32 SDK Developer Support] There are 2 Replies. #: 16567 S4/API-User Interface 04-Nov-92 04:31:33 Sb: #16554-Win16 on NT Fm: David Risack 75300,3712 To: Jerry Drain [Microsoft] 72350,2056 (X) Nope we aren't accessing any of the hardware directly. I'm not exactly sure what you mean by doing anything tricky with memory, but I don't think so. I don't believe I have access to QCWin. So currently I have no debugger to debug my win-16 apps under NT. (Which I shouldn't have to do anyway because it should work :) There must be someproblems with the WOW layer. -->Dave There is 1 Reply. #: 16689 S4/API-User Interface 04-Nov-92 21:41:11 Sb: #16567-Win16 on NT Fm: Jerry Drain [Microsoft] 72350,2056 To: David Risack 75300,3712 (X) Hi Dave, >>(Which I shouldn't have to do anyway because it should work :) There must be someproblems with the WOW layer. Well, it's pretty early to make that call yet; although I'm sure that the WOW layer isn't 100% yet. How big is this code, what's it doing? Are you getting anywhere at all with it? Is it something you can narrow down to a page or two and demonstrate the crash? If you can narrow this down to something small and concrete that I can take to the developers, I'd like to have a look at it. Thanks, -- Jerry [Win32 SDK Developer Support] There is 1 Reply. #: 16704 S4/API-User Interface 05-Nov-92 05:12:49 Sb: #16689-Win16 on NT Fm: David Risack 75300,3712 To: Jerry Drain [Microsoft] 72350,2056 (X) Unfortunately we haven't tracked down the crash under NT yet (we don't have a debugger to help us so it's tougher to find.) Since the app runs fine under Windows 3.0 and 3.1 and the debug versions of either I'd have to presume it's a bug in the WOW layer. (Because if it fully emulated it properly it would run right? :) I believe MS has a copy of our application on file that they can test with windows NT. -->Dave There is 1 Reply. #: 16742 S4/API-User Interface 05-Nov-92 13:02:40 Sb: #16704-Win16 on NT Fm: Jerry Drain [Microsoft] 72350,2056 To: David Risack 75300,3712 (X) Hi David, Sorry, can you tell me who here might have a copy of the application? Were you in the porting lab at one time, or have you worked with someone specifically with this application in the past. Thanks, -- Jerry [Win32 SDK Developer Support] #: 16683 S4/API-User Interface 04-Nov-92 20:54:10 Sb: #16554-Win16 on NT Fm: Sheldon Fox 70162,3422 To: Jerry Drain [Microsoft] 72350,2056 (X) Jerry, PMJI, I have tried QCWin under NT and while it comes up OK and loads my project (which works fine in Win 3.1 on the same machine), once I press F5 ("GO") the whole QCWin sessions hangs. F8 (to step into the first line of the program and then stop does the same thing.) Trying to kill it ends up with errors saying WOW can't get it to stop. If you can remember who told you that QCWin worked, maybe you can ask them if they have any secret incantations that are usuful in getting it to work . Sheldon #: 16748 S4/API-User Interface 05-Nov-92 14:15:22 Sb: UnpackDDElParam poor Fm: Bruce Cowan 73650,32 To: all I think UnpackDDElParam is a little bit broken: It will crash with an access violation if called to unpack the lParam received in a WM_DDE_ACK message which is an answer to a WM_DDE_INITIATE. Now, I can obviously write my code to check whether I am currently sending a WM_DDE_INITIATE and not call UnpackDDElParam in this case, and I agree that UnpackDDElParam is only supposed to be called to process a posted DDE message, not a sent one, but, this makes the winproc ugly: ... if (!InSendMessage()) UnpackDDElParam(...) else unpack it the old way ... Why can't this code be put into UnpackDDElParam itself? (Previously I reported what I thought were more problems with UnpackDDElParam, in the wrong forum section, but I've now found that they were all fixed in the October release. I at first thought they weren't, but that was because I had some screwy code in my routine to get around problems in the July release.) Bruce #: 16684 S4/API-User Interface 04-Nov-92 21:33:00 Sb: MMSYSTEM.DLL Fm: John M. Brown 76437,2222 To: all I am trying to use the MMSYSTEM.DLL for Multi-Media stuff under Oct-MIPS. I get and error saying the file is an incompatible file format type... Is this file corrupted?? jmbrown DiagSoft, Inc. There are 2 Replies. #: 16737 S4/API-User Interface 05-Nov-92 12:47:54 Sb: #16684-MMSYSTEM.DLL Fm: Jerry Drain [Microsoft] 72350,2056 To: John M. Brown 76437,2222 (X) Hi John, First think, make sure you're using the correct MMSYSTEM.DLL file (that it's not the one for your Win 3.1 machine or something. The file I have listed on the MIPs machine is dated 10/10/92, and has a file size of 14848 bytes. If it's the correct file, please repost this question in the GDI section (Section 5). We have some expertise on both MM and MIPs in that section right now. Thanks, -- Jerry [Win32 SDK Developer Support] #: 16773 S4/API-User Interface 05-Nov-92 17:06:39 Sb: #16684-MMSYSTEM.DLL Fm: Jerry Drain [Microsoft] 72350,2056 To: John M. Brown 76437,2222 (X) Hi John, In an earlier message to you, I wrote: >> First think, make sure you're using the correct MMSYSTEM.DLL file ..... A colleague of mine, who knows me to generally not to be so abrupt, pointed this out and wondered if I had made a typing error. I'm sorry, this was a typing error, I meant to type: >> First thing, make sure you're using ..... Sorry if I sounded short to you, it was not intended. (Thanks for the catch Paul). Thanks, -- Jerry [Win32 SDK Developer Support] #: 16808 S4/API-User Interface 06-Nov-92 04:39:47 Sb: MFC Comm Dlgs Fm: Trip Tucker 76054,2250 To: Sysop (X) It seems that there are still some problems with the Print Common Dialog and MFC in the new release. While this worked with the previous release, the dialog does not initialize with the correct page numbers from the structure and then does not return with the page numbers that were set. The Font Dialog worked with the last release, but now tells you that there are no fonts installed for the printer. This is the same code that works for WIN16 and worked with the July release. Have you seen this? Trip. #: 16448 S4/API-User Interface 03-Nov-92 06:07:48 Sb: GetNextWindow() Fm: Howard Myers 76711,462 To: Microsoft I have been unable to get GetNextWindow() to work at all. I try both of the following: PrevHwnd = GetNextWindow( hwnd, GW_HWNDPREV ); NextHwnd = GetNextWindow( hwnd, GW_HWNDNEXT ); where hwnd is my main window handle. There are other active windows. While one of these may legitamately return NULL, at least one of them should return a valid handle, shouldn't it? (This code works under 3.x) The only sample program that uses this call is the OLE Client demo. I tried running it under debug to see if it could successfully use this call. However, it will only call GetNextWindow() if it first gets something returned from GetTopWindow(). This call was always returning NULL. That may be correct, since the sample is passing in a handle, thus, it should only get handles for child windows. Additionally, if you call FindWindow() using your own class and NULL for the Window string, are you guaranteed that it will return a window handle other than yourself? I.e., can this function really be used to determine if there are other instances of yourself running, as suggested in some of the porting docs I have? By reading the description of the function, it seems like your own handle might get returned. Since there is no way to tell it to then give you the next occurrence, you'd be stuck, not knowing if there are others or not. (Might this only work for truely previous instances? It goes down the list in order, so if it started before you, you'll get the other handle, if it started after you, you'll get your own?) Thanks! There is 1 Reply. #: 16736 S4/API-User Interface 05-Nov-92 12:47:48 Sb: #16448-GetNextWindow() Fm: Jerry Drain [Microsoft] 72350,2056 To: Howard Myers 76711,462 (X) Hi Howard, I've been working with a sample that uses the GetNextWindow(). I'm not running into the same problems you are; however, it's not acting like I would expect it to. I've sent some mail to development to clear up a few behavior and/or documentation errors. As soon as some of these issues are cleared up, I'll try and provide you with a sample or some information to help. This might take another day or two. Thanks, -- Jerry [Win32 SDK Developer Support] There is 1 Reply. #: 16741 S4/API-User Interface 05-Nov-92 12:57:39 Sb: #16736-GetNextWindow() Fm: Howard Myers 76711,462 To: Jerry Drain [Microsoft] 72350,2056 (X) Thanks! I'll wait for your next reply. Howard Myers There is 1 Reply. #: 16857 S4/API-User Interface 06-Nov-92 16:09:12 Sb: #16741-GetNextWindow() Fm: Jerry Drain [Microsoft] 72350,2056 To: Howard Myers 76711,462 Hi Howard, I've just uploaded a very small sample called getnxt.zip into MSWIN32, Lib4. It's very simple, but it uses GetNextWindow() to list all of the windows on the desktop by their caption bar title, and the window handle value. Please compare this with your code, and let me know what questions arise. Thanks, -- Jerry [Win32 SDK Developer Support] #: 16798 S4/API-User Interface 05-Nov-92 20:46:03 Sb: Status of upload Fm: murray snowden 76620,702 To: API-User Interface I uploaded a file that has problem after program exits - sharb.zip. How do i get status. Did you receive it? File was sent on 10-31-92. I placed it under item 4 - API - User Interface. There is 1 Reply. #: 16858 S4/API-User Interface 06-Nov-92 16:10:28 Sb: #16798-Status of upload Fm: Steve Firebaugh [MS] 75430,412 To: murray snowden 76620,702 Murray, We have not seen sharb.zip. I aksed our sysop, and he did not have any record of it. That is the first time I have heard of this happening in this forum. If you would, please upload it again, and post a message here with information about it. Please note that almost all of our work here is driven by message postings. We do not watch the libraries for work to do. If you want someone to look at an issue, please post a message. Thanks. Steve Firebaugh #: 16334 S5/API-GDI/Graphics 02-Nov-92 10:04:05 Sb: #16265-Screen Saver Fm: Petrus Wong [Microsoft] 70743,3355 To: Mark Gamber 76450,2754 (X) Hello Mark, Thanks for the feedback, I will pass that along to our developers... petrus #: 16268 S5/API-GDI/Graphics 31-Oct-92 06:52:30 Sb: Vid driver for ATI UPro? Fm: Chuck Venter 76424,2626 To: all I'm guessing at the section here, I'm assuming video drivers are part of graphics? Has anyone got (supported or not) a driver for the october release that supports the ATI Ultra Pro video accelerator card? ATI hasn't but i heard that some people at Microsoft were using that card and wondered if they have something we can use till ATI delivers (or hell freezes over, whichever comes first). :) There is 1 Reply. #: 16335 S5/API-GDI/Graphics 02-Nov-92 10:04:10 Sb: #16268-Vid driver for ATI UPro? Fm: Petrus Wong [Microsoft] 70743,3355 To: Chuck Venter 76424,2626 (X) Hello Chuck, >>Has anyone got (supported or not) a driver for the october release that supports the ATI Ultra Pro video accelerator card?<< No, Windows NT is not supporting the card yet... ...petrus #: 16484 S5/API-GDI/Graphics 03-Nov-92 12:02:41 Sb: MIPS blt is sloooow!!! Fm: Petrus Wong [Microsoft] 70743,3355 To: Lanre Amos 70550,2360 (X) Hello Lanre, >>Blt functions on the MIPS (StretchBlt etc) is slooooooooow on the October build. Much slower than on the July PDC! Any ideas what's going on here?<< Can you provide us with some more information so that we can look into this? 1. Is this regarding screen update? With wallpaper? If so, how many bits per pixel? 2. Is this memory to memory blt? thanks, petrus #: 16331 S5/API-GDI/Graphics 02-Nov-92 09:24:58 Sb: #16212-Bug in StretchBlt Fm: Len Gray 71630,1703 To: Steve Firebaugh [MS] 75430,412 (X) Hi, Steve I checked out the STREBLT example and find the same problem that I reported. Try this: - Open STREBLT. Set the Source and Target values to 0, 0, 32, 32. This is the setting for a straight transfer of a 32x32 bitmap. - Select New Source and open up the EGYPT.BMP bitmap that comes with Windows. - You should have identical bitmaps on each side. Note that there is tan (color) along the bottom edge of the bitmap. - Change the target Y to 31, and the target height to -32. These are the values that I have found do the correct 'flip' since Windows 3.0. Hit ENTER to activate the new settings. - Now you see a vertically flipped bitmap in the dest area, but notice that there is no color along the top edge. This is the same result that I get in my application when it runs under Windows NT. Please let me know what you see when you try it on your end. Thanks! Best regards, Len There is 1 Reply. #: 16602 S5/API-GDI/Graphics 04-Nov-92 11:03:28 Sb: #16331-Bug in StretchBlt Fm: Steve Firebaugh [MS] 75430,412 To: Len Gray 71630,1703 (X) Len, Sorry about taking so long to respond. I too have seen the 1 pixel difference between Windows 3.1 and Win32. I've been trying to track down someone who can tell me which version is the correct behavior. No luck yet, but I'll keep trying. As you've observed, this mirroring functionality is not especially well documented, although I do not see any place where it is actually wrong. In any case, you are getting a vertically mirrored image right? For now, can't you use Y=32 (rather than 31) as a work around? I find that this gives me the upside down image exactly where I would want it. Thanks. Steve Firebaugh #: 16497 S5/API-GDI/Graphics 03-Nov-92 14:59:55 Sb: Bug in StretchBlt Fm: Len Gray 71630,1703 To: All I posted a bug in StretchBlt here last week. Steve Firebaugh suggested that there was no bug and to check out the STREBLT sample app. I have checked out STREBLT and it reproduces the error I reported. I noticed that Steve hasn't received messages here since last week - perhaps someone from Microsoft could read my message #16331 for the bug report. (The original report is in message #16197). Thank you! Len There is 1 Reply. #: 16527 S5/API-GDI/Graphics 03-Nov-92 17:34:17 Sb: #16497-Bug in StretchBlt Fm: Steve Firebaugh [MS] 75430,412 To: Len Gray 71630,1703 (X) Len. I am working on the problem which you posted yesterday. I will have an answer for you here by tomorrow morning. Steve Firebaugh #: 16491 S5/API-GDI/Graphics 03-Nov-92 13:56:41 Sb: Video Cards Fm: Absoft Corporation 76360,3551 To: All 92.11.03 Absoft is preparing some Windows NT graphically oriented demonstration programs that will be displayed in one of the Intel hosted areas of COMDEX. Many of our demos, written in FORTRAN and compiled with our 32-bit Windows NT compiler, are best when run on high resolution displays with a bare minimum of 8 bit color. We have been contacting 24bit card manufacturers to see if any of them have drivers for Windows NT. So far we've been drawing a blank. Can Microsoft supply a list of video card vendors supporting drivers for the July build of Windows NT? We would prefer either MCA or EISA cards. If we can't find appropriate drivers/cards we'll probably change our operating system choice to DOS or Unix; we'd prefer to show-off under Windows NT (some of our demo apps make nice use of threads). Jeff Knaggs Absoft 2781 Bond St. Rochester Hills, MI 48309 (313) 853-0050 voice (313) 853-0108 fax (313) 853-0000 bbs 2400,8,1,n There is 1 Reply. #: 16603 S5/API-GDI/Graphics 04-Nov-92 11:03:35 Sb: #16491-Video Cards Fm: Steve Firebaugh [MS] 75430,412 To: Absoft Corporation 76360,3551 (X) Jeff Knaggs, If you have the October beta CD, there is a file named \readme\oct92hcl.txt. This is the "hardware compatibility list." (I think that this is also in a CIS library somewhere. If you need it, let me know, and I can either find it for you, or upload it.) In any case, the supported video cards from that file are listed below. I hope that this helps, Steve Firebaugh [Excerpt from \readme\oct92hcl.txt...] Display Adapters ---------------The following display adapters have been tested. Actix Graphics ENGINE (1024x768, 800x600, 640x480, 256 colors) Diamond Stealth VRAM (1024x768, 800x600, 640x480, 256 colors) Diamond SpeedStar (1024x768, 800x600, 16 colors) IBM VGA (640x480, 16 colors) Headland Technology/Video 7 VRAM VGA (1024x768, 800x600, 16 colors) Headland Technology/Video 7 VRAM II (1024x768, 800x600, 16 colors) Orchid Fahrenheit 1280 (1024x768, 800x600, 640x480, 256 colors) Orchid ProDesigner IIs (1024x768, 800x600, 16 colors) TSENG Labs ET4000 (1024x768, 800x600, 16 colors) IBM XGA(R) (1024x768, 640x480, 256 colors) DELL DGX (1280x1024, 1152x900, 1024x768, 800x600, 640x480, 256 colors) S3 86C911 (1024x768, 800x600, 640x480, 256 colors) Native display adapter on MIPS ARC/ R4000 systems from ACER, MIPS and Olivetti (1280x1024, 1024x768, 256 colors) #: 16734 S5/API-GDI/Graphics [MSWIN32] 05-Nov-92 12:21:23 Sb: Fm: To: >>Thanks for getting back to me, Bob. I was aware that critical sections are an extreme solution there, and I used them so I was sure I understood the sequencing of the threads. That was while I still suspected the problem was related to multi-threaded "clients", to use the X terminology.<< >>But remember we saw a rarer version of the same bug in Perfmonitor (July rel) which is a single threaded app. This make me much more suspicious of a sequencing problem in the "server": i.e Window Manager.<< >>>> Hmm...looks like a base issue. Would you please repost this in the >>>>API-Base/Section. >>>> thanks, >>>> petrus Done, Dave There is 1 Reply. #: 16659 S5/API-GDI/Graphics 04-Nov-92 16:35:31 Sb: Remember Perfmonitor Fm: David Arathorn 70412,2772 To: Bob Landau, 70744,21 (X) Thanks for getting back to me, Bob. I was aware that critical sections are an extreme solution there, and I used them so I was sure I understood the sequencing of the threads. That was while I still suspected the problem was related to multi-threaded "clients", to use the X terminology. But remember we saw a rarer version of the same bug in Perfmonitor (July rel) which is a single threaded app. This make me much more suspicious of a sequencing problem in the "server": i.e Window Manager. Good Luck, Dave There is 1 Reply. #: 16734 S5/API-GDI/Graphics 05-Nov-92 12:21:23 Sb: #16659-Remember Perfmonitor Fm: Petrus Wong [Microsoft] 70743,3355 To: David Arathorn 70412,2772 (X) David, >>Thanks for getting back to me, Bob. I was aware that critical sections are an extreme solution there, and I used them so I was sure I understood the sequencing of the threads. That was while I still suspected the problem was related to multi-threaded "clients", to use the X terminology.<< >>But remember we saw a rarer version of the same bug in Perfmonitor (July rel) which is a single threaded app. This make me much more suspicious of a sequencing problem in the "server": i.e Window Manager.<< Hmm...looks like a base issue. Would you please repost this in the API-Base/Section. thanks, petrus #: 16538 S5/API-GDI/Graphics 03-Nov-92 18:59:32 Sb: Real-time animation?? Fm: Petrus Wong [Microsoft] 70743,3355 To: Joe Peterson[CyberStore] 76150,1714 (X) Hi Joe: Can you be a little bit more specific what kind of support you are looking for? thanks, petrus There is 1 Reply. #: 16557 S5/API-GDI/Graphics 03-Nov-92 22:43:42 Sb: #16538-Real-time animation?? Fm: Joe Peterson[CyberStore] 76150,1714 To: Petrus Wong [Microsoft] 70743,3355 (X) If by support, you mean via NT, then here's what we need the ability to do: 1) Real-time animation -- not pre-recorded frame playing, or bitblt-ing. We need to generate graphical objects on the fly, and animate them. That is, move them around in a three-demensional world created by the software. (This world must be as dynamic as the real world, where objects in the world come and go based on changes in the usage of the software, and what it is supposed to do.) 2) Full-motion video, both full-screen and in a window. This video must be of a quality to parallel that of your average television picture. This video must be played back smoothly and without and jerkiness. 3) Full-stereo sound (or at least to the quality of available sound cards). Just as for video, this sound playback must be smooth and without and break-up. 4) The system must support a high-speed (9600 bps or better) communications session while the "virtual world"/multimedia presentation per above is running. 5) And lastly, the system must be capable of upcoming technologies, such as voice recognition and generation. Of course, we'd like all this to happen is a standard GUI environment as well. Currently, our view on Windows and OS/2 is that the GDI and GPI, respectively, cannot handle the requirements I've outlined. The implication is that we must use "home-brewed" graphics routines in a full-screen environment and by-pass all the functionality offered by the Windows or PM interface. My question to Microsoft is: can NT fill these requirements, both *currently* (1 year or less down the pike), and into the future (two or more years along). Given that we don't expect to complete development much earlier than 12 to 18 months from now, would say that we could count on support for our requirements in NT? jep There is 1 Reply. #: 16767 S5/API-GDI/Graphics 05-Nov-92 16:07:59 Sb: #16557-Real-time animation?? Fm: Petrus Wong [Microsoft] 70743,3355 To: Joe Peterson[CyberStore] 76150,1714 Hello Joe, Thanks for the feedback. I will pass them along to our development. For point 3, full-stereo sound, Windows NT already support that. As for the other requirements, it really depends on what kind of hardware you want to run on. We do have plans to support multimedia areas other than sound, but I don't have any time frame available for you. thanks, petrus #: 16589 S5/API-GDI/Graphics 04-Nov-92 10:29:52 Sb: GetDC returns NULL-why? Fm: Mike Payne 71062,3342 To: Anyone Could someone please give me quick summary of why GetDC would return a value of NULL. My aplication runs along fine, until, after about 30000 calls to GetDC (And also calls to releaseDC ), GetDC just starts returning 0 for any windows (it seems) which were not created before that point. All the windows are of the same class, and have CS_OWNDC set . (Is this the problem?). Thanks Chris Hall (71062,3342) There is 1 Reply. #: 16799 S5/API-GDI/Graphics 05-Nov-92 20:58:58 Sb: #16589-GetDC returns NULL-why? Fm: Petrus Wong [Microsoft] 70743,3355 To: Mike Payne 71062,3342 Hi Chris, >>Could someone please give me quick summary of why GetDC would return a value of NULL. My aplication runs along fine, until, after about 30000 calls to GetDC (And also calls to releaseDC ), GetDC just starts returning 0 for any windows (it seems) which were not created before that point. All the windows are of the same class, and have CS_OWNDC set . (Is this the problem?).<< Is it possible that you are running out of memory or hard disk space for the swapping file? A memory leak somewhere? petrus #: 16483 S5/API-GDI/Graphics 03-Nov-92 12:02:35 Sb: S3 Drivers Fm: Petrus Wong [Microsoft] 70743,3355 To: Howard Myers 76711,462 (X) Hello Howard: >>First, the SetDIBits with a 256 color logical palette is extremely slow. (Approx. 10 times slower than 3.1) (Naturally, this assumes use of the DIB_PAL_INDICES option. Using DIB_PAL_COLORS it is extremely fast, but obviously gives incorrect results when 256 colors are required.)<< This will be fixed in Beta 2. Think you might want to know... ...petrus There is 1 Reply. #: 16492 S5/API-GDI/Graphics 03-Nov-92 13:57:08 Sb: #16483-S3 Drivers Fm: Howard Myers 76711,462 To: Petrus Wong [Microsoft] 70743,3355 (X) Thanks! I'm looking forward to Beta 2 already! There is 1 Reply. #: 16697 S5/API-GDI/Graphics 04-Nov-92 22:56:16 Sb: #16492-S3 Drivers Fm: Petrus Wong [Microsoft] 70743,3355 To: Howard Myers 76711,462 (X) Hello Howard, Wonder if you have found a work around for the color problem already. Here is what I have discovered; you might want to use it as a temporary work around for the Comdex. (I have already raised a bug against it, but might take a while for it to be fixed.) Repeating the rendering process on a OWNDC will give the correct color. (Actually, the second time and thereafter will work.) By repeating the process, I meant selecting the palette, realizing the palette and then blting the bitmap again. If this is not an OWNDC, that wonldn't work. (Let me know if you like to see my test code.) Hope that helps, petrus There are 2 Replies. #: 16706 S5/API-GDI/Graphics 05-Nov-92 05:31:59 Sb: #16697-S3 Drivers Fm: Howard Myers 76711,462 To: Petrus Wong [Microsoft] 70743,3355 (X) I'll give it a try. I have worked around the color problem using SetPixelV(), but this is quite slow. Maybe your technique will help me get some speed back. Thanks! #: 16803 S5/API-GDI/Graphics 05-Nov-92 22:33:56 Sb: #16697-S3 Drivers Fm: Brian Moura 76702,1337 To: Petrus Wong [Microsoft] 70743,3355 (X) We noted that the S3 640x480x256 driver works but the video is displayed incorrectly. Makes the screen image look like 800x600 and not 640x480. Had to switch back to the NT VGA driver for correct video. There is 1 Reply. #: 16844 S5/API-GDI/Graphics 06-Nov-92 14:02:54 Sb: #16803-S3 Drivers Fm: Petrus Wong [Microsoft] 70743,3355 To: Brian Moura 76702,1337 Brian, >>We noted that the S3 640x480x256 driver works but the video is displayed incorrectly. Makes the screen image look like 800x600 and not 640x480. Hmmm...I tried switching back and forth among the 1024x768, 800x600 and 640x480 via regedit, the video looks fine to me... ...petrus #: 16834 S5/API-GDI/Graphics 06-Nov-92 11:40:34 Sb: Video Adapters Fm: Absoft Corporation 76360,3551 To: all 92.11.06 Steve Firebaugh, We appreciate your response, but we will be using the the July release of Windows NT at COMDEX. Therefore, we need a list of display adapters which will work under the July release. We would like a display having 24 bit color, but please give us a complete list of display adapters. Bruce Meier Absoft Corporation There are 2 Replies. #: 16845 S5/API-GDI/Graphics 06-Nov-92 14:02:58 Sb: #16834-Video Adapters Fm: Petrus Wong [Microsoft] 70743,3355 To: Absoft Corporation 76360,3551 Hi Bruce, Sorry for jumping in...but the October Beta don't have display driver support for 24 bit color yet. Our DDK was just released though. thanks, petrus #: 16846 S5/API-GDI/Graphics 06-Nov-92 14:41:11 Sb: #16834-Video Adapters Fm: Steve Firebaugh [MS] 75430,412 To: Absoft Corporation 76360,3551 Bruce Meier, The July release offered fewer choices for video display. There were only two display adapters for X86 platform which offered more than four bit color: the IBM XGA and the DELL DGX, both with eight bit color. All of this is covered in the release notes and hardware compatibility list which you should have received at the same time that you got the July SDK itself. Steve Firebaugh #: 16673 S5/API-GDI/Graphics 04-Nov-92 18:59:06 Sb: Making .FON files Fm: Steve Firebaugh [MS] 75430,412 To: Bruce Cowan 73650,32 (X) Bruce, Last week we were discussing the fact that EnumFonts() returns an application FOUR versions of a bitmap font from a FON file. (plain, italic, bold, italic&bold). I have finally heard back from the people responsible for the EnumFonts() code. They consider it a bug, and will fix the next release so that it only returns one. Thought you'd be interested. Hope your work is going well. Steve Firebaugh There is 1 Reply. #: 16747 S5/API-GDI/Graphics 05-Nov-92 14:14:22 Sb: #16673-Making .FON files Fm: Bruce Cowan 73650,32 To: Steve Firebaugh [MS] 75430,412 (X) Thanks Steve. Yes, otherwise things are going quite well. The inability to draw on DIBs will shortly cause me to have to rewrite quite a bit of code, but ... The only real problem is the time it takes to report all the bugs to you folks ! Bruce There is 1 Reply. #: 16818 S5/API-GDI/Graphics 06-Nov-92 09:37:08 Sb: #16747-Making .FON files Fm: Steve Firebaugh [MS] 75430,412 To: Bruce Cowan 73650,32 (X) Hi Bruce. >The only real problem is the time it takes to report all the bugs >to you folks ! We really would like to be doing the best that we can on receiving such bug reports, but we are subject to certain constraints. I you have time, I'd appreciate it if you could expand your comment. 1. Do you consider the medium of communication the problem here? If so, what would be better? 2. Do you consider the response times you've been getting inadequate? 3. Do you consider the structure of this forum a problem? Your comments appreciated. Steve Firebaugh There is 1 Reply. #: 16835 S5/API-GDI/Graphics 06-Nov-92 11:41:46 Sb: #16818-Making .FON files Fm: Bruce Cowan 73650,32 To: Steve Firebaugh [MS] 75430,412 (X) Steve, the problem is that I have this rather large application which I am porting to NT. All I really want to do it get it working in NT. However, NT is quite buggy (no condemnation of NT intended) at this point, so when I run into a problem I have to 1) make sure it isn't my bug, 2) isolate it into a small piece of code if possible, 3) report it to you folks. It is really the second activity which causes all the time "wasting" - sometimes it is really quite time-consuming to extract bits of things to figure out what is going on. I'll give you an example. Right now I have a problem where the system colors get clobbered and I do not ask to have them clobbered. Actually, only the button shadow color gets clobbered. I've spent some time trying to figure out what is going on, with no luck so far. Everything I've extracted from my program works just fine in isolation! Of course, if you know much about using animated palettes you nkow that this is a complex situation anyway. Hopefully I'm getting somewhere (I just had another idea) and will be able to report something useful on this problem soon. (Needless to say, it all works fine in Win 3.1.) Bruce There is 1 Reply. #: 16847 S5/API-GDI/Graphics 06-Nov-92 14:41:16 Sb: #16835-Making .FON files Fm: Steve Firebaugh [MS] 75430,412 To: Bruce Cowan 73650,32 Bruce, Thanks for the feedback. Fundamentally, this seems like the problem which any project is going to run into developing software on a beta system. The only slightly useful tool we would have to help you would be the "Knowledge Base" which should contain a list of all of the bugs that we are aware of. You can 'GO MSKB' on Compuserve, but I have found that interface rather clumsy. We try to include the most up-to-date Knowledge Base on each CD, but most of the bugs come in right AFTER the CD is released, not before. Steve Firebaugh #: 15692 S6/API-Base/Security 23-Oct-92 12:06:27 Sb: Fm: To: Note that GetCurrentProcess() only returns a psuedo handle which *only* has meaning in the context of the calling process. hpsuedoHandle = GetCurrentProcess(); retCode = DuplicateHandle( hpsuedoHandle, ... >> I have been playing a little bit with 'DuplicateHandle' but I could not resolve my problem this way (I wanted to send a message to another application). Now I am using the function 'PostThreadMessage' which works satisfactorily, for the thread ids are unique in the system. Christian Betrisey. There is 1 Reply. #: 16267 S6/API-Base/Security 30-Oct-92 22:20:40 Sb: Info on 'GetProcAddress' Fm: Eric Sassaman/Microsoft 75430,411 To: Neil Goldstein 70632,3566 (X) << I need info on GetProcAddress. We have created our own DLL and am not getting the address back from a procedure point that we know is in the DLL, but we are not getting an error either. >> We'll need a bit more to go on... if you could post a small sample that reproduces the problem we'll be glad to track the problem down. If it's a bit to big to post (please avoid this at all costs - the smaller the better) you can upload the sample to the forum library as per page 51 of the Oct release notes. Thanks! Eric Sassaman Microsoft NT Support #: 16312 S6/API-Base/Security 01-Nov-92 16:31:17 Sb: GetUserName ??? Fm: Eric Brown 71214,1006 To: ALL (Tech. Support) Every time I try to compile my program which calls "GetUserName", I get a link error - GetUserNameA8@ undefined. GetUserName is defined in winbase.h, but I guess it is not in and of the libraries as far as I can tell. It is in the help files. Also, MAX_USERNAME_LENGHT is undefined. Currently, I am using libc.lib, ntdll.lib and kernel32.dll. Is there anyway to get at this function - or at least to get the user name of the current user in this release of windows NT? Thanks in advance, Eric. There is 1 Reply. #: 16385 S6/API-Base/Security 02-Nov-92 15:22:02 Sb: #16312-GetUserName ??? Fm: Bruce Ramsey/Microsoft 70324,2742 To: Eric Brown 71214,1006 Hi Eric - >> Every time I try to compile my program which calls "GetUserName", I get a link error - GetUserNameA8@ undefined. GetUserName is defined in winbase.h, but I guess it is not in and of the libraries as far as I can tell. It is in the help files. Also, MAX_USERNAME_LENGHT is undefined. Currently, I am using libc.lib, ntdll.lib and kernel32.dll. Is there anyway to get at this function - or at least to get the user name of the current user in this release of windows NT? Add in advapi32.lib to what you link in and it should work OK. Thanks for the tip on the undefined define - I'll let the appropriate people know if they hadn't already noticed, and either the define will get added or the doc will get fixed Bruce #: 16409 S6/API-Base/Security 02-Nov-92 16:44:25 Sb: #16215-Signal Anomaly Fm: David Taniguchi [MS] 72350,2054 To: Andrew Potter 71075,614 (X) Hi Andrew, Thanks for providing the sample. We are looking into the problem. We have reproduced what you describe. It appears to work sometimes (the second SPC> doesn't display periodically) We will post here when we get more information. Thanks, Dave #: 16280 S6/API-Base/Security 31-Oct-92 12:05:53 Sb: FormatMessage bug? Fm: Rex Conn (4DOS) 71333,202 To: Sysop (X) There seems to be a bug in the October release version of FormatMessage. The lower 8 bits of the first argument are supposed to specify the maximum line length; instead FormatMessage with a FORMAT_MESSAGE_FROM_SYSTEM argument truncates the first line to < 10 characters, inserts a CR/LF, and then fails to add a trailing CR/LF on the second line. I've kludged around it for the moment by using a 0 in the lower byte & hoping the system error messages won't be longer than 80 characters! Also, I was under the impression that national language support was supposed to be added in the October release, but the language ID field in FormatMessage still isn't supported - is it supposed to be or was that put off until the next beta? - Rex There is 1 Reply. #: 16432 S6/API-Base/Security 02-Nov-92 20:12:20 Sb: #16280-FormatMessage bug? Fm: Eric Sassaman/Microsoft 75430,411 To: Rex Conn (4DOS) 71333,202 (X) << There seems to be a bug in the October release version of FormatMessage. The lower 8 bits of the first argument are supposed to specify the maximum line length; instead FormatMessage with a FORMAT_MESSAGE_FROM_SYSTEM argument truncates the first line to < 10 characters, inserts a CR/LF, and then fails to add a trailing CR/LF on the second line. I've kludged around it for the moment by using a 0 in the lower byte & hoping the system error messages won't be longer than 80 characters! >> Thanks for catching this one! I've reported this bug to development. I'll let you know when they get back to me on this. << Also, I was under the impression that national language support was supposed to be added in the October release, but the language ID field in FormatMessage still isn't supported - is it supposed to be or was that put off until the next beta? >> Not sure about the NLS issue, but this particular problem against FormatMessage is still open - this bug hasn't been fixed. I don't have any kind of timetable on when this will be fixed, hopefully by the next release. For now you'll have to keep using 0 for the usPrimaryLangage parameter to the MAKELANGID macro. Eric Sassaman Microsoft NT Support #: 16438 S6/API-Base/Security 02-Nov-92 23:29:02 Sb: Communications timeouts Fm: Eric Sassaman/Microsoft 75430,411 To: Bruce Cowan 73650,32 (X) << That doesn't work (at least in the current release) because the shortest timeout you can specify is 1ms. and the internal interpretation of that seems to be that it will wait until the next timer tick, which, on an Intel 486 machine seems to be about 10ms. I know this because I have tried it and this is the behavior it exhibits. I really would like you to seriously consider the extension. It should be easy to implement and it does provide significant additional functionality. I suppose you could fix the interval timeout to actually work at 1 ms which would work for me, but I can't see an efficient way to do that. >> Yes, the shortest timeout interval will be rounded up to the clock interval. I don't think there's a way to get around this problem, I believe (and I'm no hardware guru) it is largely a limitation of the speed of the CPU and the length of a clock tick. << Further, if I use the short interval, then I can't read all the data that has already arrived with a single call, since any read that specifies a count greater than the available data will wait until a new character arrives and then the timeout interval. It could possibly be a *LONG* time until the next character arrives. I suppose I could use a call to find out how much data has arrived and then read exactly that much, but that is getting complex and it will suffer from the problem that because there are two separate calls, some data could possibly arrive between the calls and it won't be read. >> The interval timeout begins when the driver receives a character reguardless of whether or not you are trying to read from the port at that moment - in other words, if a char arrives, timing begins, then the timeout expires, then you try a read, you will return right away with the available data since the timeout has already expired. Give this a go and let me know how it works out - it worked this way for me. << BTW, lots of communications port stuff is broken in the July release, but I am not bothering to report it because I am expecting the next release any moment [More] There is 1 Reply. #: 16439 S6/API-Base/Security 02-Nov-92 23:29:08 Sb: #16438-Communications timeouts Fm: Eric Sassaman/Microsoft 75430,411 To: Eric Sassaman/Microsoft 75430,411 (X) [Continued] now and thought I'd test things there before reporting problems. >> By now you should have the Oct. release - if the specific problems aren't fixed, please let me know. Best if you could forward small sample(s) to demo the problems - this would be the fastest way to get these to the attention of development and fixed in time for the next release. Thanks! Eric Sassaman Microsoft NT Support #: 16446 S6/API-Base/Security 03-Nov-92 05:46:38 Sb: Pointers Fm: Dan Sullivan 76327,1534 To: SYSOP (X) I have a general question about pointers. All the Win32 functions declare pointers as long pointers (LPSTR for example). However all NT programming is really small model so it seems that pointers should be just P and not LP. Was this done to make porting of existing Win16 apps more straight forward? Is there a concious style decision to retain LP as the standard for the future? I would like to know so I can use similar naming of pointers in my NT app's. Thanks Dan #: 16463 S6/API-Base/Security 03-Nov-92 09:39:52 Sb: Validating path names Fm: David Manns 100112,2773 To: ALL Is there an API call to validate a path and/or file name? If we have to validate our own, do we get any clues about the allowable name characters (esp w.r.t FAT vs NTFS names)? #: 16321 S6/API-Base/Security 02-Nov-92 06:09:12 Sb: #16211-Portable File Commit Fm: Robin 100025,604 To: David Taniguchi [MS] 72350,2054 (X) Thanks Dave, I'll see and get back to you. Robin. #: 16445 S6/API-Base/Security 03-Nov-92 05:36:05 Sb: #16211-Portable File Commit Fm: John Hardman 100023,2532 To: David Taniguchi [MS] 72350,2054 (X) Dave, I thought fflush only flushed out the stream buffers (fopen type stuff) as opposed to direct file i/o (open, sopen etc) stuff. Even with the stream buffers I thought that fflush would flush the run-time library stream buffers only rather than flushing any caching device such as Smartdrv ie. the stream buffers would be flushed to smartdrv but smartdrv itself would not be flushed to disk. Is this incorrect ? I haven't looked at FlushFileBuffers yet - will do that this evening. Obviously using FlushFileBuffers will mean using a #ifdef so that FlushFileBuffers gets used for NT and the DOS3CALL (function 0Dh) used for Win 3.1 - a compromise (assuming it works) but that's not a problem. Many thanks, John Hardman (UK). There is 1 Reply. #: 16488 S6/API-Base/Security 03-Nov-92 12:37:49 Sb: #16445-Portable File Commit Fm: Pete Grey [Microsoft] 70744,22 To: John Hardman 100023,2532 >> I thought fflush only flushed out the stream buffers (fopen type stuff) as opposed to direct file i/o (open, sopen etc) stuff. Even with the stream buffers I thought that fflush would flush the run-time library stream buffers only rather than flushing any caching device such as Smartdrv ie. the stream buffers would be flushed to smartdrv but smartdrv itself would not be flushed to disk. Is this incorrect ? << This is correct, it is necc. to use c-runtime file I/O to use fflush. >> I haven't looked at FlushFileBuffers yet - will do that this evening. Obviously using FlushFileBuffers will mean using a #ifdef so that FlushFileBuffers gets used for NT and the DOS3CALL (function 0Dh) used for Win 3.1 - a compromise (assuming it works) but that's not a problem. << FlushFileBuffers() will do this, as well as specifying FILE_FLAG_WRITE_THROUGH when doing a CreateFile(). -pete #: 16326 S6/API-Base/Security 02-Nov-92 07:49:34 Sb: DLL global var. Fm: Christian Betrisey 76600,1450 To: John Stoddard John, <> Eventually everything is working fine. I just have to initialize all the global variables. It is enough to initialize only the 1st element of an array or structure. I do not need to export any variables, for I access them through procedures. Thank you for your help. Christian There is 1 Reply. #: 16516 S6/API-Base/Security 03-Nov-92 16:47:04 Sb: #16326-DLL global var. Fm: John Stoddard 72400,2551 To: Christian Betrisey 76600,1450 (X) Good! I'm glad you got it working. Eric has been very helpful with these issues - it's definitely rather confusing... John #: 16458 S6/API-Base/Security 03-Nov-92 09:39:33 Sb: Unicode printf Fm: David Manns 100112,2773 To: ALL Is there an equivalent to "printf" and/or "fprintf" for printing a Unicode text string with a Unicode format string (including %ts format)? There is 1 Reply. #: 16534 S6/API-Base/Security 03-Nov-92 18:56:07 Sb: #16458-Unicode printf Fm: Bruce Ramsey/Microsoft 70324,2742 To: David Manns 100112,2773 Hi David - >> Is there an equivalent to "printf" and/or "fprintf" for printing Unicode text string with a Unicode format string (including %ts format)? Coud you please repost this in MSWIN32/section 14, which is the Unicode section? I'm sorry, I don't know the answer to this one. The people who are most focussed on Unicode issues are in section 14. Thanks! Bruce #: 16574 S6/API-Base/Security 04-Nov-92 07:06:57 Sb: XACTION file Atrb Fm: ken condal 72617,572 To: sysop (X) I noticed a new file attribute for transaction processing and am very interested in its uses. The help simply says that you must use 'transaction semantics'. Where can I get info/examples and is this supported for all file systems or just NTFS? Thanks, Ken #: 16579 S6/API-Base/Security 04-Nov-92 09:10:40 Sb: GetFileTitle bug Fm: Len Gray 71630,1703 To: All I reported a bug with GetFileTitle in the July release whereby the filename returned had the first character truncated. Thus the following result would occur: -> Feed GetFileTitle the string "C:\WINDOWS\FILE.EXT" -> It returned "ILE.EXT" in the filename buffer. I had to have a working function so I wrote my own function GetFileName to do the proper processing (DOS/FAT filenames only). Since loading the October release I uncommented out the old GetFileTitle call to see if this had been corrected. However now it appears that GetFileTitle is returning an error code of 0 (successful) without doing a darn thing in the return filename buffer. The buffer has the exact same thing in it after the call as it had before. Please let me know if I can help further with this. Len #: 16601 S6/API-Base/Security 04-Nov-92 10:59:40 Sb: UnpackDDElParam problems Fm: Pete Grey [Microsoft] 70744,22 To: Bruce Cowan 73650,32 (X) Bruce, Would you please re-post this in the API-USER section, as DDE is considered to be part of the User API. Thanks, -pete #: 16514 S6/API-Base/Security 03-Nov-92 16:44:59 Sb: DLL: shared data Fm: John Stoddard 72400,2551 To: Eric Sassaman/Microsoft 75430,411 (X) Thanks, I'm mostly just curious, it's no big deal. JS #: 16560 S6/API-Base/Security 04-Nov-92 00:04:17 Sb: DLL: shared data Fm: Eric Sassaman/Microsoft 75430,411 To: John Stoddard 72400,2551 (X) << Yup, that's just what I was doing - accessing the variable from an .EXE, and resetting its value there... This whole area was very confusing, I just wrote a bunch of test programs until I found out what worked and what didn't in my environment. The key item I wouldn't have guessed on my own was that it was a *pointer* to the item that was exported - any idea why this was done? It sure has been the source of most of my problems in porting from OS/2! >> John, I've gotten word back on why this was done. The reason is that if the loader has to do fixups all over the code pages to set these data references to the correct addresses in the DLL, it dirties the code pages, which means that the pager can't just toss them out and restore from the .exe image, but must page the dirtied code pages out to the paging file. If, instead, only a small indirection table (the Import Address Table) is used, the system won't need to page out gobs of modified code pages. Applying these fixups takes time, too, so there was a performance decision there. Eric Sassaman Microsoft Developer Support There is 1 Reply. #: 16609 S6/API-Base/Security 04-Nov-92 11:19:34 Sb: #16560-DLL: shared data Fm: John Stoddard 72400,2551 To: Eric Sassaman/Microsoft 75430,411 (X) .. well, that makes sense. I was hoping it wasn't just arbitrary ;-) JS #: 16635 S6/API-Base/Security 04-Nov-92 14:10:18 Sb: RegQueryKeyValue Fm: Eric Sassaman/Microsoft 75430,411 To: Marc Singer 72130,2546 (X) << The function is RegQueryValue, yes. The parameter that is a problem is lpcb. The type of pointer is different for NT than for Win 3.1. One is signed and one is unsigned. >> Marc, I'm looking into this to see if this can be changed to avoid this warning in the future. I'll get back to you as soon as I hear anything more on this. Thanks for reporting this problem to us. Eric Sassaman Microsoft Developer Support #: 16639 S6/API-Base/Security 04-Nov-92 14:53:47 Sb: Atom table Fm: Christian Betrisey 76600,1450 To: sysop (X) What is the max. length of an null-terminated string I can pass to the function GlobalAddAtom? Thank you Christian. #: 16308 S6/API-Base/Security 01-Nov-92 14:45:00 Sb: DLL addresses Fm: neil colvin 71650,3517 To: John Hall [MS SDE] 70750,2341 (X) If my DLL builds large, interlinked, data structures (using C++ new), do I have t write an entire new set of memory allocation routines to make all of the allocations occur in the shared memory?? The implication is that I cannot use any of the normal WIN32 calls (GlobalAlloc, LocalAlloc, VirtualAlloc, or new) with this approach. What I really want is to share ALL of my data space. I have no need for process local storage for this DLL. The WIN3.1 model is perfect for the application. Is there an easy solution?? There are 2 Replies. #: 16347 S6/API-Base/Security 02-Nov-92 11:00:39 Sb: #16308-DLL addresses Fm: John Hall [MS SDE] 70750,2341 To: neil colvin 71650,3517 (X) Use one process with lots of threads. That way you can share all the data you want to. There is 1 Reply. #: 16427 S6/API-Base/Security 02-Nov-92 19:32:49 Sb: #16347-DLL addresses Fm: neil colvin 71650,3517 To: John Hall [MS SDE] 70750,2341 (X) I am writing a general purpose DLL which will be used by as many as 5 or 6 different applications on the same system at the same time. It must present a coherent data structure to them all, and all changes in it runtime data must be reflected in all of the processes using it. Sine 32bit DLLs running under WIN32s have TOTAL shared data, and the WIN32s documentation says that this behavior can be duplicated with the SAME CODE under NT, there must be someway of doing it!!! #: 16647 S6/API-Base/Security 04-Nov-92 15:57:49 Sb: #16308-DLL addresses Fm: Eric Sassaman/Microsoft 75430,411 To: neil colvin 71650,3517 (X) << If my DLL builds large, interlinked, data structures (using C++ new), do I have t write an entire new set of memory allocation routines to make all of the allocations occur in the shared memory?? The implication is that I cannot use any of the normal WIN32 calls (GlobalAlloc, LocalAlloc, VirtualAlloc, or new) with this approach. What I really want is to share ALL of my data space. I have no need for process local storage for this DLL. The WIN3.1 model is perfect for the application. Is there an easy solution?? I am writing a general purpose DLL which will be used by as many as 5 or 6 different applications on the same system at the same time. It must present a coherent data structure to them all, and all changes in it runtime data must be reflected in all of the processes using it. Sine 32bit DLLs running under WIN32s have TOTAL shared data, and the WIN32s documentation says that this behavior can be duplicated with the SAME CODE under NT, there must be someway of doing it!!! >> Yes, there is, but there is no easy solution. You will need to use file mapping to create and use shared memory under NT. Take a look at the overviews for file mapping on how to do this. If you are going to be using pointers in your shared memory, for something like a linked list, or structures with pointers in them, you've got to use MapViewOfFileEx, which allows you to specify a base address for the view of the file you are mapping. This will help ensure that your pointers in your structures are valid pointers in all processes that use this shared memory. However, it can't guarantee this address for all processes - if that virtual address space is already begin used in that process memory space, you've got problems. It would be best to write a simple memory manager that allocates a large chunk of shared memory then doles out smaller pieces to your application as you need them, something similar to the run-time malloc() call. I also strongly suggest that you take a look at John Hall's shrmem.zip sample in library 10. This might fit your needs with a few tweaks here and [More] There is 1 Reply. #: 16648 S6/API-Base/Security 04-Nov-92 15:57:54 Sb: #16647-DLL addresses Fm: Eric Sassaman/Microsoft 75430,411 To: Eric Sassaman/Microsoft 75430,411 (X) [Continued] there. Win32s is a bit out of my area of expertise - which win32s doc were you referring to? Eric Sassaman Microsoft Developer Support #: 16556 S6/API-Base/Security 03-Nov-92 22:35:19 Sb: #16220-DLL addresses Fm: Eric Sassaman/Microsoft 75430,411 To: David Edge 75170,1461 (X) << Maybe I am missing something here, but please be patient with me for a moment. Is it not true that the DLL's data space resides within the 2GB virtual address space of the caller? I understand that these are virtual addresses. Here is a scenario: process 1 loads a DLL and the DLL gets its data space mapped at virtual addresses 10000 - 11000. Now, process 2 starts up and uses virtual addresses 10000 - 20000, then loads the same DLL. How will the DLL access its data using those same virtual addresses when they are taken by the caller? Does the DLL have its address space outside of the address space of the caller? Perhaps this seems irrelevant but I am wanting to make extensive use of DLLs from multiple applications and I want to understand what is going on. Thanks. >> I failed to mention this case. If the DLL can't load at it's preferred virtual address, the loader has to make a copy of the DLL's code pages, put them somewhere else, and fixup all references to global data in those code sections to point to the proper shared pages. So you'll essentially wind up with separate copies of the DLL's code, each with different virtual addresses for each shared data item, but both pointing to the same area of physical memory. Eric Sassaman Microsoft Developer Support #: 16288 S6/API-Base/Security 31-Oct-92 22:16:15 Sb: NTFS questions Fm: KENNETH R SCHROCK 70621,1521 To: Bruce Ramsey/Microsoft 70324,2742 (X) I am getting odd behavior in the editor I built for NT. If I load a file into the editor, and then go to file manager and try to move or delete the file, it says "File in use by another process". This is fine, and as it was on the July version. The problem is, I get the same message for every file my editor has touched since I started it up. It doesn't matter if the file was only opened/closed, or if it was opened/modified/saved. It even holds for .exe, object and other files that were created by the compiler and linker shelled from my program. Everything is fine after I close my program. I don't know if this is something different in this build, or because I am using NTFS this time. Why does the system think these files are in use? I compared this to notepad, and after loading a file into note pad, I can go to file manager and move the file. This doesn't seem optimal either, as when you then try to save the file from notepad, you get a "File doesn't exist" error. Second problem. I modified the MANUAL.BAT and didn't load the samples in when I installed this time. NT + SDK + MYWORK left me with 20 meg free on the partition. Tonight I loaded all the samples, built each, and examined each for any differences. Not finding much new, I deleted each after I was done. All the samples are off the drive. I now only have 18 meg free. Where'd the other 2 meg go? I added up the files in the directories and get 55 meg. On an 80 meg partition that should leave me more than 18 meg. What happened? There is 1 Reply. #: 16535 S6/API-Base/Security 03-Nov-92 18:56:19 Sb: #16288-NTFS questions Fm: Bruce Ramsey/Microsoft 70324,2742 To: KENNETH R SCHROCK 70621,1521 (X) Hi Kenneth - >> ...It doesn't matter if the file was only opened/closed, or if it was opened/modified/saved. It even holds for .exe, object and other files that were created by the compiler and linker shelled from my program. Everything is fine after I close my program. I don't know if this is something different in this build, or because I am using NTFS this time Since you say "this time" I assume you're seeing this on the October build >> Why does the system think these files are in use? If there are any handles open on the files this would be normal behavior. It's possible a problem in the system makes it think there are open handles to these files, even though there are none (although this is not a known problem). Do you see this behavior equally on both NTFS and FAT partitions? >> I compared this to notepad, and after loading a file into note pad, I can go to file manager and move the file. This doesn't seem optimal either, as when you then try to save the file from notepad, you get a "File doesn't exist" error Notepad closes all it's handles to the file being edited after it uses the handle to load the file into workspace. So, from the OS's point of view, Notepad retains no ownership that would keep the user from moving the file with file manager (or doing other operations to the file). Notepad gets new handles to the file when the user does a save/new/exit Again, are you seeing this behavior on both NTFS and FAT partitions, or are you seeing different behavior on NTFS versus FAT? >> Second problem. I modified the MANUAL.BAT and didn't load the samples in when I installed this time. NT + SDK + MYWORK left me with 20 meg free on the partition. Tonight I loaded all the samples, built each, and examined each for any differences. Not finding much new, I deleted each after I was done. All the samples are off the drive. I now only have 18 meg free. Where'd the other 2 meg go? I added up the files in the directories and get 55 meg. On an 80 meg partition that should [More] There is 1 Reply. #: 16536 S6/API-Base/Security 03-Nov-92 18:56:25 Sb: #16535-NTFS questions Fm: Bruce Ramsey/Microsoft 70324,2742 To: Bruce Ramsey/Microsoft 70324,2742 (X) [Continued] leave me more than 18 meg. What happened? This sounds like a problem the OS file system folks will want to hear about over in WINNT/9/File-systems. If you could please post the details on this over there, we'd apprciate it - thanks! I'm assumming this happened for you on an NTFS partition, but please clarify this for the folks in WINNT/9. Also, if you haven't already noticed, "DIR C:\ /S >c.lst" is a handy way to get the system to add up all the files for you on the C partition Although adding a lot of files to a partition, then deleting them and not recovering all the space, does sound like a file system issue, one other thing to check (if you haven't already) is temp files the compiler or linker may have failed to delete. If you find temp files left around that shouldn't have been, please let the tools folks know in MSWIN32/7-Tools - thanks! Bruce There is 1 Reply. #: 16657 S6/API-Base/Security 04-Nov-92 16:14:23 Sb: #16536-NTFS questions Fm: KENNETH R SCHROCK 70621,1521 To: Bruce Ramsey/Microsoft 70324,2742 (X) This was not a problem with the july build, and is not a problem under fat. It only started happening this build, and only after I reinstalled on a NTFS partition. My editor is a hack, open\close is Win16 functions, save is Win32 that creates a new file. That is what makes this so strange. This is occuring with files that only were open\closed, those that were saved, and those created by the compiler and linker. I did every check I could, and can't find the 2 meg. I think it has to be directory space for all the samples that didn't get washed when the file left. #: 16662 S6/API-Base/Security 04-Nov-92 17:55:22 Sb: Threads context switch Fm: Koby 71172,2722 To: sysop (X) Hi. I ran a benchmark testing the threads context switch speed. I got a very surprisingly results. The typical thread context switch on a 486 33Mz machine that has not background tasks except the system, id performed in 1.3 Milliseconds. The benchmark process was based on two threads and two event objects. The thread synchronized themselves using the event objects to perform round robin. We are trying to build a real-time application for control systems. A slow context switch will prevent us from building fast response control system as required by real-time control system. I wonder how we can control the context switching time. What is the expected switch time, is it an order of miliseconds or microsecends. By the way the same benchmark on SUN sparc system perform hread switch in 180 micro seconds. When the process is running the CPU thermometer shows 60% application 40% Kernel. This shows me that even if we will the process priority to real-time priority will not change to context switch time, Am I right?. Thanks Koby. #: 16327 S6/API-Base/Security 02-Nov-92 07:51:31 Sb: Message to other proc. Fm: Christian Betrisey 76600,1450 To: bob landau #: 16663 S6/API-Base/Security 04-Nov-92 18:32:44 Sb: #16327-Message to other proc. Fm: Pete Grey [Microsoft] 70744,22 To: Christian Betrisey 76600,1450 (X) Christian, No text was uploaded with this post. Could you please re-post. Thanks, -pete #: 16517 S6/API-Base/Security 03-Nov-92 16:54:55 Sb: _CRT_INIT not in CRTDLL Fm: John Stoddard 72400,2551 To: sysop (X) When I got the October SDK, I read on page 31 of the Win32 SDK Release Notes that "If you need to call C-Runtime functions in a DLL, then the entry point for the .DLL must be _CRT_INIT. ... If you have your own entry point, then you must call the _CRT_INIT function with the same parameters that were passed to your entry point". Each of my .DLLs has a LibMain, which is the entry point - this was a requirement in the July SDK. I tried calling _CRT_INIT from LibMain, with the same parameters. I get an unresolved external. Same thing if I code "CRT_INIT" instead of "_CRT_INIT". I did a hex dump of all of your libraries, and I see a _CRT_INIT entry point only in LIBC and LIBCMT, not CRTDLL, which is what I'm linking to. 1) Is LibMain still needed? 2) Should I not try to call _CRT_INIT if I link to CRTDLL? 3) Is it better if I link my .DLLs to LIBCMT? Thanks, John Stoddard Micro Decisionware There are 2 Replies. #: 16559 S6/API-Base/Security 04-Nov-92 00:04:10 Sb: #16517-_CRT_INIT not in CRTDLL Fm: Eric Sassaman/Microsoft 75430,411 To: John Stoddard 72400,2551 (X) << 1) Is LibMain still needed? 2) Should I not try to call _CRT_INIT if I link to CRTDLL? 3) Is it better if I link my .DLLs to LIBCMT? >> John, the release notes failed to mention that calling _CRT_INIT is unnecessary if you are using crtdll.lib. This is because the run-time gets initialized on the process/thread attach in crtdll.dll's entry point - it's all done for you automagically when using crtdll.dll. There is no need to explicitly call _CRT_INIT in your DLL's code. Eric Sassaman Microsoft Developer Support There is 1 Reply. #: 16608 S6/API-Base/Security 04-Nov-92 11:18:29 Sb: #16559-_CRT_INIT not in CRTDLL Fm: John Stoddard 72400,2551 To: Eric Sassaman/Microsoft 75430,411 (X) That's what I figured, but I wanted to get some comfirmation. Thanks! BTW, is there any particular reason to use or not use CRTDLL? Do I gain anything? lose anything? John There are 2 Replies. #: 16666 S6/API-Base/Security 04-Nov-92 18:35:33 Sb: #16608-_CRT_INIT not in CRTDLL Fm: Eric Sassaman/Microsoft 75430,411 To: John Stoddard 72400,2551 << That's what I figured, but I wanted to get some comfirmation. Thanks! BTW, is there any particular reason to use or not use CRTDLL? Do I gain anything? lose anything? >> The nice thing abouit crtdll.dll is that you can have one copy of the C run-time for any .exe and associated .dll's that link to it. Cuts down on code size. It also enables you to share run-time resources between an .exe and a .dll, for instance, open a stream handle in the .exe an pass it to the dll for processing. If you link both the .exe and .dll with libcxxx.lib, they both have separate copies of the run-time, and you can't mix run-time resources between them like that. Another gotcha is allocating memory in one module and passing to another to be freed, like you might do with some type of uer-defined IPC mechanism. Allocate a packet, stuff it, send it to a DLL, the DLL frees it. Unfortunately, if you're using separate copies of the run-time data, a nulclear explosion will result. If, however, both modules are using crtdll.dll, only one copy of the run-time resources is involved, so this type of resource sharing will work. Eric Sassaman Microsoft Developer Support #: 16693 S6/API-Base/Security 04-Nov-92 21:57:32 Sb: #16608-_CRT_INIT not in CRTDLL Fm: John Hall [MS SDE] 70750,2341 To: John Stoddard 72400,2551 >> Do I lose anything? ctype.h function macros like isalpha() are not currently in crtdll #: 16584 S6/API-Base/Security 04-Nov-92 09:36:41 Sb: #16517-_CRT_INIT not in CRTDLL Fm: dan white 70324,3147 To: John Stoddard 72400,2551 (X) John, I've been wondering about the same thing. I notice that the TLS sample program in the SDK does not call _CRT_INIT. --Dan There is 1 Reply. #: 16611 S6/API-Base/Security 04-Nov-92 11:24:11 Sb: #16584-_CRT_INIT not in CRTDLL Fm: John Stoddard 72400,2551 To: dan white 70324,3147 (X) The samples are a real life-saver. I'm always a lot more confident when I can see working code illustrating something, rather than reading about it in some manual. Good work including these, Microsoft. JS There is 1 Reply. #: 16667 S6/API-Base/Security 04-Nov-92 18:35:37 Sb: #16611-_CRT_INIT not in CRTDLL Fm: Eric Sassaman/Microsoft 75430,411 To: John Stoddard 72400,2551 << The samples are a real life-saver. I'm always a lot more confident when I can see working code illustrating something, rather than reading about it in some manual. Good work including these, Microsoft. >> Thanks for your compliments. We're glad to hear that the samples we wrote were helpful. I'll pass your compliments on to the rest of the team. Thanks! Eric Sassaman Microsoft Developer Support #: 16575 S6/API-Base/Security 04-Nov-92 07:09:06 Sb: Shared Memory Question Fm: Andrew Potter 71075,614 To: all Here's a shared memory question. I've found some documentation regarding shared memory under NT and was successful in getting 2 processes to initialize and use the same shared memory. I used CreateFileMapping, MapViewOfFile, CloseHandle and found documentation for OpenFileMapping, CreateFile, ReadFile, and WriteFile. Is this the entire list? There doesn't seem to be a key associated with the file mapping object, only a name. Is this true? Here's my problem: I have several executables in my application all of which communicate via shared memory. I should clarify this. The user sees this product as 1 application. The user doesn't realize that it's really several executables communicating among one another. If I start up the 'application', the controlling executable creates the shared memory segment and all other executables attach to the segment. If I wanted to start up the 'application' a second time in the NT world, the controlling executable would connect to the shared memory segment from the first invocation of the application. Is there a way for my controlling executable to get a name at runtime via some system call that would be different for each invocation of the application? I can probably come up with some kludgy way to do this but it would be nice if there were a nice system call I could make???? Thanks, Andrew There is 1 Reply. #: 16694 S6/API-Base/Security 04-Nov-92 22:02:59 Sb: #16575-Shared Memory Question Fm: John Hall [MS SDE] 70750,2341 To: Andrew Potter 71075,614 (X) Suggestions: 1) look up shrmem.zip. You won't want to do things that way but it should be helpful. It is in the library for 'Porting from OS/2' 2) consider making this one exe with multiple threads. then you wouldn't have to mess with shared memory at all. 3) In Shrmem you will notice I hit the registry on Process Attach. You could probably place legacy information in the registry and get some of what you want, just an idea. #: 16622 S6/API-Base/Security 04-Nov-92 13:20:21 Sb: MapViewOfFileEx Fm: Kent Cedola - Kitware 72230,1451 To: SYSOP (X) Has anyone (including the fine folks at Microsoft) got MapViewOfFileEx to work between two (or more) processes with shared memory at the same address? If so, what address did you use? I keep getting an error 8 on the second MapViewOfFileEx() API. Thanks, Kent There is 1 Reply. #: 16695 S6/API-Base/Security 04-Nov-92 22:06:19 Sb: #16622-MapViewOfFileEx Fm: John Hall [MS SDE] 70750,2341 To: Kent Cedola - Kitware 72230,1451 (X) Porting from OS/2 Library: shrmem.zip By default I use 16Mb, though it is registry configurable. You can create a 'virtual memory snooper' by looping through all your virtual memory blocks. One of the Virtual*() functions will give you details. It is how I picked my default, along with other notes in the readme's in shrmem.zip. #: 16447 S6/API-Base/Security 03-Nov-92 05:47:47 Sb: IO Control Fm: Dan Sullivan 76327,1534 To: SYSOP (X) I would like to make use of the IO Control funtions to manipulate floppies in my applications. Will this capability be added to NT for the final release? When will it be available to developers? Will complete IO Control functionality be provided? Will existing DOS app's that make use of IO Control for floppies be supported? This lack of IO Control is the only fly in the ointment for the large NT app we are in the process of developing. We must be able to write floppies in non-DOS formats. We are faced with having our users reboot DOS or have a separate DOS system to write these floppies. Needless to say neither is really a desirable solution. Thanks in advance Dan There is 1 Reply. #: 16665 S6/API-Base/Security 04-Nov-92 18:32:58 Sb: #16447-IO Control Fm: Pete Grey [Microsoft] 70744,22 To: Dan Sullivan 76327,1534 (X) >> I would like to make use of the IO Control funtions to manipulate floppies in my applications. Will this capability be added to NT for the final release? When will it be available to developers? Will complete IO Control functionality be provided? Will existing DOS app's that make use of IO Control for floppies be supported? This lack of IO Control is the only fly in the ointment for the large NT app we are in the process of developing. We must be able to write floppies in non-DOS formats. We are faced with having our users reboot DOS or have a separate DOS system to write these floppies. Needless to say neither is really a desirable solution. << It is possible to open a floppy for direct reads and writes by doing a CreateFile("\\\\.\\a:",.....). This will allow you to use ReadFile() and WriteFile() to do transfer data from/to the floppy in chunks at offsets. I believe this will be going into VDM support as well. -pete There is 1 Reply. #: 16703 S6/API-Base/Security 05-Nov-92 04:28:52 Sb: #16665-IO Control Fm: Dan Sullivan 76327,1534 To: Pete Grey [Microsoft] 70744,22 (X) I don't think that this will work for what we want to do, but I don't know all the details, I will not be writing this portion of our app. However I don't think this would allow us to work with disk with sector sizes other than some default assummed by NT or that might be interleaved in some strange fashion. It seems like allowing IO Control for floppies could be done without violating any of the security provisions for NT. I am sure your work around would work for some cases, but my question remains, what is the status of IO Control for floppies in NT, especially the DOS subsystem? Thanks Dan #: 16363 S6/API-Base/Security 02-Nov-92 13:04:05 Sb: Shared mem. question Fm: Megatronics, Inc. 74020,2170 To: SYSOP (X) Hi there. I'm trying to prepare for Windows/NT today figuring I'll save myself some effort later. Along these lines, can you give me some details on how Windows/NT handles shared memory? I know in Windows 3.1 that a pointer is a pointer -- all applications share a single LDT. What about NT? Thanks in advance. John There is 1 Reply. #: 16481 S6/API-Base/Security 03-Nov-92 11:57:49 Sb: #16363-Shared mem. question Fm: Pete Grey [Microsoft] 70744,22 To: Megatronics, Inc. 74020,2170 (X) >> I'm trying to prepare for Windows/NT today figuring I'll save myself some effort later. Along these lines, can you give me some details on how Windows/NT handles shared memory? I know in Windows 3.1 that a pointer is a pointer -- all applications share a single LDT. What about NT? Thanks in advance. << In Win32/NT, each process is in its own address space and therefore cannot share memory in the old Win31 fashion. Memory must be mapped into each process' space in order to have shared memory. The shared memory overview details this. -pete There is 1 Reply. #: 16490 S6/API-Base/Security 03-Nov-92 13:25:07 Sb: #16481-Shared mem. question Fm: Megatronics, Inc. 74020,2170 To: Pete Grey [Microsoft] 70744,22 (X) Pete >> The shared memory overview details this. Not to sound silly, but where would I locate this? I have the Developer's Network CD, and I haven't got it up and running yet -- will I find it there? If not, please advise. Thanks for the reply. John There is 1 Reply. #: 16643 S6/API-Base/Security 04-Nov-92 15:52:06 Sb: #16490-Shared mem. question Fm: Pete Grey [Microsoft] 70744,22 To: Megatronics, Inc. 74020,2170 (X) >> Not to sound silly, but where would I locate this? I have the Developer's Network CD, and I haven't got it up and running yet -- will I find it there? If not, please advise. << Yes, they are on there, but these are the old version (July release). The newest help files are on the Oct. release of the SDK. -pete There is 1 Reply. #: 16707 S6/API-Base/Security 05-Nov-92 05:42:11 Sb: #16643-Shared mem. question Fm: Megatronics, Inc. 74020,2170 To: Pete Grey [Microsoft] 70744,22 (X) Ah. I didn't realize the CD files were that out of date. I guess I'll have to check into the Oct. SDK release. John #: 16709 S6/API-Base/Security 05-Nov-92 06:59:16 Sb: MFC Comm Dlgs Fm: Trip Tucker 76054,2250 To: Sysop (X) It seems that there are still some problems with the Print Common Dialog and MFC in the new release. While this worked with the previous release, the dialog does not initialize with the correct page numbers from the structure and then does not return with the page numbers that were set. The Font Dialog worked with the last release, but now tells you that there are no fonts installed for the printer. This is the same code that works for WIN16 and worked with the July release. Have you seen this? Trip. #: 16548 S6/API-Base/Security 03-Nov-92 21:36:08 Sb: #16198-GlobalSize fails on dlg Fm: Eric Sassaman/Microsoft 75430,411 To: Len Gray 71630,1703 (X) << The code that I have been using to checksum a dialog is now failing under Windows NT (Oct). This code works fine under Windows 3.1. Under Windows NT, I seem to be getting values back for my find, load, and lock calls. The GlobalSize, however, returns 0. >> Len, I'm looking into this problem and will get back to you as soon as I can get some more information on this. Thanks! Eric Sassaman Microsoft Developer Support #: 16668 S6/API-Base/Security 04-Nov-92 18:35:43 Sb: #16198-GlobalSize fails on dlg Fm: Eric Sassaman/Microsoft 75430,411 To: Len Gray 71630,1703 (X) << The code that I have been using to checksum a dialog is now failing under Windows NT (Oct). Under Windows NT, I seem to be getting values back for my find, load, and lock calls. The GlobalSize, however, returns 0. >> The problem is that the underlying resource object returned by LoadResource is not compatible with a memory object. Calling GlobalSize on resource handles worked under 3.1 because these objects were in fact implemented as GlobalAlloc'd blocks of memory. This kind of object type casting is depending on implementation specific behavior that was subject to change, which it has under Win32. There are some similar issues with certain GDI objects, from what I understand. Best to turn on STRICT and make sure that you're not mixing types like this. This will help you catch incompatible types being used in this way. Eric Sassaman Microsoft Developer Support There is 1 Reply. #: 16711 S6/API-Base/Security 05-Nov-92 07:22:25 Sb: #16668-GlobalSize fails on dlg Fm: Len Gray 71630,1703 To: Eric Sassaman/Microsoft 75430,411 (X) Thank you for your research and explanation, Eric. Checksumming the dialog is important to insure that the original copyright and registration request remains on the shareware that I produce. Resource editors readily available to the public make it too easy for this information to be changed inside a Windows application. If you could suggest the proper way to examine a dialog resource at runtime under Windows NT I would appreciate it very much. Thanks, Eric. Len #: 16296 S6/API-Base/Security 01-Nov-92 10:05:47 Sb: Setting comm queue sizes Fm: Jeff Thomson 71460,3222 To: All Under Win 3.x, the size of the transmit & receive queues for the comm port are specified as arguments to the OpenComm() function. Under NT, the size of the queues can be queried with the GetCommProperties() function, but there appears to be no way to set the size of the queues. Is there an API function under NT which allows one to specify the size of the transmit and receive queues, or are they of fixed size ? Thanks, -- Jeff There is 1 Reply. #: 16547 S6/API-Base/Security 03-Nov-92 21:36:03 Sb: #16296-Setting comm queue sizes Fm: Eric Sassaman/Microsoft 75430,411 To: Jeff Thomson 71460,3222 (X) << Under Win 3.x, the size of the transmit & receive queues for the comm port are specified as arguments to the OpenComm() function. Under NT, the size of the queues can be queried with the GetCommProperties() function, but there appears to be no way to set the size of the queues. Is there an API function under NT which allows one to specify the size of the transmit and receive queues, or are they of fixed size ? >> You can give suggested buffer sizes with the SetupComm API. Eric Sassaman Microsoft Developer Support There is 1 Reply. #: 16712 S6/API-Base/Security 05-Nov-92 07:36:43 Sb: #16547-Setting comm queue sizes Fm: Jeff Thomson 71460,3222 To: Eric Sassaman/Microsoft 75430,411 (X) Eric, > You can give suggested buffer sizes with the SetupComm API. That's just what I was looking for. Thanks. -- Jeff #: 16629 S6/API-Base/Security 04-Nov-92 13:37:33 Sb: hel Fm: Rob Tessier 76666,3633 To: all Are POSIX apps written to a specific API set that excludes the Win32 APIs? The sample POSIX app that came with the Oct SDK makes no Win32 API calls. Is that the idea? If so, what is the GUI for POSIX apps? Or are POSIX apps treated like DOS apps, they run in a windows but don't actually directly call into Win32? Rob There is 1 Reply. #: 16664 S6/API-Base/Security 04-Nov-92 18:32:50 Sb: #16629-hel Fm: Pete Grey [Microsoft] 70744,22 To: Rob Tessier 76666,3633 (X) >> Are POSIX apps written to a specific API set that excludes the Win32 APIs? The sample POSIX app that came with the Oct SDK makes no Win32 API calls. Is that the idea? If so, what is the GUI for POSIX apps? Or are POSIX apps treated like DOS apps, they run in a windows but don't actually directly call into Win32? << There are no cross-subsystem calls allowed at all, be it from POSIX to Win32 or DOS to Win32 or Win32 to DOS, etc. This would limit POSIX applications to character mode running in a window. -pete There is 1 Reply. #: 16731 S6/API-Base/Security 05-Nov-92 12:18:40 Sb: #16664-hel Fm: Rob Tessier 76666,3633 To: Pete Grey [Microsoft] 70744,22 (X) Thanks, that is what I suspected all along. -- RPT #: 16823 S6/API-Base/Security 06-Nov-92 11:01:49 Sb: Queues in NT Fm: Poul Erik Hansen 70642,2047 To: SYSOP (X) Hi: I could not find any Queues API in Win32. Are named pipes the only way to implement message queues in NT? Thanks, --Prashant #: 16778 S6/API-Base/Security 05-Nov-92 17:57:19 Sb: InvalidRect problem Fm: David Arathorn 70412,2772 To: ms #: 16840 S6/API-Base/Security 06-Nov-92 13:38:09 Sb: #16778-InvalidRect problem Fm: Pete Grey [Microsoft] 70744,22 To: David Arathorn 70412,2772 (X) David, No text was sent with your message. Please re-send, but judging from the title, it looks as if this is either a USER or GDI type query, so it needs to go to the appropriate section. Thanks, -pete #: 16867 S6/API-Base/Security 06-Nov-92 17:11:47 Sb: TCPIP problems Fm: Joseph E. Greer 70253,3130 To: Joe Greer 70253,3130 I have recently installed the October Windows/NT. I recompiled a program that I used with the july release to transfer files from our unix system to the micro running NT via a TCPIP interface. I was surprised to discover that the program now runs only 15 min (functioning correctly) before the system hangs. Has anyone else encountered a problem like this? If so, how did you get around it? This program I mentioned is rather central to our product generation process and could really use a work around. PS I did notice that the services are now called TCPIP and TELNET (contrary to the release notes) -- Joe #: 16281 S7/Tools-Microsoft 31-Oct-92 14:16:50 Sb: Dialog Editor- !Win32 Fm: Jake Kirk 76207,1403 To: SYSOP (X) I am getting the following error message when I read my RES file into the dialog editor- "test.res is not a valid Win32 format resource file." Is something wrong with my nmake makefile file? I think it is OK. Here are the 3 files necessary to recreate the problem: nmake 'makefile' as follows: ----------------------------- #: 16299 S7/Tools-Microsoft 01-Nov-92 12:06:28 Sb: #16258-C++ syntax errors? Fm: Julie Solon [Microsoft] 71075,107 To: Bill Cohagan 74375,313 (X) Bill, You are welcome. I appreciate you narrowing your code and confirming that we are receiving the same error messages. We are passing compiler problems on as fast as we can. I will get back to you when I know more. Julie Solon Microsoft Developer Support #: 16300 S7/Tools-Microsoft 01-Nov-92 12:06:35 Sb: _CRT_INIT failure Fm: Julie Solon [Microsoft] 71075,107 To: Ted C. Ho 73310,150 (X) Ted, I assume that you are just trying to compile this single-threaded application as if it were a multi-threaded application - i.e., you changed only switches, not code. First, I tested the code with the existing makefile. When I exited WinDbg, I received an access violation. Next, I made the following changes compiled select with -D_MT, demo with -D_MT linked select with $(guilibsdll), demo with $(guilibsmt) The resulting behavior didn't seem any different than before. Finally, I changed the entry point as you had done to -entry:_CRT_INIT@12. Again, I get the access violation when exiting WinDbg. I will look into this problem that I am seeing with the unmodified sample. If you are seeing different symptoms... like everything works as it should unmodified (but not any other way) or the access violation happens any other time but at exit of WinDbg, let me know. In that event, we would need to narrow down cause and effect so that I can understand and reproduce exactly what you are seeing here. Thank you for your report, Julie Solon Microsoft Developer Support #: 16301 S7/Tools-Microsoft 01-Nov-92 12:06:40 Sb: Compiler Bug Fm: Julie Solon [Microsoft] 71075,107 To: David J. Plunkett 71163,2122 David, Thank you for reporting this problem. I was able to reproduce it (the only thing I did was change %f to %lf where you were printing doubles). There does seem to be a growing number of problem reports against this compiler. We are attempting to track them down and get out work-arounds and fixes. I will file a report right away and post more information as it becomes available. Thanks, Julie Solon Microsoft Developer Support #: 16306 S7/Tools-Microsoft 01-Nov-92 13:52:36 Sb: #16227-fatal compiler error Fm: Julie Solon [Microsoft] 71075,107 To: Orin Eman 71102,705 (X) Orin, There have been postings on this error occuring with the October compiler and not the July compiler. Unfortunately, sometimes when known problems are fixed, others crop up. We are very interested in receiving test cases and pursuing these problems. We have a sample of this particular error, so hopefully, in fixing the compiler for one test case, we will fix it for most others. If you could submit a reasonably sized test case, with compiler options and whatever work-arounds you find, we would be happy to look into it. Julie Solon Microsoft Developer Support #: 16305 S7/Tools-Microsoft 01-Nov-92 13:52:30 Sb: Fatal Compiler Error Fm: Julie Solon [Microsoft] 71075,107 To: Bill Cohagan 74375,313 (X) Bill, Thank you for your efforts in narrowing down the code. I could not find any minor code or compile option changes that would work around the problem immediately. I will report this problem right away and I will let you know what happens. Thanks, Julie Solon Microsoft Developer Support There is 1 Reply. #: 16325 S7/Tools-Microsoft 02-Nov-92 07:25:30 Sb: #16305-Fatal Compiler Error Fm: Bill Cohagan 74375,313 To: Julie Solon [Microsoft] 71075,107 (X) Julie- Thanks for looking into this. This compiler error has brought me to a stop on this project -- so a rapid fix/workaround would be *greatly* appreciated. Thanks, Bill #: 16353 S7/Tools-Microsoft 02-Nov-92 11:53:49 Sb: #16257-Oct/July NT coexist? Fm: Colin Stuart [Microsoft] 70744,25 To: Bill Cohagan 74375,313 (X) Bill, You can install both the July and October versions of Win32/NT on one drive, with the following considerations: -Place them in separate directories. :) -If you installed NT with the CD setup, BOOT.INI will use the SCSI()disk()rdisk()partition(n)\... notation, which requires the SCSI miniport driver C:\NTBOOTDD.SYS, which differs between the July and October releases. You can most likely replace this with normal DOS paths (i.e. c:\nt...) in BOOT.INI, depending on how your partitions are set up (if both are in the first partition of your first drive, there should be no difficulties). Doing this would alleviate the need for NTBOOTDD.SYS. If this prodcedure fails, you must first boot into DOS and copy the appropriate NTBOOTDD.SYS to the root before switching between versions of Win32/NT. Colin #: 16358 S7/Tools-Microsoft 02-Nov-92 12:16:32 Sb: WinDbg Fm: Howard Myers 76711,462 To: Paul Tissue [Microsoft] 70744,24 (X) I've solved both problems! The file that wouldn't load into WinDbg at all was because a line was too long. There was a comment that went out to column 228. WinDbg apparently won't load any files that have line(s) beyond about 200 columns. The file that would load but wouldn't allow me to set any breakpoints was named UTIL.C. This must be reserved for WinDbg. The exact same file renamed as UTIL0.C, works fine. It would be nice if the tools/documentation gave some warning about things of this nature. Particularly the long line lengths, must be hard coded into WinDbg. When someone writes code like this, why don't they add just one line of code to pop up a message dialog that tells us why they won't load it; instead of doing absolutely nothing and leaving us totally perplexed? The UTIL.C problem isn't as clear that it's in the code and could have been flagged, however, it seems like it almost would have had to have been coded that way to exclude just one file name. Who knows? Hopefully, this can help other people from fighting the same problems! #: 16368 S7/Tools-Microsoft 02-Nov-92 14:11:01 Sb: #16255-*MAJOR* bug with new rel Fm: Colin Stuart [Microsoft] 70744,25 To: Marco Papa 70353,3446 (X) Marco, The screen update speed has been considerably improved in the October release, and I believe that product 1 will move even faster. Glad you noticed :) -c #: 16372 S7/Tools-Microsoft 02-Nov-92 14:14:40 Sb: #16206-RemoveDir() Fm: David Taniguchi [MS] 72350,2054 To: Andrew Bradnan [Erudite 70402,63 (X) Hi Andrew, >This is a MSSETUP question, point me in the right direction if I am in >the wrong place. Since this is just with the API, base (section 6) would be a better place to ask this in the future. >IAC, using Windows NT or Windows 3.1 (16bit) setup fails on RemoveDir(). >The directory is empty and DoesDirExist says it is there but it >still does not work. It is giving me the API error not BadArg1 error. >The only thing I can think of is that I am in the current directory and >the OS is dumping on me for that. Please post a response in the base section (6), but are you trying to remove a directory that you are executing the setup program? Thanks, Dave #: 16410 S7/Tools-Microsoft 02-Nov-92 16:46:27 Sb: Porttool Fm: Randy Kath 72110,2014 To: Paul Tissue [Microsoft] 70744,24 (X) Porting PortTool to 16bits will be a fairly easy process. The code is basic API code, no tricks or hacks. The only two (major) areas where the Win32 API is exploited is in the use of memory mapped files for sharing data between processes that call the PORT.DLL and the background processing that was just added in the last release. You can omit that feature in the 16 bit Windows version by commenting out a couple lines of code and not compiling and linking the file PTBKPORT.C. The memory mapped file issue is easy to solve in 16bit land since global memory can be allocated by the DLL for all to use by default (unlike Win32/NT Paul is correct in the 30,000 character limit. PortTool uses the standard multiline edit control for the editor interface, so this would be a limitation. Yet, as an alternative, the DLL supplied with PortTool exports the porting functionality. Refer to the file PORT.H for more information on how to make another application call into the DLL #: 16411 S7/Tools-Microsoft 02-Nov-92 16:46:58 Sb: Porttool Fm: Randy Kath 72110,2014 To: Paul Tissue [Microsoft] 70744,24 (X) Continue from previous message And speaking of limitations, the current version of PortTool included in the October release of the NT SDK also has the 30,000 character limit by default This can be fixed by simply sending the EM_LIMITTEXT message to the edit control immeditely after creation. Specifically, in the file PORTTOOL.C after line 215 and before line 229 add the following line SendMessage (hWndEdit, EM_LIMITTEXT, 0xFFFFFFF, 0); The default behavior of the edit control is to limit text to 30,000 characters. Yet, the edit control is capable of handling up to 0x0FFFFFFF characters. Also, sending the above message with a wParam value of 0 has the same effect of allowing the maximum number of characters. Also note: the latest documentation available at time of the October releae did not accurately portray this edit control behavior. It will be accurate in the release The version of PortTool found on the NT SDK has this limitation. The version supplied on the next Microsoft Developer Network CD will not have this problem. Also, in the interim you can find a fixed version of PortTool placed in the MSDNLIB forum in section 1. It should be placed there tonight Thanks Randy Kath, Microsoft Developer Network #: 16414 S7/Tools-Microsoft 02-Nov-92 17:00:48 Sb: #16253-problems with Oct rel. Fm: Petrus Wong [Microsoft] 70743,3355 To: Marco Papa 70353,3446 (X) Hello Marco: >>3. When compiling resources built with the Windows 3.1 dialog editor, the NT resource compiler will break immediately when processing multiple .DLG files that all contain an initial line of DLGINCLUDE. The July release of the NT Resource compiler did not have this pnroblem. Also this makes .DLG files from Windows 3.1's Dialog Editor totally imcompatible with NT.<< There are some changes regarding the usage of DLGINCLUDE since the July PDC build of Windows NT. Please check out the following article for detail: thanks, petrus ps. Please post questions regarding the installation of the OS itself on the WINNT forum. ------------------------------------------------------------------------------- Summary: The Windows 3.1 SDK Dialog Editor needs a way to know what include file is associated with a resource file that it opens. Rather than prompt the user for the same of the include file, the name of the include file is embedded in the resource file in most cases. Embedding the name of the include file is done with a resource of type RCDATA with the special name DLGINCLUDE. This resource is placed into the .res file and contains the name of the include file. The dialog editor looks for this resource when it loads a .res file. If it is found, then the include file is opened also. If not, the editor prompts the user for the name of the include file. In some Windows 3.1 build environments, the dialog editor was used to create dialogs which were placed in more than one .dlg file. These different .dlg files were then included in one .rc file that was compiled with the resource compiler. So the resource file gets multiple copies of a RCDATA type resource with the same name, DLGINCLUDE, but the resouce compiler and dialog editor do not complain. In the Win32 Preliminary SDK, changes were made so that this resource has its own resource type. It was changed from an RCDATA type resource with the special name, DLGINCLUDE, to a DLGINCLUDE resource type whose name can be specified. The dialog editor would look for resources of the type DLGINCLUDE. [More] There is 1 Reply. #: 16415 S7/Tools-Microsoft 02-Nov-92 17:00:56 Sb: #16414-problems with Oct rel. Fm: Petrus Wong [Microsoft] 70743,3355 To: Petrus Wong [Microsoft] 70743,3355 (X) [Continued] For the October beta, there was a change made to CvtRes so that it gives an error if it finds a resource that has the same type, name, and language as another resource in the file. We are being stricter about the resources needing to be unique in the Win32 SDK than the Windows 3.1 SDK. This is good, because there was never any guarantee at run time which of the two or more resources would be returned by LoadResource(). This means that some apps being ported to Windows NT will give an error when their resources are compiled, because they will have duplicate RCDATA type resources with the same name (DLGINCLUDE). This error is by design. The work-around is straightforward: delete all the DLGINCLUDE RCDATA type resource statements from all the .dlg files. Finally, because it does not make much sense to have the DLGINCLUDE type resources in the executable at all, CvtRes will strip them out so that they do not get linked into the exe. #: 16275 S7/Tools-Microsoft 31-Oct-92 09:26:04 Sb: cout?? Fm: Ben Laurie 100014,1235 To: Doug Olson Microsoft 72350,2635 (X) Doug, Much delayed result ... yep, using ...\MFC\?.LIB worked fine, and, furthermore, this version of C++ did NOT display the bug that the DOS version does. Ben. There is 1 Reply. #: 16422 S7/Tools-Microsoft 02-Nov-92 18:49:10 Sb: #16275-cout?? Fm: Doug Olson [Microsoft] 72350,2635 To: Ben Laurie 100014,1235 (X) Ben: Glad to hear that it worked. Doug Olson Microsoft Developer Support #: 16380 S7/Tools-Microsoft 02-Nov-92 14:52:28 Sb: #16254-problems (continued) Fm: Petrus Wong [Microsoft] 70743,3355 To: Marco Papa 70353,3446 (X) Hello Marco, >>Given the two above problems, debugging on MIPS with the October release is practically impossible. I could only build my app by taking out both the -g2 flag from compilation and the -c flag from MIp2coff. The file got linked OK, but of course it has no debugging info any more. Given this, the October release is worse than the July one as far as debugging on MIPS is concerned.<< Thanks for the bug report. This is a known problem with "mip2coff". We are looking into this currently. thanks, petrus There is 1 Reply. #: 16428 S7/Tools-Microsoft 02-Nov-92 19:36:46 Sb: #16380-problems (continued) Fm: Marco Papa 70353,3446 To: Petrus Wong [Microsoft] 70743,3355 (X) One way I was able to do some debugging ona a large program on MIPS, was to compile *everything* without debuggingh stuff, and then one file at a time add the -g2 and -c switches. It is really not completely useful, since you must know the file where the problem resides, but it is better than nothing. You're welcome, BTW, Petrus. Since our Online account has expired, I'm glad to see you her on CIS. -- Marco #: 16359 S7/Tools-Microsoft 02-Nov-92 12:26:01 Sb: WinDbg won't load source Fm: Howard Myers 76711,462 To: Brent Ingraham (M&M) 70550,2303 (X) I had the same problem. In my case it was the result of a line of code (a comment) that was 228 characters long. WinDbg appears to not be able to load any files that have a line over approx. 200 chars. I don't know the exact cut-off, but it is around 200. (Hardly seems like a magic number, does it?) Hope this helps! #: 16369 S7/Tools-Microsoft 02-Nov-92 14:11:05 Sb: WinDbg won't load source Fm: Colin Stuart [Microsoft] 70744,25 To: Brent Ingraham (M&M) 70550,2303 (X) Brent, I'm looking into this. Colin #: 16487 S7/Tools-Microsoft 03-Nov-92 12:37:25 Sb: WinDbg won't load source Fm: David Edge 75170,1461 To: Brent Ingraham (M&M) 70550,2303 (X) I had the same problem. After a bit of probing around, I looked at the debug options. Look at Options/Debug, Debug DLLs. There is an item for the expression evaluator. It needs to be changed to EECXX.DLL. That fixed the problem for me. #: 16460 S7/Tools-Microsoft 03-Nov-92 09:39:41 Sb: LIB discards Locals Fm: David Manns 100112,2773 To: ALL Putting C object files into a library using LIB appears to discard all the debugging information. When you link those libraries into your app, you no longer have access to the locals within those objects. It works okay if you link the objects directly without putting them into a library. Is this a known problem which will be (or has been) addressed? There is 1 Reply. #: 16489 S7/Tools-Microsoft 03-Nov-92 13:23:25 Sb: #16460-LIB discards Locals Fm: Doug Olson [Microsoft] 72350,2635 To: David Manns 100112,2773 David: For debugging information to be preserved it is necessary to use the -DEBUGTYPE:{COFF|CV|BOTH} switch on the LIB (or COFF -LIB) command line. Please let me know if you are already using this switch. Sincerely, Doug Olson Microsoft Developer Support #: 16503 S7/Tools-Microsoft 03-Nov-92 15:16:17 Sb: DLGINCLUDE Fm: Colin Stuart [Microsoft] 70744,25 To: Thomas Kerr 75140,3326 Thomas, This posting is in regards to your bug report uploaded in the WINNT forum. Please report other SDK bugs in this forum, and end-user bugs in the WINNT templates. Thank you, Colin From MDKB on CIS (GO MDKB): Summary: .TITLE: INF: Use of DLGINCLUDE in Resource Files The Windows 3.1 SDK Dialog Editor needs a way to know what include file is associated with a resource file that it opens. Rather than prompt the user for the same of the include file, the name of the include file is embedded in the resource file in most cases. Embedding the name of the include file is done with a resource of type RCDATA with the special name DLGINCLUDE. This resource is placed into the .res file and contains the name of the include file. The dialog editor looks for this resource when it loads a .res file. If it is found, then the include file is opened also. If not, the editor prompts the user for the name of the include file. In some Windows 3.1 build environments, the dialog editor was used to create dialogs which were placed in more than one .dlg file. These different .dlg files were then included in one .rc file that was compiled with the resource compiler. So the resource file gets multiple copies of a RCDATA type resource with the same name, DLGINCLUDE, but the resouce compiler and dialog editor do not complain. In the Win32 Preliminary SDK, changes were made so that this resource has its own resource type. It was changed from an RCDATA type resource with the special name, DLGINCLUDE, to a DLGINCLUDE resource type whose name can be specified. The dialog editor would look for resources of the type DLGINCLUDE. For the October beta, there was a change made to CvtRes so that it gives an error if it finds a resource that has the same type, name, and language as another resource in the file. We are being stricter about the resources needing to be unique in the Win32 SDK than the Windows 3.1 SDK. This is good, because there was never any guarantee at run time which of the two or more resources would be returned by LoadResource(). [More] There is 1 Reply. #: 16504 S7/Tools-Microsoft 03-Nov-92 15:16:24 Sb: #16503-DLGINCLUDE Fm: Colin Stuart [Microsoft] 70744,25 To: Colin Stuart [Microsoft] 70744,25 (X) [Continued] This means that some apps being ported to Windows NT will give an error when their resources are compiled, because they will have duplicate RCDATA type resources with the same name (DLGINCLUDE). This error is by design. The work-around is straightforward: delete all the DLGINCLUDE RCDATA type resource statements from all the .dlg files. Finally, because it does not make much sense to have the DLGINCLUDE type resources in the executable at all, CvtRes will strip them out so that they do not get linked into the exe. #: 16403 S7/Tools-Microsoft 02-Nov-92 16:34:06 Sb: WinDbg problems Fm: Jeff Thomson 71460,3222 To: All I'm having two major problems trying to debug my application using WinDbg. The application is written using MFC, although I'm uncertain whether that has anything to do with the problems I'm encountering. Problem #1 is that no matter where I set a breakpoint in my program, the Unresolved Breakpoints dialog box pops up when I try to run my program. Problem #2 is that it looks like my app's InitInstance() function is not being executed. When I step through WinMain(), the following output appears in the Command window: Module Load: C:\projects\myprog\MYPROG.EXE (symbols loaded) Thread Create: Process=0, Thread=0 Module Load: C:\WINNT\SYSTEM\NTDLL.DLL (symbols loaded) Module Load: C:\WINNT\SYSTEM\GDI32.DLL (symbols loaded) Module Load: C:\WINNT\SYSTEM\KERNEL32.DLL (symbols loaded) Module Load: C:\WINNT\SYSTEM\USER32.DLL (symbols loaded) ERROR: [0583] Class does not exist. ERROR: [0583] Class does not exist. > The two error messages above appear to occur just as I single-step over the call in WinMain() to my applications's InitInstance(). My question is, what do these messages mean ? They say "Class does not exist."; but *which* class(es) ?!? Thanks in advance. -- Jeff There is 1 Reply. #: 16520 S7/Tools-Microsoft 03-Nov-92 17:22:18 Sb: #16403-WinDbg problems Fm: Colin Stuart [Microsoft] 70744,25 To: Jeff Thomson 71460,3222 (X) Jeff, I'm corresponding with development on these. I'll get back to you. Colin #: 16525 S7/Tools-Microsoft 03-Nov-92 17:29:04 Sb: C compiler bug Fm: Doug Olson [Microsoft] 72350,2635 To: Jeff Robbins 70303,1570 (X) Jeff: Your observation regarding the .asm code is correct. One of our language developers checked into it and confirmed the problem. I will let you know when more information is available. Sincerely, Doug Olson Microsoft Developer Support #: 16348 S7/Tools-Microsoft 02-Nov-92 11:11:23 Sb: Tools Survey: what do yo Fm: Colin Stuart [Microsoft] 70744,25 To: ALL Folks, Please respond in this thread, please. -Colin --------------------------This message is being posted as part of our effort to get more information towards developing great new Windows systems. We want to move to object and component based development and our Windows group is aiming to create the ultimate programmable system. This is only a small scale survey, amongst people we believe are active professional developers that we have higher than average contact with. We'd really appreciate your feedback in answer to the questions below. Being small scale, we'd ask doubly that you reply to this! Many of the questions are deliberately free-form, and might deserve longer answers. We really appreciate you time, and hope that you'll enjoy helping us make Windows a better product Alistair Banks Developer Relations Group Microsoft _____TOOLS_____ 1. If you could choose your IDEAL SDK available from Microsoft, what would it include? 2a. What Tools do you use currently for Windows development? 2b. What specifically do you like about these tools? 2c. What would you change? 3a. What graphical programming environments have you used? 3b. What was good, what was bad? 4. Do you anticipate using Microsoft Foundation Classes Microsoft Visual Basic Borland OWL C++ Other (please specify) _____ C vs. C++_____ 5. If your application is in C, would you consider doing new development in C or C++... In the same application? In a new application? 6. If you have a handle in C, and a class pointer in C++ do they have to be the same (Y/N)? would you call an API to convert them (Y/N)? 7. How often do you anticipate passing handles between C and C++? _____Object/Component Programming_____ 8a. What Object Oriented programming have you done? 8b. Which environments? 9a. With the direction toward object and component programming, what specifically would you like to see? 10. What components would you need? There are 3 Replies. #: 16365 S7/Tools-Microsoft 02-Nov-92 13:54:45 Sb: #16348-Tools Survey: what do yo Fm: William Lees 75300,250 To: Colin Stuart [Microsoft] 70744,25 (X) _____TOOLS_____ 1. If you could choose your IDEAL SDK available from Microsoft, what would it include? An integrated C/C++ development environment similar in look and feel to QCWIN, but without bugs, nd including version control software and a MAKE facility capable of building multiple targets. A highly graphical and interactive help file development system with lightning performance 2a. What Tools do you use currently for Windows development? QCWIN 3.1 SDK Spy, CodeView Turbo C++ 3.1 Brief 2b. What specifically do you like about these tools? 2c. What would you change? I find QCWIN a very productive tool to work with, almost entirely because of its integrated debug er. However I can't get used to the editor and I still shell out to Brief. So I would change the editor to mak it a Windows clone of Brief, more or less. Spy is essential, but it could do with some improvements: it hasn't been changed in years. It ne ds more specific 'tuning' of messages to log, decoding of message parameters to English, and perhaps the ability t define one's own messages and parameters so that they can be decoded too. I would like the debugger to be able to descend into machine code, when I want it to. I would like a list of ALL the messages that can appear in the debug kernel's debug window, with brief explanation of each. For example the other day I had 'local heap is busy' Well so I would expect :-), but w ere do I look for an explanation of this message??? 3a. What graphical programming environments have you used? 3b. What was good, what was bad? As above. Turbo C++ just doesn't feel right to me, but it is faster than QCWIN, and the debugger is less buggy., The editor is IMO much worse. 4. Do you anticipate using Microsoft Foundation Classes yes Microsoft Visual Basic no. PLEASE don't force us to learn Basic, visual or otherwise. Borland OWL no. C++ yes Other (please specify) _____ C vs. C++_____ 5. If your application is in C, would you consider doing new development in C or C++... yes C/C++ There is 1 Reply. #: 16374 S7/Tools-Microsoft 02-Nov-92 14:16:24 Sb: #16365-Tools Survey: what do yo Fm: William Lees 75300,250 To: William Lees 75300,250 (X) (continued) yes C/C++ In the same application? yes C/C++ In a new application? yes C/C++ 6. If you have a handle in C, and a class pointer in C++ do they have to be the same (Y/N)? no would you call an API to convert them (Y/N)? always 7. How often do you anticipate passing handles between C and C++? I'll live with whatever you think is most sensible. _____Object/Component Programming_____ 8a. What Object Oriented programming have you done? 8b. Which environments? none. 9a. With the direction toward object and component programming, what specifically would you like to see? sorry, don't understand the question well enough to comment. 10. What components would you need? << #: 16475 S7/Tools-Microsoft 03-Nov-92 11:01:24 Sb: #16348-Tools Survey: what do yo Fm: Scott Alexander 76556,557 To: Colin Stuart [Microsoft] 70744,25 (X) colin Can we upload a WFW file as a reply (where to so it will get found) I can not do the questions justice in a message (or 5). If not, should we send private to reduce scroll (I would prefer an upload so I can include diagrams). There is 1 Reply. #: 16519 S7/Tools-Microsoft 03-Nov-92 17:22:14 Sb: #16475-Tools Survey: what do yo Fm: Colin Stuart [Microsoft] 70744,25 To: Scott Alexander 76556,557 (X) Scott and ALL, We would like to hear from as many of you folks as possible, so here are some other alternitives that may be more convenient for you: You may post your answer publicly or privately in this section, or you may email, fax, or mail your responses to: Windows Cairo Design-Preview Feedback Fax: 206-93-MS-FAX One Microsoft Way, Redmond, WA 98052 Internet: cairodp@microsoft.com Thanks, Colin #: 16529 S7/Tools-Microsoft 03-Nov-92 17:53:52 Sb: #16348-Tools Survey: what do yo Fm: Daniel Green 70312,3470 To: Colin Stuart [Microsoft] 70744,25 (X) Regarding the survey: All I want (right now :-) is a robust, reliable, high speed, C++ compiler with a debugger that fully supports C++ and a post-mortem tool similar to the Winspecter shipping with Borland CPP for Win 3.1. I don't care about an IDE environment or any other bells and whistles; command line is fine. What I do care about is that the compiler is generating good code and not crashing with internal compiler errors or CVPACK errors. In regards to speed, the compiler is too slow right now (this is exacerbated by the extra steps of cvpack, converting the resource file, etc). The traditional edit compile link cycle takes much longer under NT then it does under Win 3.1. #: 16297 S7/Tools-Microsoft 01-Nov-92 11:02:17 Sb: Compiling SDK Samples Fm: Jeff Richardson 73670,1672 To: ALL When I try to run the MAKEALL.BAT file, the system seems to compile the samples ok, but does not LINK. The linker reports error U1077 and a return code of '128'. I look in the docs and could not find a return code of 128. HELP, what is wrong? thanks jeff> There is 1 Reply. #: 16386 S7/Tools-Microsoft 02-Nov-92 15:23:30 Sb: #16297-Compiling SDK Samples Fm: David Taniguchi [MS] 72350,2054 To: Jeff Richardson 73670,1672 (X) Hi Jeff, Look in the file makeall.err to see which sample it fails on. Also, run the setenv.bat file to set your environment variables (setupsdk should do this for you, but the changes will not occur until you reboot your machine). What type of machine is this on? (MIPS or x86) Thanks, Dave There is 1 Reply. #: 16555 S7/Tools-Microsoft 03-Nov-92 22:20:57 Sb: #16386-Compiling SDK Samples Fm: Jeff Richardson 73670,1672 To: David Taniguchi [MS] 72350,2054 (X) Dave, Thanks for the reply, after checking and re-checking all the things you mentioned. I re-installed the SDK and the problem seems to be fixed. When I installed the SDK the first time, there was a error concerning one of the readme.txt files. The only thing I can guess, is that my CD-ROM drive needs to be cleaned (I have only had it a couple months :-( ). Jeff. #: 16570 S7/Tools-Microsoft 04-Nov-92 05:53:08 Sb: Installation Problem Fm: Bob Soodak 71055,542 To: All Problem Installing NT from MS Win32 SDK CD-ROM. Motherboard: IBM PS/2 Model 40SX SCSI Adapters: TCM-850IBM manufactured for IBM by Future Domain; IBM 3.5 inch rewritable optical dreve attached, installed internally. Adaptec AHA-1510 (from Corel Systems); Corel (manufactured by Toshiba) CD-ROM attached, installed externally. (Blockbuster Bundle) Problem: MS boot disk (dated July 1992) Setup program fails to recognize the presence of the CD-ROM. Note: In all other respects, the CD-ROM functions normally. I had no problem installing the MS-Developer Network Software from the same drive. Is it possible that the Setup program assumes only one SCSI card is installed and fails to search the second for valid CD-ROMs? #: 16356 S7/Tools-Microsoft 02-Nov-92 12:04:46 Sb: Oct. SPY Fm: Mark Gamber 76450,2754 To: 70744,25 (X) Colin, One more *little* thing... : ) October's SPY's icon shows up everywhere except when the program is minimized on the desktop. The icon text appears, but the icon itself is invisible. At least on my desktop. It never ends, does it? Mark There is 1 Reply. #: 16590 S7/Tools-Microsoft 04-Nov-92 10:34:25 Sb: #16356-Oct. SPY Fm: David Taniguchi [MS] 72350,2054 To: Mark Gamber 76450,2754 (X) Hi Mark, Thanks for the problem report. You are correct, the icon disappears when SPY is minimized. I'll report this to the proper people. Thanks, Dave #: 16614 S7/Tools-Microsoft 04-Nov-92 11:56:19 Sb: New DLL build procedure Fm: neil colvin 71650,3517 To: Sysop (X) What is the rationale behind the two changes required to build a DLL in the October release? I refer to the requirement that ALL of the object files for the DLL be passed to LIB, and that a -base:0x1c000000 be specified for the DLL. Both are done in the samples, but are they both required, and what do they do? #: 16465 S7/Tools-Microsoft 03-Nov-92 09:42:51 Sb: RC Issues Fm: Samuel Feldman 70403,432 To: Colin Stuart (MSFT) 70744,25 (X) Colin, Two more RC issues: 1) When compiling a font resource, there seems to be 32 bytes of extra junk at the end of the font directory, where I expect the szFaceName[] to be. Is that a bug? 2) When compiling a Version-Information resource, there seems to be a new WORD in the version-information block structure, besides the cbBlock and the cbValue. Can you tell me what it is? Thanks -- Samuel There is 1 Reply. #: 16620 S7/Tools-Microsoft 04-Nov-92 13:16:29 Sb: #16465-RC Issues Fm: Colin Stuart [Microsoft] 70744,25 To: Samuel Feldman 70403,432 (X) Samuel, I'm looking into this. By the way, we've just uploaded a document you may be interested in. It's in library 4, RESFMT.ZIP. It contains descriptions of the binary formats of win32 resources. Enjoy, Colin #: 16518 S7/Tools-Microsoft 03-Nov-92 17:22:09 Sb: Windbg problems Fm: Colin Stuart [Microsoft] 70744,25 To: Bruce Cowan 73650,32 (X) Bruce, I'll talk with development about these, and post new info as I get it. Thanks for the post. Colin #: 16628 S7/Tools-Microsoft 04-Nov-92 13:35:38 Sb: Windbg problems Fm: Kevin Quinn 75430,255 To: Carl H Bache [PS Norway] 100010,2257 (X) Bruce, et al: I've also seen the lock problem with NTSD; pview shows a process still running that can't be 'killed'. However, all the windows have long since gone... Kevin Desktop Systems Ingres #: 16651 S7/Tools-Microsoft 04-Nov-92 16:00:49 Sb: Too Many S Fm: Doug Olson [Microsoft] 72350,2635 To: Thomas Holaday, GAT Corp 76244,1532 (X) Thomas: Would you upload the EXE file that was being CVPACK'ed when this error occurred? We are working to track down this problem but need a sample. To make sure you have the right one, you can type CVPACK PROGRAM.EXE from the command prompt and see if the error occurs. Thanks for the help, Doug Olson Microsoft Developer Support #: 16420 S7/Tools-Microsoft 02-Nov-92 18:48:59 Sb: #16242-CVPACK Error Fm: Doug Olson [Microsoft] 72350,2635 To: Koby 71172,2722 (X) Koby: Thanks for the problem report. I am looking into this error. I'll get back with you as soon as I discover the cause. Sincerely, Doug Olson Microsoft Developer Support There is 1 Reply. #: 16636 S7/Tools-Microsoft 04-Nov-92 14:16:10 Sb: #16420-CVPACK Error Fm: Koby 71172,2722 To: Doug Olson [Microsoft] 72350,2635 (X) Hi Doug. I think that I found the resone why we got CVPACK error. It based on the copmilation switch -Zp1 that we are using. Please try using this switch Thanks' Koby #: 16652 S7/Tools-Microsoft 04-Nov-92 16:00:54 Sb: #16242-CVPACK Error Fm: Doug Olson [Microsoft] 72350,2635 To: Koby 71172,2722 (X) Koby: Could you upload a sample EXE file that produces this error? To verify you have the right EXE, you can type CVPACK PROGRAM.EXE at the command prompt. You should see the error from CVPACK. Thanks, Doug Olson Microsoft Developer Support #: 16585 S7/Tools-Microsoft 04-Nov-92 09:36:49 Sb: C runtime in DLL Fm: dan white 70324,3147 To: sysop (X) I'm still a little confused about how NT and the C runtime are handling DLL's. If everyone links to CRTDLL.DLL, do a DLL and the main routine see the same state of the C runtime if they are in the same thread? I think the answer is yes, but I'm not sure? Now, how about there are multiple threads. Is there only one heap? If so, there must be mutex's and other synchronization code hidden in the heap manager. Does this make a big speed difference? Is the C runtime still faster than than the corresponding OS calls (HeapAlloc and so on). If there are multiple heaps, then can we pass allocated memory between threads? What happens if a thread terminates and other threads have pointers to memory it allocated? Thanks Dan There is 1 Reply. #: 16669 S7/Tools-Microsoft 04-Nov-92 18:37:32 Sb: #16585-C runtime in DLL Fm: Doug Olson [Microsoft] 72350,2635 To: dan white 70324,3147 (X) Dan: I will put together some information on this topic and get back with you shortly. Sincerely, Doug Olson Microsoft Developer Support #: 16294 S7/Tools-Microsoft 01-Nov-92 09:26:21 Sb: #16230-CFile destructor Fm: Jim Young 72760,725 To: Julie Solon [Microsoft] 71075,107 (X) The CFile object is declared as a local variable in the InitInstance() function. I do not use new to create it from the heap. Thanks! --Jim There is 1 Reply. #: 16477 S7/Tools-Microsoft 03-Nov-92 11:15:18 Sb: #16294-CFile destructor Fm: Julie Solon [Microsoft] 71075,107 To: Jim Young 72760,725 (X) Jim, I took the HELLOAPP sample and added the following statement in InitInstance() CFile nfile; I also compiled with the -Fc option to get a mixed source and assembler listing. I see that the constructor (CFile::CFile) is called when the variable is created and that the destruction (CFile::~CFile) is called as part of the exit from the InitInstance routine. How are you verifying that the object is not deleted? Under what circumstances does this happen? You might try my small case and work up from there if your example is large. Julie There is 1 Reply. #: 16692 S7/Tools-Microsoft 04-Nov-92 21:47:05 Sb: #16477-CFile destructor Fm: Jim Young 72760,725 To: Julie Solon [Microsoft] 71075,107 (X) The only verification that the destructor is not being called until the end of the app is that is when I get the assert message. There is no particulare circumstance that causes this - it always happens. I'll try to isolate the problem. --Jim #: 16716 S7/Tools-Microsoft 05-Nov-92 08:47:33 Sb: I've had it! Fm: Bill Cohagan 74375,313 To: sysop (X) I would appreciate your forwarding to me (via email if that's appropriate) the name and email (or USmail) address of your manager for developer relations. I am a developer with whom MS relations have been less than satisfactory and I would like to communicate this to someone who might care about such things. I regret that it has come to this, but I have no more slack to cut. Thank you, Bill #: 16687 S7/Tools-Microsoft 04-Nov-92 21:40:34 Sb: NT boot failure Fm: Daniel Pugh 70244,533 To: sysop (X) I'm trying to install the NT SDK in a BOOT MANAGER partition so that it can cohabitate with both OS/2 2.0 and DOS 5. Am I destined for failure? Since my CD-ROM is not supported, I performed the steps in the DOS2NT.BAT file by hand. When I try to boot, I get the following error: OS Loader V2.10 loading file multi(0)disk(0)partition(1)/nt/system/ntoskrnl.exe the system did not load because it cannot find the following file \system\ntoskrnl.exe. Please re-install a copy of the above file . Boot failed Does NT do some low-level partition stuff that conflicts with Boot Manager? It seemed so safe because each OS gets its own partition. I hate to have to buy yet another computer. thanks in advance -- dp There is 1 Reply. #: 16723 S7/Tools-Microsoft 05-Nov-92 10:15:29 Sb: #16687-NT boot failure Fm: David Taniguchi [MS] 72350,2054 To: Daniel Pugh 70244,533 (X) Hi Daniel, Please post this in section 3 (Windows NT Setup) of forum WinNT. They are more experienced with these types of issues. Thanks, Dave #: 16615 S7/Tools-Microsoft 04-Nov-92 11:58:54 Sb: "Dual-boot"? Fm: Kent Rollins 72040,1607 To: all Does Windows NT have a "dual-boot" capability whereby I can boot either Windows NT or DOS 5.0? I know Windows NT supports DOS itself, but I want to know whether I can boot both from the same machine. Kent There are 2 Replies. #: 16642 S7/Tools-Microsoft 04-Nov-92 15:45:07 Sb: #16615-"Dual-boot"? Fm: Chuck Atwood 76004,1602 To: Kent Rollins 72040,1607 (X) The default installation of NT on my DOS 5.0 system created a dual-boot configuration. A BOOT.INI file defines the name of each operating system. You can even have multiple version of NT if you want. I have dual booted with both the July and October release of NT. #: 16724 S7/Tools-Microsoft 05-Nov-92 10:15:34 Sb: #16615-"Dual-boot"? Fm: David Taniguchi [MS] 72350,2054 To: Kent Rollins 72040,1607 Hi Kent, Yes, "Flexboot" is installed on your system when you load it. When you boot your machine you see a list of both operating systems. You can choose whichever operating system to boot into. Flexboot will time out and boot into whichever you choose as the default in a settable amount of time. If you have further setup questions, please post in section 3 (Windows NT setup) of forum Winnt. Thanks, Dave #: 16459 S7/Tools-Microsoft 03-Nov-92 09:39:37 Sb: LINK w multi resources Fm: David Manns 100112,2773 To: ALL Is there a way of linking multiple resource objects in an executable? LINK accepts them without giving any errors, but you cannot then access the resources. It only works if there is one and only one resource object. Is this a known problem that will be addressed? There is 1 Reply. #: 16732 S7/Tools-Microsoft 05-Nov-92 12:19:27 Sb: #16459-LINK w multi resources Fm: David Taniguchi [MS] 72350,2054 To: David Manns 100112,2773 Hi David, I am checking to seeing if this is by design or a bug. With Windows 3.0 they recommend you using one single .RC file. If you have your resources spread out among many resources, create an empty rc file and #include the contributing .rc files. Thanks, Dave #: 16607 S7/Tools-Microsoft 04-Nov-92 11:15:48 Sb: No I/O fncs found Fm: Ken Lidster 71051,1323 To: sysop (X) When porting our product to Windows NT, we get unresolved external messages for the I/O functions (_lseek, _open, _write, etc.). The release notes indicate that the file mstools\lib\symbols.txt would list all of the symbols available in the libraries. We could only find the above functions listed as _llseek, _lopen, _lwrite, and so forth in kernel32.lib. As a test, we changed our modules to match the names in symbols.txt, but they still remain undefined (all forms). Where can we find these functions? Ken There is 1 Reply. #: 16733 S7/Tools-Microsoft 05-Nov-92 12:19:34 Sb: #16607-No I/O fncs found Fm: David Taniguchi [MS] 72350,2054 To: Ken Lidster 71051,1323 (X) Hi Ken, >When porting our product to Windows NT, we get unresolved external >messages for the I/O functions (_lseek, _open, _write, etc.). The release >notes indicate that the file mstools\lib\symbols.txt would list all of >the symbols available in the libraries. We could only find the above >functions listed as _llseek, _lopen, _lwrite, and so forth in kernel32. >As a test, we changed our modules to match the names in symbols.txt, but >they still remain undefined (all forms). I was able to use the _lopen() and _lread(). To do this, I modified the sample "simple". I simply added hFile = _lopen("simple.c", OF_READ) _lread(hFile, szTemp, 80). As you mentioned, these should be included in kernel32.lib. Are you using the October Release? When you are linking, do you see kernel32.lib on the link line? Is your LIB environment variable set to the proper location? Thanks, Dave #: 16653 S7/Tools-Microsoft 04-Nov-92 16:01:00 Sb: CVPACK breakpoint (?) Fm: Doug Olson [Microsoft] 72350,2635 To: Bruce Cowan 73650,32 (X) Bruce: Our developer who wrote CVPACK has been looking into this problem and has asked me for a sample. Could you upload a sample EXE file? This would be the EXE file that is left over after the breakpoint exception. To make sure you have the correct EXE file, you can type CVPACK PROGRAM.EXE at the command prompt. You should see the same breakpoint exception. Thanks for the help, Doug Olson Microsoft Developer Support There is 1 Reply. #: 16743 S7/Tools-Microsoft 05-Nov-92 13:04:24 Sb: #16653-CVPACK breakpoint (?) Fm: Bruce Cowan 73650,32 To: Doug Olson [Microsoft] 72350,2635 (X) Doug: I'm afraid I can't recreate that problem any more. Now when I try the same thing (compiling my entire application with /Zi) I get a "instruction at 0x00015da5 couldn't write the memory at 0x0058e1b0" or "instruction at 0x00015d12 couldn't write the memory at 0x00559000" (paraphrased). The crash is quite repeatable although the failure seems to move around a bit. Do you really want me to upload a 2.6MB file (compressed, the .EXE file is 6.1MB)? If so, I think you should pay! Bruce #: 16753 S7/Tools-Microsoft 05-Nov-92 14:39:14 Sb: MFC Linker Problem Fm: Bruce W. Johnson 73170,3076 To: all I am having linking trouble with MFC. Symmptoms can be replicated as follows: 1. Take the sample program hello.cpp. 2. Move the constructor CMainWindow::CMAinWindow() to a separate source module. 3. Add this new module to the list of objects in the makefile. 4. The make will fail in linking with the message that ??0CMainWindow@@QAG@XZ is an undefined external. What is the magic incantation to give the compiler/linker to keep it from choking on this? I am using the October beta. Thanks in advance. #: 16417 S7/Tools-Microsoft 02-Nov-92 18:08:11 Sb: Oct NT and C++ Debugging Fm: John H Soucie 71052,3516 To: SYSOP (X) Sysop, I have just recieved my October release of NT and the preliminary SDK. I have a C++ test application that compiles under the new version and runs beginning to end. The application is a short console application. I attempted to run the application through Windbg but found that while I now view C++ code, I still cannot view C++ class data. Whenever I attempt to expand a class listed in the Locals or Watch window by clicking on the "+" symbol, the debugger replies with: Error: error in OMF type information. With this error, it seems as though the Windbg expression evaluator is still unable to recognize C++ class information. Is this true or am I doing something wrong. The release notes state that the new Windbg supports C++, but I would have to take exception to this statement if C++ class data cannot be viewed. Additionally, if the problem is with Windbg, why isn't there a readme file somewhere that discusses its limitations? Any insight would be great. Thanks. John Soucie There is 1 Reply. #: 16760 S7/Tools-Microsoft 05-Nov-92 15:38:39 Sb: #16417-Oct NT and C++ Debugging Fm: Colin Stuart [Microsoft] 70744,25 To: John H Soucie 71052,3516 (X) John, We've reproduced the problem and have informed development. I'll post any new information I get. Thanks for your post. Colin #: 16761 S7/Tools-Microsoft 05-Nov-92 15:38:46 Sb: Bug in C++/C comp. Fm: Ken Joyner 75140,256 To: Sysop (X) /******************************************************************************* * The code below shows a bug in the compiler: * This code works with Borland, AT&T, and Zortech compilers. * * The compiler can't figure out what operator it should use, but * it should use the operator precedents to figure it out. * * This is a serious bug. What is your policy on bug fixes, do you make * patches? * * This bug shows up in the July and October Releases. * * Bryan Sera * * *******************************************************************************/ class Astr { public: Astr(char* CString); }; class AstrArrLtr { }; class AstrArr { public: AstrArr(unsigned int size); operator AstrArrLtr*(); operator Astr*(); Astr& operator[](unsigned int index); }; void main() { AstrArr array(5); array[0] = Astr("Zero"); } #: 16762 S7/Tools-Microsoft 05-Nov-92 15:39:51 Sb: another Bug in C++ comp. Fm: Ken Joyner 75140,256 To: Sysop (X) /******************************************************************************* * The code below shows a bug in the compiler: * This code works with Borland, AT&T, and Zortech compilers. * * This bug shows up in the July and October Releases. * * Bryan Sera * * *******************************************************************************/ class A { virtual int func() { return 0; }; }; typedef A thisisA; class B : public A { /* Since thisisA is a A, the function below should work */ /* the same as this->A::func(); */ virtual int func() { return this->thisisA::func(); }; }; #: 16720 S7/Tools-Microsoft 05-Nov-92 09:54:31 Sb: Resource Formats Fm: Samuel Feldman 70403,432 To: Steve Firebaugh [MS] 75430,412 (X) Steve -- Another question occurred to me after reading the resource formats document, and that is, what are these new resource types with predefined type numbers? RT_NEWRESOURCE is a flag 0x2000, and gets applied to other types to create RT_NEWBITMAP, RT_NEWMENU, and RT_NEWDIALOG. Can you tell me what these are? And doesn't Microsoft documentation say that if you want to use a user-defined type with an ID, make sure it's above 256, since numbers below 256 are reserved for Microsoft? Then wouldn't the use of 0x2000 create problems of ID conflicts? -- Samuel There is 1 Reply. #: 16763 S7/Tools-Microsoft 05-Nov-92 15:40:14 Sb: #16720-Resource Formats Fm: Steve Firebaugh [MS] 75430,412 To: Samuel Feldman 70403,432 (X) Samuel, First of all, thank you for the constructive feedback on resfmt which you posted in section 4. We are reviewing it for inclusion in the next release of that document. I will try to have answers to the questions which you raised there sometime tomorrow. I will post them in section 4. Secondly, I do not know what the RT_NEWRESOURCE flag is for. From everything that I've seen, it is not used. Please notice that I sort of "went out on a limb" to upload resfmt. It has not been through a formal doc. review process. Although this is being worked on, I don't think that anyone would have seen the information in that document before the next beta release. You may want to interpret the contents with this in mind. It is NOT an official document. Thanks for your interest and feedback. Steve Firebaugh #: 16328 S7/Tools-Microsoft 02-Nov-92 08:10:41 Sb: DLL/FP/Linker Fm: Thomas Nielsen 100065,2167 To: Dough Olson 10000,34 Doug, once again thanks for your help on our linking problem (msg. #15574, using floating point in a DLL), but another one has arisen. Since we are now using the _CRT_INIT function as entry-point in the DLL, our old LibMain function isn't getting called on DLL startup. The problem is, that we are defining a couple of window classes in the LibMain function, at the time the DLL is loaded, and since this function isn't being called anymore, the window classes aren't defined. Our question is: Is it possible to call the _CRT_INIT function from within our own LibMain, and still get the desired effect (no reference to a _main function), or do we have to define a special function, to be called on program (EXE) startup, which will define our window classes. Many thanks in advance, Thomas There are 2 Replies. #: 16423 S7/Tools-Microsoft 02-Nov-92 18:49:15 Sb: #16328-DLL/FP/Linker Fm: Doug Olson [Microsoft] 72350,2635 To: Thomas Nielsen 100065,2167 (X) Thomas: Interesting question, I am checking into it. Doug Olson Microsoft Developer Support #: 16775 S7/Tools-Microsoft 05-Nov-92 17:32:30 Sb: #16328-DLL/FP/Linker Fm: Doug Olson [Microsoft] 72350,2635 To: Thomas Nielsen 100065,2167 Thomas: It is possible to call the _CRT_INIT function from within your own entry point function. Below I have included a sample DLL entry point which calls the _CRT_INIT correctly and invokes LibMain(). _CRT_INIT must be called on PROCESS_ATTACH before any C run-time routines are called, and it must be called on PROCESS_DETACH after all run-time routines are called. Please note the two lines that are only for the October beta. The calling of _heap_init will not be necessary in the next release of the libraries. I hope that this helps, Doug Olson Microsoft Developer Support BOOL _stdcall DllEntryPoint(HANDLE hDll, DWORD dwReason, LPVOID lpReserved) { /* These two lines are only needed with the October Beta */ void __cdecl _heap_init(void); if ( dwReason == DLL_PROCESS_ATTACH ) _heap_init(); /* Initialize the C Run-Time *BEFORE* any user code initialization */ if ( dwReason == DLL_PROCESS_ATTACH || dwReason == DLL_THREAD_ATTACH ) if (!_CRT_INIT(hDll, dwReason, lpReserved)) return fFalse ; /* Call LibMain() on process initialization */ if ( dwReason == DLL_PROCESS_ATTACH ) LibMain(); /* This is a place for more user code DLL init/term */ /*Terminate the C Run-Time *AFTER* all user code termination */ if ( dwReason == DLL_PROCESS_DETACH || dwReason == DLL_THREAD_DETACH ) if (!_CRT_INIT(hDll, dwReason, lpReserved)) return fFalse ; return fTrue; } #: 16596 S7/Tools-Microsoft 04-Nov-92 10:38:23 Sb: CL386 Problems Fm: Chuck Atwood 76004,1602 To: sysop (X) Problems with CL386 compiler for NT. I am lumping them altogether since they seem to be related to the evaluation of macro variables during prototyping or compilation. I had the same problem prototyping which has already been reported: INTERNAL COMPILER ERROR in WINUSER.H at line 111. After commenting out lines 111 and 112 (wsprintf*) I was able to prototype most c files. However, the evaluation of macro expressions created other compiler errors. Problem evaluating #if expressions during prototyping (-Zg). We use the standard #if (WINVER >= 0x030a) contruct used in windows.h for a define CADVER. This generates a fatal error: "integer error evaluation expression". #if (CADVER >= 300) ... #endif Also simple tests fail, e.g.: #define DEBUGDB 0 #if DEBUGDB ... #endif CL386: Problem evaluation simple expressions. The following code fragment generates a "overflow in floating point constant arithmetic": #define CHORD_TOLERANCE 0.00015 double d; d = CHORD_TOLERANCE * 1440; CL386: -Zs compile switch not implemented. CL386: INTERNAL COMPILER ERROR when compiling YACC generated source file. Always fails in ?????.A which is a large switch statement. Chuck Atwood (Foresight Resources Corporation) There is 1 Reply. #: 16776 S7/Tools-Microsoft 05-Nov-92 17:32:38 Sb: #16596-CL386 Problems Fm: Doug Olson [Microsoft] 72350,2635 To: Chuck Atwood 76004,1602 (X) Chuck: Thanks for the problem report. I will report these problems to development and let you know the status. Thanks, Doug Olson Microsoft Developer Support #: 16418 S7/Tools-Microsoft 02-Nov-92 18:14:16 Sb: C++/NT Oct -> July Fm: Alex Bronstein 75070,2452 To: Colin Stuart [Microsoft] 70744,25 (X) How many of you C++ programmers are dropping Oct NT and going back to July NT ? Microsoft: in light of the C++ bugs you're seeing in the Oct NT release, would you recommend going back to the July NT release (for C++) ? Everybody (and apologies if this is a heresy here): is there a good Borland C++ solution for NT development available yet? Thanks, Alex There are 2 Replies. #: 16455 S7/Tools-Microsoft 03-Nov-92 08:48:05 Sb: #16418-C++/NT Oct -> July Fm: Bill Cohagan 74375,313 To: Alex Bronstein 75070,2452 (X) Alex- I'm afraid I'm about to drop back to the July release. My application won't compile under the Oct release. Unfortunately that same app wouldn't *link* correctly under the July release, but at least I figured out a workaround. Not so with the current compiler problem (fatal compiler error.) Although I would have liked an instant fix from MS on this forum there has been no fix/workaround forthcoming. In fairness, I think MS is responding as best they can under the circumstances, but I'm still left high and dry. I'm just hoping we don't have to wait another major release cycle to get fixes for these compiler problems. Surely MS is aware of how these problems are impacting developers' efforts. Bill #: 16510 S7/Tools-Microsoft 03-Nov-92 16:23:47 Sb: #16418-C++/NT Oct -> July Fm: Peter Handsman[Inmark] 70550,2570 To: Alex Bronstein 75070,2452 (X) Alex, Most of the bugs i've seen in the oct'92 c++ compiler seem to have workarounds, not that that's an escuse - but environment wise the october version is much better. Of course, their are 3-4 bug reports i'll be making as soon as I get them written up.... Peter There is 1 Reply. #: 16670 S7/Tools-Microsoft 04-Nov-92 18:37:38 Sb: #16510-C++/NT Oct -> July Fm: Doug Olson [Microsoft] 72350,2635 To: Peter Handsman[Inmark] 70550,2570 (X) Peter: Please report any bugs to us that you find in the October release. We are investigating the bugs that have been reported so far, and would like to hear about any others that are being encountered. We are working to find solutions to these problems. Sincerely, Doug Olson Microsoft Developer Support There is 1 Reply. #: 16701 S7/Tools-Microsoft 05-Nov-92 00:04:26 Sb: #16670-C++/NT Oct -> July Fm: Peter Handsman[Inmark] 70550,2570 To: Doug Olson [Microsoft] 72350,2635 (X) Doug, I didn't mean to imply that the tools are not being worked on, and considering what the tools looked like several months ago they are improving. It's just disapointing to see things which seemed to work better in prior versions. Esp the C++ parser in cl386. Here is code for a bug where under some circumstances the constructor syntax of "classname(paramater)" is not accepted and "(classname)paramater" is accepted. class Id { public: Id(int) {} }; class Other { public: Other(Id); }; void func(int xx) { Other *off; Other ot(Id(1)); // Compiles Other ot2((Id)1); // Compiles Other ot3(Id(xx)); // Compiles Other ot4((Id)xx); // Compiles off=new Other((Id)xx); // Compiles off=new Other(Id(xx)); // Syntax error as follows: #: 16777 S7/Tools-Microsoft 05-Nov-92 17:32:41 Sb: #16701-C++/NT Oct -> July Fm: Doug Olson [Microsoft] 72350,2635 To: Peter Handsman[Inmark] 70550,2570 (X) Peter: Thank you for reporting these problems. I am going to forward this information on to development and also look into your questions. Thanks, Doug Olson Microsoft Developer Support #: 16789 S7/Tools-Microsoft 05-Nov-92 19:09:33 Sb: NMAKE problems Fm: Chuck Atwood 76004,1602 To: ssyop NMAKE doesn't check dependencies correctly for directories. The following makefile works using NMAKE from C7 with DOS 5.0 but will display "'CHUCK' is up-to-date" in the MSDOS window under NT. I am using the FAT file system. Make a directory CHUCK and type: NMAKE -f NTBAD.MAK. # NTBAD.MAK # author: cla DIRS = CHUCK $(DIRS): cd $(*) copy ..\ntbad.mak . cd.. # [EndOfMake] P.S. NMAKE needs !ELSE which is available in 3.1/C700 SDK. Chuck Atwood (Foresight Resources Corporation) #: 16606 S7/Tools-Microsoft 04-Nov-92 11:09:02 Sb: Backup/Restore problem Fm: Apostolos Karmirantzos 75070,2354 To: all I was trying to copy a few files from one NT machine to another. The files were some executables and a few archive files. Everything seemed to work fine, but when I tried to execute the file, I got the error "Failed to initialize properly, address 0xc000000005" or something like that. When I looked at the files, they seemed to have the correct size. The way I finally managed to copy the files was to use backup and restore from simple DOS (not DOS under NT). So I assume that backup/restore under NT has some problems. Apostolos There is 1 Reply. #: 16794 S7/Tools-Microsoft 05-Nov-92 19:51:34 Sb: #16606-Backup/Restore problem Fm: David Taniguchi [MS] 72350,2054 To: Apostolos Karmirantzos 75070,2354 (X) Hi Apostolos, Thanks for the problem report. Please post this in section 7 (Utilities/ Applets) of Forum WinNT. They are more familiar with the system utilities and non-sdk issues. Thanks, Dave #: 16594 S7/Tools-Microsoft 04-Nov-92 10:36:57 Sb: NTSD and SECOND CHANCE Fm: Mike Payne 71062,3342 To: Anyone Hi, Perhaps someone can help me. I want to turn off the way the ntsd comes up on "SECOND CHANCE" after my application crashes. THis is so that I can run an automated series of tests, which note what crashes anyway. Thanks, Chris Hall There is 1 Reply. #: 16795 S7/Tools-Microsoft 05-Nov-92 19:51:40 Sb: #16594-NTSD and SECOND CHANCE Fm: David Taniguchi [MS] 72350,2054 To: Mike Payne 71062,3342 Hi Mike, One way you could do this is put a try-except around your whole program. And then the exception handler does nothing. If I find a better way to do this, i'll post here. Thanks, Dave #: 16809 S7/Tools-Microsoft 06-Nov-92 04:49:31 Sb: Test for NT Fm: Dan Sullivan 76327,1534 To: SYSOP (X) I am interested in when Test for NT will be available? Any predictions? Thanks in advance Dan #: 16814 S7/Tools-Microsoft 06-Nov-92 08:35:33 Sb: Windbg crashes Fm: David Edge 75170,1461 To: all I am using Windbg to debug a C++ application. So far, I have had reasonably good success. I am having two problems though. The first problem is expanding classes in the local variable window. Many times, when I try to do this, I get the application error message: "The instruction at 0x400133F5 referenced memory at 0x00000000. The memory could not be "read"." This is very frustrating to get deep into debugging, need to look at the values in a class and know that if I click on it to expand it it may be the end of the debug session. The second problem is that it seems like the local variable window is not always updated properly. Sometimes when I reach a breakpoint, the values do not look correct so I go over to the command window and look at them there and the values are different. Closing and reopening the window causes the values to be correct again. Overall, I really like Windbg. It took a while to get used to it because I really liked two monitor debugging in Win 3.1 with an 8514 and a 50 line VGA codeview screen. The debugger seems very sluggish. Single stepping takes a long time and startup is pretty slow. It also takes a long time to restore all of the windows on startup. Are there plans to improve its performance? #: 16384 S7/Tools-Microsoft 02-Nov-92 15:21:01 Sb: link/problem with errno Fm: Apostolos Karmirantzos 75070,2354 To: all When I try to link my program using the predefined $(guilibsmt) symbol, link reports that _errno is undefined. When I link with c:\mstools\lib\*.lib, the link works fine. Unfoprtunatly, if you link with all the libraries, there is a conflict between some libraries about errno, so errno isn't updated correctly. My program uses errno to decide if some file operations are succesfull or not, and if not takes the neccesary corrective actions. So there isn't the option of not using errno. Which are the correct libraries to link with so errno is defined? (My program is a windows multithread application) Thanks in advance Lakis There are 2 Replies. #: 16523 S7/Tools-Microsoft 03-Nov-92 17:28:52 Sb: #16384-link/problem with errno Fm: Doug Olson [Microsoft] 72350,2635 To: Apostolos Karmirantzos 75070,2354 (X) Apostolos: I am checking into this problem and will get back with you as soon as possible. Sincerely, Doug Olson Microsoft Developer Support #: 16661 S7/Tools-Microsoft 04-Nov-92 16:55:16 Sb: #16384-link/problem with errno Fm: Doug Olson [Microsoft] 72350,2635 To: Apostolos Karmirantzos 75070,2354 (X) Apostolos: Please make sure you are doing the following: 1. Using $(guilibsmt) on your link line 2. Using $(cvarsmt) instead of $(cvars) on your compile line 3. Do not extern errno in your code With the above configuration I was able to successfully link. Please let me know what happens. Thanks, Doug Olson Microsoft Developer Support There is 1 Reply. #: 16819 S7/Tools-Microsoft 06-Nov-92 10:22:25 Sb: #16661-link/problem with errno Fm: Apostolos Karmirantzos 75070,2354 To: Doug Olson [Microsoft] 72350,2635 (X) Thanks for the info. I managed to compile/link with no problem about errno. The problem was that I was using $(cvars) indetad of $(cvarsmt). Next time I should read the manual first. Thanks again. Lakis #: 16408 S7/Tools-Microsoft 02-Nov-92 16:42:05 Sb: Efficient try/except ?? Fm: Bill Cohagan 74375,313 To: sysop (X) How efficient is it to use try/except? I'm considering using a write protected page as a boundary for some memory being managed by my app. This way I can just use the write protect violation as a signal that I need to increase the memory size. The alternative is to do a threshold compare each time I allocate a chunk. Due to other considerations it appears that I'd have to wrap the individual assignment operation with the try/except; e.g., try{ i += len; foo[i] = 0; } except ( handle write protect fault ) { ... } The idea is that I can unprotect that page and let the operation proceed in *most* cases. So is the above likely to be more efficient than: i += len; if (i > threshold) ...handle overflow... else foo[i] = 0; My guess is that the threshold check is quicker 'cause the try has got to do some setup in order to arrange to catch the protection violation, right? Bill There are 2 Replies. #: 16524 S7/Tools-Microsoft 03-Nov-92 17:28:57 Sb: #16408-Efficient try/except ?? Fm: Doug Olson [Microsoft] 72350,2635 To: Bill Cohagan 74375,313 (X) Bill: I have someone looking into this issue. I will get back with you as soon as possible. Sincerely, Doug Olson Microsoft Developer Support #: 16824 S7/Tools-Microsoft 06-Nov-92 11:06:39 Sb: #16408-Efficient try/except ?? Fm: Doug Olson [Microsoft] 72350,2635 To: Bill Cohagan 74375,313 (X) Bill: This has been an interesting question. >>My guess is that the threshold check is quicker 'cause the try has got to do some setup in order to arrange to catch the protection violation, right? I think you are right, the threshold check would be quicker. To use try/except more efficiently, I suggest putting the try/except around the loop. This method would be faster than the threshold check. Sincerely, Doug Olson Microsoft Developer Support #: 16626 S7/Tools-Microsoft 04-Nov-92 13:33:18 Sb: help Fm: Rob Tessier 76666,3633 To: all I am in the process of porting and .exe and several supporting .dlls to the Oct release of Win32/NT. Supposedly, only one set of sources is needed to support both win32 and win3.x but as I get more into it, it seems almost unavoidable to have separate sources. How do I combine the differences in the makefiles and .def file? Win32 doesn't require the stack and heapsize fields in the def file, how do I exclude them when building for win32? Also, it appears that the hardbound docs I got with the July release is just a copy of the win3.x/C7.0 docs with NT stuck in place of 3.x/c7.0. Some of the NMAKE options described are not supported as the docs say they are plus the cl386 compiler doc still talks segmentation and memory models. Is there some accurate docs for the SDK tools so I can reference legit info? I tried using !ifdef $(win32) in my make script to isolate win32 builds from win16 builds and nmake keeps complaining about that line. Rob There is 1 Reply. #: 16793 S7/Tools-Microsoft 05-Nov-92 19:51:29 Sb: #16626-help Fm: David Taniguchi [MS] 72350,2054 To: Rob Tessier 76666,3633 (X) Hi Rob, Try looking at the Generic sample (in mstools\samples\generic). This provides a good example of how to isolate the Win16/Dos differences. What it boils down to is the following statements: !IF "$(CPU)" != "" OS=NT ENV=WIN32 !ELSE OS=DOS ENV=WIN16 !ENDIF !include <$(OS)$(ENV).MAK> In NT environments, the NTWIN32.MAK environment specific details can be defined (like the libraries, (guilibs, conlibs), etc.) In DOS, DOSWIN16.MAK will be included. You also mentioned the module definition file. In the Generic example makefile, they handle this in the following way: !IFDEF CPUTYPE # This is for Windows NT $(cvtobj) $(cvtdebug) *.obj link -link $(linkdebug) $(guiflags) $(proj).obj .... !ENDIF !IFNDEF CPUTYPE # This is for Windows DOS $(link) $(guiflags) ... $(proj).def rc ... !ENDIF As for documentation, try looking at the Compiler Tools Help (In general the Win Help files are more recent than the hard copy docs). (To start it, type winhlp32 :\mstoools\bin\tools.hlp) Hope this helps, Dave You mentioned the There is 1 Reply. #: 16836 S7/Tools-Microsoft 06-Nov-92 11:45:43 Sb: #16793-help Fm: Rob Tessier 76666,3633 To: David Taniguchi [MS] 72350,2054 (X) Your suggestions were very helpfull, Thanks. But one thing I didn't deduce from the GENERIC samples is what should the output filenames be named if I intend to support both 16 and 32 bit Windows environments? Because I'm guessing the import libs, and dlls I built under NT are not compatible with Win16, I appended 'w32' to the filename to keep them separate from their 16 bit counterparts. For instance, my dirs have: [c:\lvwindev\lib] [c:\lvwindev\exe] util.lib utilw32.lib util.dll utilw32.dll graph.lib graphw32.lib graph.dll graphw32.dll Is this recommended? I found if I tried to use the import lib built from Win3.0, the NT linker complains. I haven't tried the reverse yet to see if I try to link under Win3.0 with the import libs generated from NT. Could you please clarify the recommended procedure for developing from the same source code, yet supporting Win3.x and Win32? Rob T. #: 16838 S7/Tools-Microsoft 06-Nov-92 13:25:52 Sb: Windbg and "mdb" Fm: Jake Kirk 76207,1403 To: sysop (X) I can seem to be able to "mdb" or memory dump bytes under the debugger. this feature is available under codeview. how would one do this under NT? for example I want to dump bytes starting at the address of a variable for a length of 100 bytes. ? mdb my_var L 100 < cr > would be the way to do it under codeview. //Jake. #: 16839 S7/Tools-Microsoft 06-Nov-92 13:28:55 Sb: NT Windbg Fm: Chester L. Kloss 72311,743 To: sysop (X) I'm attempting to debug an application which uses the MFC library. When I enter windbg with my app it (windbg) can't find the source files for MFC because (I guess) they are in the MFC\src directory. How do I solve this. I thought maybe I have to copy all the sources for MFC into my codes directory. Or, do I have to recompile the MFC library in with some "new" targets, models etc. ? Also is quickwatch disabled in october release of NT? Finally, how do I get windbg when it starts up an app to automatically put the current program location cursor on winmain, I usually execute trace into to get me there? Thanks #: 16848 S7/Tools-Microsoft 06-Nov-92 14:44:03 Sb: MIPS Link Fm: Howard Myers 76711,462 To: Microsoft I'm trying to port our app. to the MIPS platform. I've ported several small apps. successfully, but in porting our largest app., I can't link it. Every time I try, after all the link commands have been issued, including the list of files, I very quickly get the following messages: LINK() : error 0102: Out of memory Extended Error: No such file or directory NMAKE : fatal error U1077: ....... I am linking approximately 90 files that comprise 8.5 MBytes of disk space. These have been converted using MIP2COFF without the -c option. Initially, I tried using the -c option with the same results, but the files ate 22 MBytes, so I tried it without. I have 32 Meg of RAM, 100 Meg of free disk space, and 90 Meg allocated for the System Paging File. Needless to say, I can't understand why I'm running out of memory! I'm assuming that the Out of memory error, since it comes first, is the legitimate error, and that the no such file or directory is a side effect. I'm supposed to have this up to demo in the Microsoft 3rd party booth at COMDEX, so admittedly, I'M DESPARATE! I'm posting this in both the Tools and MIPS sections, because I can't afford the delay to re-post it if I put it in the wrong place. (I don't want to waste your time by having 2 people simultaneously tracking the same problem. Likewise, please don't drop it, assuming the other section's taking care of it.) Thanks!!!!!! #: 16640 S7/Tools-Microsoft 04-Nov-92 15:08:32 Sb: KD and NTSD setup Fm: Tim Stevenson [NCD PC-X] 71041,1465 To: Sysop (X) I have just completed setting up my second NT system with KD and have verified that the two systems talk to one another via the serial ports. I start KD on the "debugger" machine and get a KD: waiting to connect message On the target machine I start NTSD with ntsd -d hooks.exe I immediately get an Application Error dialog from NTSD.EXE stating the following: The exception "Breakpoint A breakpoint has been reached in the source program. "(0x80000003) occured in the application at location 0x6010a3f4 When I click on Cancel I am in NTSD.EXE at ntdll!_DebugService+0x14. I believe I have followed what instructions I have been able to find for the setup of KD and NTSD but I have not been able to get it to work. If you are asking why I am going to all this trouble, its because I need to be able to debug system hooks and NTSD is not able to debug code that hooks keyboard and mouse events. There is 1 Reply. #: 16774 S7/Tools-Microsoft 05-Nov-92 17:08:26 Sb: #16640-KD and NTSD setup Fm: Petrus Wong [Microsoft] 70743,3355 To: Tim Stevenson [NCD PC-X] 71041,1465 Hello Tim: Is your target machine sending debug messages to the i386kd (or Mipskd) when it boots? You may run the i386kd on the debugger with the -v option to put it in the verbose mode. (Do i386kd -? for detail.) If the -v option is used, then you should see screenful of debug messages as the target machine boots. If you don't see the debug message as the target boots, then you should check to make sure if the target machine is running in debug mode. On x86, the boot.ini file should not specify the /NODEBUG on the os line. On MIPS, check the environment variable on the bottom of your blue screen before booting. If the target is indeed in debug mode, check the serial line. On x86, the target graps the highest com port for debugging. You will see the message "debugger using com2" displayed on the top of the blue screen if your machine has two com port. This com port should be connected to com1 on the debugger (or to the port specified by the _nt_debug_port environment variable. Finally check to make sure you have the right cable and no hardware conflict on the com port. Hope this helps, petrus There is 1 Reply. #: 16852 S7/Tools-Microsoft 06-Nov-92 15:42:41 Sb: #16774-KD and NTSD setup Fm: Scott Alexander 76556,557 To: Petrus Wong [Microsoft] 70743,3355 (X) Petrus > Finally check to make sure you have the right cable and no hardware conflict on the com port. < Since I am about to start dual machine debugging code, what is the right cable? I heard this mentioned at the DDC also, but never found out if I needed a modem cable, a straight thru or a "null modem" cable. Is this documented somewhere ? (I do not have the 3.1 ddk, only the NT ddk, which has no docs yet) Thanks #: 16714 S7/Tools-Microsoft 05-Nov-92 08:19:34 Sb: New NT RC? Fm: Mike Bailey 70760,112 To: Tools-Microsoft Will the released version of the NT SDK still use the Windows 3.X resource compiler along with the CVTRES.EXE program? We have a situation where we replace the resource in a executable. Will we still be able to do this under NT or will we have to re-link all of the objs again? There is 1 Reply. #: 16859 S7/Tools-Microsoft 06-Nov-92 16:10:34 Sb: #16714-New NT RC? Fm: Steve Firebaugh [MS] 75430,412 To: Mike Bailey 70760,112 Mike Bailey, At the moment, there is no utility to modify the resources built into an executable without relinking. There may be a tool, by the final release of the SDK, but the decision has not been made yet one way or the other. Please notice that there are three new API's which would make writing your own utility pretty straightforward: BeginUpdateResource, UpdateResource, and EndUpdateResource. Incidentally, the resource compiler included on the Windows/NT SDK is different from the 3.1 resource compiler. For example, the new RC writes out strings in unicode. The CVTRES program is used to convert the RES file to an object module format compatible with what the linker expects. I believe that CVTRES will continue to exist through the first release, but there are some people who would like to incorporate its function into either the linker or into RC. Steve Firebaugh #: 16813 S7/Tools-Microsoft 06-Nov-92 08:30:30 Sb: New NT RC.EXE Fm: Mike Bailey 70760,112 To: sysop (X) Will the released version of the Win32 SDK contain a Win32 version of the resource compiler, instead of the Win 3.X version and the CVTRES.EXE program? We need to replace a resource after the executable has been linked. Will there be a way of doing this in the future without re-linking the entire program? There is 1 Reply. #: 16860 S7/Tools-Microsoft 06-Nov-92 16:10:38 Sb: #16813-New NT RC.EXE Fm: Steve Firebaugh [MS] 75430,412 To: Mike Bailey 70760,112 Mike Bailey, I just answered these questions in reponse to your message # 16714. In general, if you do not hear a reponse within two working days, it is a good idea to re-post the question. If it has been less than two working days, we are probably still working on the answer. Thanks, Steve Firebaugh #: 16865 S7/Tools-Microsoft 06-Nov-92 16:55:58 Sb: MFC & CRTDLL Fm: dan white 70324,3147 To: sysop (X) I am unsure how to compile MFC for use with CRTDLL.DLL. If I compile with _DLL defined, MFC will think its in a DLL, right? Also, MFC uses routines like __doserrno which bon't seem to be in in CRTDLL.LIB. Which compile switches am I supposed to use? Also, how do I view TRACE output under NT? Is it supposed to come up in Windbg (it doesn't seem to)? Thanks, Dan #: 16431 S7/Tools-Microsoft 02-Nov-92 20:03:44 Sb: MFC/DLL Fm: Jim Conger 73220,324 To: ALL I'm having trouble converting a DLL based on the MFC classes from Windows 3.1 to Windows NT. Although both the DLL and EXE compile fine, the linker can't find the exported function from the DLL. I'm using the following linker command line to create the DLL and export library: coff -link -DLL -base:0x1C000000 \ -subsystem:windows -entry:WinMainCRTStartup -machine:i386 \ -out:MYDLL.dll MYDLL.obj nafxcw.lib libc.lib ntdll.lib \ winspool.lib kernel32.lib user32.lib gdi32.lib comdlg32.lib \ version.lib lib -out:MYDLL.lib -def:MYDLL.def -machine:i386 (I'm suspicious that the entry point should be something other than WinMainCRTStartup. I tried LibMain, but that resulted in an unresolved external.) The real problem surfaces when I try to link the EXE that uses the DLL. The linker does not find the exported function in MYDLL.LIB. Here is the linker call for the EXE: coff -link -subsystem:windows -entry:WinMainCRTStartup \ -out:MYEXE.exe MYEXE.obj nafxcw.lib libc.lib ntdll.lib \ winspool.lib kernel32.lib user32.lib gdi32.lib comdlg32.lib \ version.lib MYDLL.lib I verified that the MYDLL.lib export library exists and contains the correct name for the exported function (declared _extern "C" to avoid name mangling). Clearly the linker is not finding that exported function name, but I cannot figure out why. Any ideas? Jim Conger There is 1 Reply. #: 16825 S7/Tools-Microsoft 06-Nov-92 11:06:54 Sb: #16431-MFC/DLL Fm: Doug Olson [Microsoft] 72350,2635 To: Jim Conger 73220,324 (X) Jim: It appears that my previous message did not get to you. In any case, I have a few questions regarding this problem. 1. What is the exported name in your MYDLL.LIB file? 2. Is your calling program C or C++? 3. How is the imported function prototyped in your calling program? I look forward to your answers, Doug Olson Microsoft Developer Support #: 16868 S7/Tools-Microsoft 06-Nov-92 17:36:37 Sb: MFC/DLL Fm: Jim Conger 73220,324 To: 72350,2635 (X) Thanks for getting back to me. Here are the answers to your questions: 1. The exported function name is StrRev(). The full function declaration: extern "C" BOOL FAR PASCAL _export StrRev (LPSTR lpSource, int nLong) ; 2. The calling program is a C++ program with a CPP extenstion. However, the function declaration extern "C" is visible to the calling program. 3. The imported function is prototyped as shown in 1. above. I think the point of your questions is that C++ name mangling could be the reason for the problem. I did make this error initially, but caught it when doing the Windows 3.1 development. The DLL and calling program work fine when compiled using MS C++ 7.0 and the MFC classes for under Windows 3.1. The programs also compile fine under Windows NT, but the function StrRev() is not found by the calling program during the linking step. Running DEBUG on the export library shows that the StrRev() function is listed in the export library without name mangling. I lose the trail at this point. #: 16591 S7/Tools-Microsoft 04-Nov-92 10:34:32 Sb: RC Anomaly Fm: David Taniguchi [MS] 72350,2054 To: David Taniguchi [MS] 72350,2054 (X) Hi Samuel, We just uploaded RESFMT.ZIP to Lib 4. This contains a description for the resource formats. In this document, it states 0x8000 (as you originally guessed it should be) as the last entry. Thanks, Dave There is 1 Reply. #: 16869 S7/Tools-Microsoft 06-Nov-92 17:49:10 Sb: #16591-RC Anomaly Fm: Samuel Feldman 70403,432 To: David Taniguchi [MS] 72350,2054 David, > We just uploaded RESFMT.ZIP to Lib 4. This contains a description for the > resource formats. In this document, it states 0x8000 (as you originally > guessed it should be) as the last entry. [ In reference to Accelerator table items ] Yes, I see that is what the document says. However, this is NOT what RC does. This is either a bug in the doc, or a bug in RC, your choice. Could you please let me know which it is? Thanks a lot -- Samuel #: 16871 S7/Tools-Microsoft 06-Nov-92 18:11:00 Sb: WINDBG and My app Fm: Marc Singer 72130,2546 To: Sysop (X) I just went through the past two weeks worth of messages to find a solution for my problem. I did find that I needed to use -Zi instead of -Zd to get the right symbols. Unfortunately, after making this change, I can no longer get WINDBG to load my app. I get an error during a call to malloc which calls ntdll!_RtlExAllocateHeap (NTSD provided a stack backtrace) while WINDBG is loading my application. I checked that I have sufficient swap space, and I rebooted the machine to make sure there were no lingering sub-system problems. The error as reported is the instruction 0x6010fbad referenced memory at 0x002cea55 which could not be read. I am getting many more run-time applications errors than I did from the July release. I hope to use WINDBG to dump my symbols because we all know how fun using ntsd is. Is there any advice for getting WINDBG to load? My app is mostly C++. The disk image is 1.5 megabytes. What more can I give you all to help find a solution? I would like to use WINDBG before the show, but I will do whatever it takes to get the app running by then. Marc Singer -- Straylight Software #: 16322 S8/CPU-MIPS Specific 02-Nov-92 06:15:16 Sb: Porting to NT Fm: John Hardman 100023,2532 To: ALL Just a quickie to confirm how easy all you Microsoft people have made porting to NT. I have been working on a Win 3.1 project for the past 6 months and have been listening to all the stuff about message crackers etc. as you give the information out. The C source for the project is now around the 1.7 Mb mark. Having recently won a MIPs machine I thought I'd try rebuilding the project under NT. Starting from scratch with a partly configured NT machine but with none of my usual tools installed I ported the project successfully in under seven hours. About half of this was taken up rewriting my make file to work with the MIPs compiler and linker, maybe an hour fixing the few non-portable API calls that crept into the code, and the rest trying to edit large files using Notepad (which isn't quite up to it yet). To say that I'm pleased is an understatement. The system is up and running now. Well done guys. However .... at run time there are a few problems, but then NT isn't even in beta yet so that's not really surprising. The question is, do you want to hear about problems now or wait until beta starts ? I'll list a few things here, but as I haven't had time to check out the MIPs debugger yet it is possible that some of these are down to our source as opposed to NT ... (1) "unknown software exception" (0xc0000037) at location 0x6012f630. (2) Complete hang of system requiring push of reset button. (3) Error loading icon. (4) Error posting message to window. (5) Combo box sorting behaving in unexpected manner. Nice one guys ... John Hardman (UK). There is 1 Reply. #: 16341 S8/CPU-MIPS Specific 02-Nov-92 10:38:34 Sb: #16322-Porting to NT Fm: Petrus Wong [Microsoft] 70743,3355 To: John Hardman 100023,2532 (X) Hello John, We are glad to hear that you are pleased with the system and the ease of porting... >>However .... at run time there are a few problems, but then NT isn't even in beta yet so that's not really surprising. The question is, do you want to hear about problems now or wait until beta starts ?<< We have just released the Beta a few weeks ago. Yours is probably on the way. So, we really prefer the latter especially it could be bugs which have been fixed. thanks, petrus #: 16323 S8/CPU-MIPS Specific 02-Nov-92 06:15:33 Sb: Porting to NT - MIPS Fm: John Hardman 100023,2532 To: All The application that I am developing is designed to be portable across all supported Windows (3.1, Workgroups and NT) platforms (Intel, MIPs, DEC etc). In the error checking & reporting layer I use a hardware specific macro to get the return address off the stack. It currently looks like ... #ifdef INTEL_80x86 #define GetReturnAddress(ReturnSegment, ReturnOffset) \ _asm \ { \ _asm mov ax,[bp+2] \ _asm mov ReturnOffset,ax \ _asm mov dx,[bp+4] \ _asm mov ReturnSegment,dx \ } #else #define GetReturnAddress(ReturnOffset, ReturnSegment) \ { ReturnOffset = 0; ReturnSegment = 0; } #endif As I haven't had the chance to look into MIPs assembler yet can anyone tell me how to do the same thing for the MIPS R4000 (and if anyone knows also the other forthcoming NT supported machines such as DEC). ( yes, I realise that flat memory architectures don't have segments ) ( - it's only a label ... ) Many thanks in advance, John Hardman (UK). There is 1 Reply. #: 16349 S8/CPU-MIPS Specific 02-Nov-92 11:13:05 Sb: #16323-Porting to NT - MIPS Fm: Petrus Wong [Microsoft] 70743,3355 To: John Hardman 100023,2532 (X) Hello John, >>As I haven't had the chance to look into MIPs assembler yet can anyone tell me how to do the same thing for the MIPS R4000 (and if anyone knows also the other forthcoming NT supported machines such as DEC).<< The return address is stored in the ra register on the R4000. Only the Load and Store instructions can reference memory on the R4000. The SW instruction is defined as follows (Kane/Heninrich: MIPS RISC Architecture): Format: SW rt, offset(base) Description: The 16-bit offset is sign-extended and added to the contents of general register base to form a virtual address. The contents of the general regiater rt are stored at the memory location specified by the effective address. Hope this helps, petrus There is 1 Reply. #: 16444 S8/CPU-MIPS Specific 03-Nov-92 05:35:38 Sb: #16349-Porting to NT - MIPS Fm: John Hardman 100023,2532 To: Petrus Wong [Microsoft] 70743,3355 (X) Petrus, Many thanks for your help. Do you by any chance know how to do the same thing (getting return address) on NT supported machines other than Intel and MIPs. I am thinking about DEC etc. Regards, John Hardman. There is 1 Reply. #: 16623 S8/CPU-MIPS Specific 04-Nov-92 13:20:47 Sb: #16444-Porting to NT - MIPS Fm: Petrus Wong [Microsoft] 70743,3355 To: John Hardman 100023,2532 Hello John, No, don't have much information on other CPUs that runs NT like DEC Alpha yet. petrus #: 16532 S8/CPU-MIPS Specific 03-Nov-92 18:43:04 Sb: Debuggin.... Fm: Mike Ketcham[Mozart] 72417,3344 To: MS Folk... As I go diving deeper ... I am running into data alignment problems. This means I need more info on how to deal with this. Where might there be on-line info as to compiler switches to turn on code printing (ala. -Fc ) so that I can see structure / data alignment? How do you folks do it? Compile and run? (I have the october build running on a MIPSpc50) Also, where might I find general info on the MIPS like there is on CL? Is this planned for later? (I have very little docs now, but plan to get whatever might be available in bookstores). There is also supposed to be a "MS front end" to the MIPS compiler. Where? thx -=mike There are 2 Replies. #: 16749 S8/CPU-MIPS Specific 05-Nov-92 14:15:45 Sb: #16532-Debuggin.... Fm: Petrus Wong [Microsoft] 70743,3355 To: Mike Ketcham[Mozart] 72417,3344 Mike, >>As I go diving deeper ... I am running into data alignment problems. This means I need more info on how to deal with this. Where might there be on-line info as to compiler switches to turn on code printing (ala. -Fc ) so that I can see structure / data alignment? How do you folks do it? Compile and run? (I have the october build running on a MIPSpc50) Also, where might I find general info on the MIPS like there is on CL? Is this planned for later? (I have very little docs now, but plan to get whatever might be available in bookstores). Use the -S compiler switch. This switch is for compiling the specified source programs and leave the symbolic assembly language output in corresponding files suffixed with '.s'. There is no online compiler info at the moment. We are looking into this. There are also the two MIPS compiler guides that I know of, the MIPS RISCompiler and C Programmer's Guide and the MIPS RISCompiler Porting Guide published by the MIPS Computer Systems. >>There is also supposed to be a "MS front end" to the MIPS compiler. Where? It is in the /mstools/bin directory, cl.exe. thanks, petrus #: 16764 S8/CPU-MIPS Specific 05-Nov-92 15:41:41 Sb: #16532-Debuggin.... Fm: Petrus Wong [Microsoft] 70743,3355 To: Mike Ketcham[Mozart] 72417,3344 Mike, >>There is also supposed to be a "MS front end" to the MIPS compiler. Where? Correction for the location of this. It is on \mstools\bin\msmips directory. petrus #: 16849 S8/CPU-MIPS Specific 06-Nov-92 14:44:42 Sb: MIPS Link Fm: Howard Myers 76711,462 To: Microsoft I'm trying to port our app. to the MIPS platform. I've ported several small apps. successfully, but in porting our largest app., I can't link it. Every time I try, after all the link commands have been issued, including the list of files, I very quickly get the following messages: LINK() : error 0102: Out of memory Extended Error: No such file or directory NMAKE : fatal error U1077: ....... I am linking approximately 90 files that comprise 8.5 MBytes of disk space. These have been converted using MIP2COFF without the -c option. Initially, I tried using the -c option with the same results, but the files ate 22 MBytes, so I tried it without. I have 32 Meg of RAM, 100 Meg of free disk space, and 90 Meg allocated for the System Paging File. Needless to say, I can't understand why I'm running out of memory! I'm assuming that the Out of memory error, since it comes first, is the legitimate error, and that the no such file or directory is a side effect. I'm supposed to have this up to demo in the Microsoft 3rd party booth at COMDEX, so admittedly, I'M DESPARATE! I'm posting this in both the Tools and MIPS sections, because I can't afford the delay to re-post it if I put it in the wrong place. (I don't want to waste your time by having 2 people simultaneously tracking the same problem. Likewise, please don't drop it, assuming the other section's taking care of it.) Thanks!!!!!! #: 16473 S9/CPU-x86 Specific 03-Nov-92 10:49:35 Sb: Memory Mapped Hardware Fm: Fred Meyer 71510,2301 To: Sysop (X) Hi! The Sysop over in WINNT sent me here in the hope that you may be able to answer our question. Here goes: We have built a number of AT bus boards that communicate via memory that is mapped into the high memory area. This works with DOS, and also with DOS extenders, (since, at bottom, they also use DOS). We are now preparing to build another board. Ultimately, this board will be part of a system running under WINNT in 32 bit mode. Here's the question: What happens to High Memory when DOS disappears and we are dealing with a straight NT kernal that has no 640k limitation. Will memory mapped boards still work? Will NT maintain the old 'high memory' architecture and address ranges for compatibility? If not, what is the officially sanctioned method for communicating with a memory-mapped board under NT? Hope you can help. We haven't been able to find anything in the documentation. --Fred There is 1 Reply. #: 16506 S9/CPU-x86 Specific 03-Nov-92 15:51:16 Sb: #16473-Memory Mapped Hardware Fm: David Taniguchi [MS] 72350,2054 To: Fred Meyer 71510,2301 Hi Fred, I assume by a "memory-mapped" board you mean a 16-bit board which has a page frame (typically between the 640-1meg area) and commonly used with XT systems? Some of these boards you can convert to map into extended memory. But, I don't believe there is an officially sanctioned method for communicating with these types of boards. Thanks, Dave #: 16588 S9/CPU-x86 Specific 04-Nov-92 10:14:38 Sb: Fatal system error Fm: Robert Tonsing 70740,3541 To: Sysop (X) Help! I had Win NT working fine on a 486 machine with 8 meg of memory, in a NTFS disk partition. This morning, when I try to boot NT, I get: *** Fatal System Error: 0x00000071 *** Session manager terminated. What is error 0x71? Where can I find error message meanings for future reference? Thanx, Robert There is 1 Reply. #: 16660 S9/CPU-x86 Specific 04-Nov-92 16:53:39 Sb: #16588-Fatal system error Fm: David Taniguchi [MS] 72350,2054 To: Robert Tonsing 70740,3541 (X) Hi Robert, Please post this in Section 3 of forum WINNT. They are more experienced with these issues. Thanks, Dave #: 16437 S9/CPU-x86 Specific 02-Nov-92 22:48:57 Sb: Oct NT & DTC3290HD Fm: Balaji - SynOptics 76330,1047 To: ALL Hi, I am using DTC3290HD EISA SCSI Harddisk Controller. October release of windows nt doesnot seem to support this card. Is there any work being done in this area? Till a driver is available, will I be able to install NT over the network and then boot from the floppy drive? When I tried it, it gives errors about copying 'setupldr', 'disk1', 'hosts', but copies the other files. What is special about these files? I guess, if I can get over this problem, then I should be able to install NT on this system. Any pointers would be appreciated. -Balaji (pbalaji@synoptics.com( There is 1 Reply. #: 16505 S9/CPU-x86 Specific 03-Nov-92 15:38:49 Sb: #16437-Oct NT & DTC3290HD Fm: David Taniguchi [MS] 72350,2054 To: Balaji - SynOptics 76330,1047 (X) Hi Balaji, Please post this in Section 3 (Windows NT setup) of Forum WINNT. They are more familiar with these issues. You should not be able to run NT if you have an unsupported hard disk controller. >> Is there any work being done in this area? You might check with the manufacturers of the card to see if drivers are being written. Thanks, Dave There is 1 Reply. #: 16721 S9/CPU-x86 Specific 05-Nov-92 09:56:37 Sb: #16505-Oct NT & DTC3290HD Fm: Balaji - SynOptics 76330,1047 To: David Taniguchi [MS] 72350,2054 (X) thanks. will post in Section 3 of WINNT. FYI, I WAS ABLE to complete the Text Portion of Setup, but when it rebooted to run the graphical portion, it displayed, I/O Initialization error... System Error : 0x00069 -Balaji. #: 16815 S9/CPU-x86 Specific 06-Nov-92 08:46:24 Sb: Apaptec Trantor & Banyan Fm: Richard Bono 73060,1573 To: sysop (X) Hi, I have some support questions. I need an answer that I can rely on. If this is not the proper forum for these questions, please let me know. If I have to, I can use our ONLINE account. 1) I need to know if there is a driver available that will support the Adaptec 6260 and 6360 SCSI chip. I am told that drivers that work with the Adaptec 1520 card will work with this. The October (340 release does not seem to have support for this (only the 1540/1740 series) Who should I talk to about getting any possible driver, and being assured that one will be available. 2) I also need information on the Trantor SCSI adapter card. Must I contact Trantor on this? 3) We use Banyan VINES, is there any beta version of something so that we can evaluate this support? Thanks, Rich #: 16419 S9/CPU-x86 Specific 02-Nov-92 18:14:22 Sb: DDK Fm: Alex Bronstein 75070,2452 To: David Taniguchi [MS] 72350,2054 (X) What is the difference between the DDK hard-copy documentation and the SDK hard-copy documentation? (In other words, if I've purchased the full-blown SDK w/ hard-copy doc, is it worth it to order the full-blown DDK doc ?) Thanks, Alex There is 1 Reply. #: 16508 S9/CPU-x86 Specific 03-Nov-92 16:07:10 Sb: #16419-DDK Fm: David Taniguchi [MS] 72350,2054 To: Alex Bronstein 75070,2452 (X) Hi Alex, It's your judgement call. If you order printed docs you get hardcopy of the NDIS spec, the DDK API's etc. All of this is on the CD (you'd have to print it out yourself). Also, you get Mary Hunton's (sp?) Kernel Mode Device Driver book (which I don't believe is on the CD). I understand this book is like a "Inside Windows NT" book in that it doesn't give you details on the API, but more of an explanation of the components, and how the system works. The SDK printed documentation will give you API specific to the SDK. (ie The Programer's Reference, Tools manual, API References, Application Design Guide, etc) I believe everything in the Win32 SDK is given on CD. The difference between the DDK and the SDK is the content. The DDK is for writing device drivers which interface with hardware and the SDK is for writing application programs. Let me know if you need more information. Thanks, Dave There is 1 Reply. #: 16756 S9/CPU-x86 Specific 05-Nov-92 15:06:25 Sb: #16508-DDK Fm: Alex Bronstein 75070,2452 To: David Taniguchi [MS] 72350,2054 (X) Thanks much David. About additional info, could you give me pricing on the DDK w/ and w/o printed doc, and any discount for people buying the SDK and DDK simultaneously? Alex There are 2 Replies. #: 16771 S9/CPU-x86 Specific 05-Nov-92 17:05:04 Sb: #16756-DDK Fm: David Taniguchi [MS] 72350,2054 To: Alex Bronstein 75070,2452 (X) Hi Alex, There is no discount for buying both. If pricing is a problem, you might consider the CD-Only options. Personally, I rarely use the hard-printed docs for the SDK. I always have Winhelp running, and just type in the API name if I need to see the parameters and descriptions. Also, all the overviews are accessible through winhelp also.. Thanks, Dave #: 16772 S9/CPU-x86 Specific 05-Nov-92 17:05:12 Sb: #16756-DDK Fm: David Taniguchi [MS] 72350,2054 To: Alex Bronstein 75070,2452 (X) Whoops, pardon me. I forgot pricing. The SDK is $69 w/o docs, $399 w/docs. The SDK docs include the Programmer's References (which are online), C and C++ Manuals (API's online), RPC manual (API's online), Tools Manual (Most have online help), Application design guide, and other misc. manuals. The Preliminary DDK will be available before the end of November. The cost is as follows: DDK CD-ROM only $69 DDK CD-ROM with hard-copy documentation $399 This pricing includes updates and the final version of the DDK. The documentation will be included on the CD in both Postscript and Write format, and all the device driver functions will be documented in Winhelp. If you are in the US and currently own the Win32SDK, you will be receiving a mailing on the DDK in mid-November which will include 4-page product datasheet and an order coupon. We encourage you to order via this coupon since we are unable to keep up with extremely high call volumes the past few weeks. You will receive this coupon in plenty of time before the product ships. International customers should contact their local MS office for ordering, pricing and availibility information. One other important change which you should be aware of. Everyone who purchases the hard-copy documentation for the DDK will receive the new book "Inside Windows NT" from MS Press. This book will not be on the CD. There is 1 Reply. #: 16832 S9/CPU-x86 Specific 06-Nov-92 11:22:34 Sb: #16772-DDK Fm: Alex Bronstein 75070,2452 To: David Taniguchi [MS] 72350,2054 (X) Thanks David, One more question (since I am $$-sensitive): what is the rule concerning my reselling a SDK or DDK CD-ROM (without keeping any copies of course, neither disk nor paper)? Say now that I have the October NT SDK, I'd like to sell the July NT SDK to someone, am I allowed to do that (just like I can resell old books)? Thanks, Alex #: 16346 S10/Porting from OS/2 02-Nov-92 10:57:42 Sb: #16266-Bitmaps on menus Fm: John Hall [MS SDE] 70750,2341 To: BENJ FAYLE 71151,1603 (X) Use a static control for ES_READONLY. yes, I know you have to fiddle with the resource file. Of all the tools, I am most embarrassed about resconv.zip. It is on my list to improve but not at the top. #: 16373 S10/Porting from OS/2 02-Nov-92 14:16:14 Sb: #16266-Bitmaps on menus Fm: Nancy Cluts - Microsoft 70744,20 To: BENJ FAYLE 71151,1603 Hi Benj. <> Unfortunately, I have not found a way to do this other than the approach that Petzold uses. <> A new flag, ES_READONLY, has been added for Windows 3.1. Another way to do this is covered in a Knowledge Base article (which I will append to the end of this message for you). <> John is a developer (the one who wrote the OS/2 to Win porting tools) and I am with Developer Support. It is my responsibility to get questions answered in this forum (It is also my good fortune that John jumps in with answers). Nancy - Microsoft Win 32 Developer Support INF: Implementing a Read-Only Edit Control In Windows [P_WinSDK] ID: Q80946 CREATED: 20-FEB-1992 MODIFIED: 20-MAY-1992 3.00 3.10 WINDOWS ENDUSER | softlib ROEDIT 3.x Summary: In some situations, an application may contain text that is displayed for the user to read, which is undesirable for the user to modify. The application can use a static control to contain this text if the message is short. However, for larger amounts of text, which require scrolling to display all the text in the allotted area, something closer to an edit control is required. This article, and its associated sample code, details how to make an edit control "readonly." ROEDIT is a file in the Software/Data Library that demonstrates the techniques discussed in this article. ROEDIT can be found in the Software/Data Library by searching on the word ROEDIT, the Q number of [More] There is 1 Reply. #: 16375 S10/Porting from OS/2 02-Nov-92 14:16:29 Sb: #16373-Bitmaps on menus Fm: Nancy Cluts - Microsoft 70744,20 To: Nancy Cluts - Microsoft 70744,20 (X) [Continued] this article, or S13284. ROEDIT was archived using the PKware file-compression utility. More Information: An application can create a read-only edit control by subclassing or superclassing a standard edit control. The subclass procedure filters out messages that change the contents of the edit control. The following code fragment demonstrates this process: FARPROC gOldProc; LONG FAR PASCAL ROEditProc(HWND hWnd, WORD msg, WORD wParam, LONG lParam) { switch (msg) { case WM_KEYUP: case WM_KEYDOWN: case WM_CHAR: case WM_CUT: case WM_COPY: case WM_PASTE: case WM_LBUTTONUP: case WM_LBUTTONDOWN: case WM_LBUTTONDBLCLK: return 1L; case WM_GETDLGCODE: return 0L; } return CallWindowProc(gOldProc, hWnd, msg, wParam, lParam); } //*** ROEditProc In the example above, the subclass procedure traps mouse clicks, keystrokes, and the cut, copy, and paste commands. It also traps the WM_GETDLGCODE message to prevent an edit control in a dialog box from receiving the input focus. The following example demonstrates superclassing an edit control to create a new ROEDIT-class control that behaves similar to a read-only edit control. A ROEDIT control implements the window procedure provided above. It also changes the cursor to an arrow instead of an I-beam, which provides an additional indication to the user that the contents of the control cannot be changed. Using an ROEDIT control eliminates the necessity of subclassing each control after it is created. When the application creates a control from the ROEDIT class, the read-only behavior is provided automatically. The following code demonstrates superclassing an edit control as described above: FARPROC gOldProc; BOOL RegisterROEdit(HANDLE hInstance) { WNDCLASS wc; if (!GetClassInfo(NULL, "EDIT", &wc)) return FALSE; gOldProc = (FARPROC)wc.lpfnWndProc; [More] There is 1 Reply. #: 16376 S10/Porting from OS/2 02-Nov-92 14:16:35 Sb: #16375-Bitmaps on menus Fm: Nancy Cluts - Microsoft 70744,20 To: Nancy Cluts - Microsoft 70744,20 (X) [Continued] wc.style &= ~CS_GLOBALCLASS; wc.lpfnWndProc = ROEditProc; wc.hInstance = hInstance; wc.hCursor = LoadCursor(NULL, IDC_ARROW); wc.lpszClassName = "ROEDIT"; return RegisterClass(&wc); } //*** RegisterROEdit In Windows 3.1, there is a new style bit for the edit control (ES_READONLY) that removes the editing capabilities of the edit control, leaving only the viewing capabilities. This style is useful when the application shows the user a body of static text that the user reads and does not modify. Additional reference words: 3.00 3.10 3.x softlib ROEDIT.ZIP #: 16757 S11/Porting from Unix 05-Nov-92 15:06:38 Sb: ISV Dev'pment Lab Unix,X Fm: Alex Bronstein 75070,2452 To: John Hall [MS SDE] 70750,2341 (X) John, Thanks for your reply. I've moved the thread to section 11 as you requested. You didn't mention any experiences on the non-character based side (i.e. X apps). Too tragic to mention? I'm wondering at what stage of porting can such things be brought to your lab. Regards, Alex There is 1 Reply. #: 16797 S11/Porting from Unix 05-Nov-92 20:22:39 Sb: #16757-ISV Dev'pment Lab Unix,X Fm: John Hall [MS SDE] 70750,2341 To: Alex Bronstein 75070,2452 (X) >> You didn't mention any experiences on the non-character based side (i.e. X apps). Too tragic to mention? I'm wondering at what stage of porting can such things be brought to your lab. The vendor with X which came already had a Windows product, so we used that. The first problem with X would be 'what interface did you use'? My experience was right on top of XLib, not using any toolkits. Offhand, I think a UI framework can be redone on Windows very quickly using various code generation facilities. The concepts are often parallel -- use DC's instead of GC's, for example. The messaging model is a bit different with plusses and minuses (every window gets its own proceedure, which is good. every window gets all messages, which is bad). Graphics and Fonts just get in and get with it. Overall, I have a lot of faith in the ability of Unix ISV's to get the work done once they decide to. There is 1 Reply. #: 16833 S11/Porting from Unix 06-Nov-92 11:22:43 Sb: #16797-ISV Dev'pment Lab Unix,X Fm: Alex Bronstein 75070,2452 To: John Hall [MS SDE] 70750,2341 (X) John, OK, thanks for the encouragement. Everyone out there: If anyone else out there is doing/considering an X -> Win32 port, would you mind sending me an email (either here or alex@gain.com) and I'd maintain a fellow-sufferer list. We're probably in different vertical markets, so I don't think sharing some info on the porting experience would hurt any of our businesses... Regards, Alex #: 16478 S11/Porting from Unix 03-Nov-92 11:38:27 Sb: Porting from Unix Fm: Alex Bronstein 75070,2452 To: John Richardson 70541,672 (X) John, Are you building the X server yourself, or using a commercial product? Also, would you (or your company) be willing to license the port of X lib to NT/Posix to someone else? This is a different strategy than the one we've decided on to port our X app to NT, but it would be well worth our time to try both approaches and compare... Alex (aka: internet>alex@gain.com) There is 1 Reply. #: 16530 S11/Porting from Unix 03-Nov-92 18:10:10 Sb: #16478-Porting from Unix Fm: John Richardson 70541,672 To: Alex Bronstein 75070,2452 (X) I plan on using the Visionware product from leads England. I am running the Windows 3.x version, and will be getting the WINSOCK interface to work under NT. (Currently they use either PC-NFS, or PC-IP) They will not have a native WIN32 version until sometime next year, so if another one comes along, I will try it. One thing in Visionwares favor is that they usally get pretty good ratings, and the company I work for allready licenses the product for Windows 3.x. As far as the basic Xlib port, I do plan on either doing it myself, or someone in my development group will do it. I am currently blocked by not being able to do any kind of IPC communication outside of the POSIX subsystem that comes with NT. As soon as I can do the IPC, I can then bounce to a WIN32 or other WINSOCK program that implements the Xserver. As far as licensing, I work for a large company (Siemens-Nixdorf), and I am just starting to try to convince management that some of what my group does should be available to outside companys. I am writing up some proposals for expanding the scope of co-operation between my group an other 3rd parties and submitting this to my various bosses. I may get lucky here, due to the need to 'leverage' everything a company does today in this economic environment. JR There is 1 Reply. #: 16577 S11/Porting from Unix 04-Nov-92 08:54:52 Sb: #16530-Porting from Unix Fm: Arthur Kreitman 76060,2677 To: John Richardson 70541,672 >As far as the basic Xlib port, I do plan on either doing it myself, or >someone in my development group will do it. I am currently blocked by >not being able to do any kind of IPC communication outside of the >POSIX subsystem that comes with NT. As soon as I can do the IPC, I >can then bounce to a WIN32 or other WINSOCK program that implements >the Xserver. I think that you'll find that the POSIX subsystem is a nonstarter as far as xlib is concerned. It unlikley that MS will upgrade the POSIX subsystem anytime soon. If fact, the "we'll study the matter" response is their way of saying NFW. The real problem with making XLIB (and all its relatives: xt, xaw, motif widgets...) available under NT is the the behaviour has to exactly match the expected behaviour. It's more then just having a select available. There is 1 Reply. #: 16686 S11/Porting from Unix 04-Nov-92 21:37:12 Sb: #16577-Porting from Unix Fm: John Hall [MS SDE] 70750,2341 To: Arthur Kreitman 76060,2677 (X) >> It unlikley that MS will upgrade the POSIX subsystem anytime soon. >> If fact, the "we'll study the matter" response is their way of >> saying NFW. As a long standing Unoid, I can understand the not of cynicism, but IMHO it is not warranted. I'm afraid I can't elaborate. Time is not yet full. There is 1 Reply. #: 16705 S11/Porting from Unix 05-Nov-92 05:16:09 Sb: #16686-Porting from Unix Fm: Arthur Kreitman 76060,2677 To: John Hall [MS SDE] 70750,2341 (X) Want to do something usefull. Make fork and links available in the win32 subsystem. Then you won't have to worry about posix. #: 16482 S11/Porting from Unix 03-Nov-92 12:01:12 Sb: Porting from Unix Fm: Steve Firebaugh [MS] 75430,412 To: John Richardson 70541,672 (X) JR, Here is the official word I've gotten back regarding the limitations on posix IPC. "The first release of the POSIX SS on Win NT will be limited to IEEE 1003.1 (FIPS 151-1) support. This is a big achievement given that POSIX was written for UNIX. However, overall utility of the SS will be limited by its lack of IPC with the other subsystems. Microsoft is looking to improve this support in the future, same as it is studying evolution of the SS." One other person suggested the following technique, which you might find helpful: "The simplest IPC that would seem to solve your problem is to start your POSIX server app from a Win32 app, and re-direct the stdin/stdout handles, giving you two way communication from the Win32 to the POSIX part of you app - you could quite simply layer your Xlib-like calls to use this two-way channel to talk to the Win32 X-Server." Steve Firebaugh There are 2 Replies. #: 16494 S11/Porting from Unix 03-Nov-92 14:52:05 Sb: #16482-Porting from Unix Fm: Seattle Lab 71202,560 To: Steve Firebaugh [MS] 75430,412 (X) How do you start a POSIX server app from within a win32 app? I thought that you couldn't mix environmental subsystems. Dennis Krueger Seattle Lab There is 1 Reply. #: 16526 S11/Porting from Unix 03-Nov-92 17:34:12 Sb: #16494-Porting from Unix Fm: Steve Firebaugh [MS] 75430,412 To: Seattle Lab 71202,560 (X) You can not mix API calls from different subsystems within the same application. However, the Win32 API CreateProcess() is able to create a posix process. Furthermore, this API includes a parameter which allows the "spawned" process to inherit the handles of the Win32 app. Steve Firebaugh There is 1 Reply. #: 16728 S11/Porting from Unix 05-Nov-92 11:32:16 Sb: #16526-Porting from Unix Fm: Seattle Lab 71202,560 To: Steve Firebaugh [MS] 75430,412 (X) Thanks Steve, We are porting the UniVerse application environment, a PICK-like operating system, from UNIX to NT and have been examining cross-over stratagies. We haven't wanted to go completly to POSIX because that wouldn't seem to allow us to implement the type of inter-application connectivity that the NT platform offers. By starting off from a WIN32 environment and spawning off specific POSIX functions, it looks like an approach which will allow us to integrate the best of both worlds. We will have more questions in the future. Dennis Krueger Seattle Lab There is 1 Reply. #: 16843 S11/Porting from Unix 06-Nov-92 14:01:17 Sb: #16728-Porting from Unix Fm: Steve Firebaugh [MS] 75430,412 To: Seattle Lab 71202,560 (X) Dennis Krueger, I just had a meeting with one of the posix developers here. He pointed out that there is a limit on the number of posix processes. It is currently set to 32, but this can be easily raised if there is proper justification. Do you think that you'll run into problems with this? If so, it may be worth your while to send us a limit which will work for you. Thanks. Steve Firebaugh There is 1 Reply. #: 16850 S11/Porting from Unix 06-Nov-92 15:33:29 Sb: #16843-Porting from Unix Fm: Seattle Lab 71202,560 To: Steve Firebaugh [MS] 75430,412 (X) Dear Steve, Thanks for your interest and follow-up. What we are porting could be viewed as a multi-user business operating system supporting up to 1000 users, supporting the entire environment over NT. This currently runs over UNIX supporting each user as a distinct process, say in the range of 1500 processes upper limit. We are now at the design stage of the port. One approach is to attempt a direct port from SCO UNIX to the POSIX standard. Our own environment sub-system and file system is the other extreme. We are leaning toward a more NT integrated approach. Your suggestion of a mix of POSIX processing from a WIN32 shell being the most attractive. Any POSIX approach would appear to require over 1000 processes unless we were able to run users as threads within one process, which brings up environmental access and scheduling issues. And we haven't raised any peripheral control issues. Enough for now, Thanks again for your interest and response. Dennis Krueger Seattle Lab #: 16509 S11/Porting from Unix 03-Nov-92 16:19:02 Sb: #16482-Porting from Unix Fm: Arthur Kreitman 76060,2677 To: Steve Firebaugh [MS] 75430,412 (X) > "The first release of the POSIX SS on Win NT will be limited to IEEE > 1003.1 (FIPS 151-1) support. This is a big achievement given that > POSIX was written for UNIX. However, overall utility of the SS will The posix subsystem cannot be considered a big achievement. The only hard part was fork, and fork support was a NT kernel design requirement. The only big part of the achievement was believing that they could get away with an implementation that had that many limitations. #: 16352 S12/API-RPC/WinNet 02-Nov-92 11:47:08 Sb: #16251-RPC nesting alls Fm: Koby 71172,2722 To: Bruce Ramsey/Microsoft 70324,2742 (X) Thank you very much, Bruce for the detailed absware. I have to digest it. I am waiting to the end result. Koby #: 16271 S12/API-RPC/WinNet 31-Oct-92 09:01:31 Sb: RPC Callbacks Fm: Daniel Norton 76050,2204 To: All I am working with a producer-consumer application whereby the consumer calls the producer to register itself so that the producer can begin providing information. As information becomes available, the producer calls back the consumer for each item of information. Unfortunately, it does not seem to me that RPC accounts for this. The concept is similar to a callback with the Windows timer. Instead of RPC here, I plan on providing a handle to something like a named pipe that the producer will write to, but I'd rather maintain the RPC model, since it is platform independent. Once I add named pipes, I become platform specific and the problem becomes more complicated. Can anyone see a simpler solution? Thanks. -- Daniel There is 1 Reply. #: 16355 S12/API-RPC/WinNet 02-Nov-92 11:59:12 Sb: #16271-RPC Callbacks Fm: David Taniguchi [MS] 72350,2054 To: Daniel Norton 76050,2204 (X) Hi Daniel, If the clients are NT machines (or other computers capable of creating RPC servers) you could create two RPC servers (both on the producer and consumer sides). However, on Dos and Windows clients, this won't work since you cannot create RPC servers on Dos and Windows 3.1 machines at this time. RPC is capable of passing parameters both to and from a client machine. However, this depends that a call is initiated by the client. You could have the client keep checking the server (by calling a remote procedure call), and when interesting information is available, the server could pass back the information in a parameter. Also, if it is possible to narrow down what types of machines you would potentially be communicating with, you can consider named pipes, Windows Sockets, or netbios. This isn't as nice as having the platform independence that RPC provides, but these mechanisms do provide the functionality you seek. Hope this helps, Dave #: 16370 S12/API-RPC/WinNet 02-Nov-92 14:13:04 Sb: #16219-network registry access Fm: Bruce Ramsey/Microsoft 70324,2742 To: Andrew Potter 71075,614 (X) Hi Andrew - >> ...If LANManager can administer user account information, can it be used to manage other user defined information? For example, project_A's design files are located on server_A_proj under the /proj_A file system and project_A's database is located on server_DB_A_proj under the /proj_A file system. If LANManager cannot do the above example, can the registry provide the above information? LAN Manager for Windows NT has a Replicator service, and can, for example, replicate the contents of the directory c:\proj_A on machine \\srvr_A_pr to the directory c:\any_dir\proj_A on the machine \\srvDB_A_pr The Repl server needs to be a LAN Manager for Windows NT machine, however the Repl client can be a Windows NT machine, that is, need not have the LAN Manager for Windows NT add-on Subdirectories of a replicated directory are also replicated, that is, the entire tree structure is maintained I threw in the sub-dir name "any_dir" as an example - the point is the disk letters on the Repl source and client machines don't have to be the same, and neither do the directory paths. You could replicate \\srvr_A_pr 's c:\proj_A to \\srvDB_A_pr 's d:\dontcare Bruce #: 16276 S12/API-RPC/WinNet 31-Oct-92 09:35:18 Sb: Win16 Netbios problem Fm: Michael Lynch 76260,1670 To: Bruce Ramsey/Microsoft 70324,2742 (X) Thanks for the response. Unfortunately, these are commercial apps, not mine. Specifically, Attachmate Extra for Windows... and Keyfile (A document managmemnt app). What is curious is that both seem to be behaving the same (ie, wait 3 second s and die). I will try removing TCP/IP and seeing if this is the problem. Can you give me an example of a commercial app that you have tried that does work, so I can attempt to duplicate this in my environment? Thanks Mike There is 1 Reply. #: 16371 S12/API-RPC/WinNet 02-Nov-92 14:13:09 Sb: #16276-Win16 Netbios problem Fm: Bruce Ramsey/Microsoft 70324,2742 To: Michael Lynch 76260,1670 Hi Mike - >> ...Unfortunately, these are commercial apps, not mine. Specifically, Attachmate Extra for Windows... and Keyfile (A document managmemnt app). What is curious is that both seem to be behaving the same (ie, wait 3 second s and die) I apologize, Mike, I misunderstood - I thought these were your apps. The people most focussed on problems with running commercial 16-bit apps on Windows NT are in WINNT/section 4 - please check with them there >> I will try removing TCP/IP and seeing if this is the problem. Can you give me an example of a commercial app that you have tried that does work, so I can attempt to duplicate this in my environment? I think removing TCP/IP is still worth trying. Unfortunately I don't have an example of an app similar to what you're trying - again, please check with the people who focus on this area in WINNT/4 Bruce #: 16390 S12/API-RPC/WinNet 02-Nov-92 15:55:51 Sb: Re: Win RPC / UNIX RPC Fm: Bruce Ramsey/Microsoft 70324,2742 To: gene saunders 72265,23 Hi Gene - I checked with the OSF. They were't able to narrow the list down to only UNIX vendors, but here is the current list of vendors shipping DCE: IBM DEC Siemens Nixdorf HP Groupe Bull SCO NCR SUN (Transarc Corporation) Stratus Gradient Technologies (MS Windows) USL Hitachi The OSF mentioned that there are a little over a dozen additional companies that are working in the area but haven't yet made announcements Microsoft is of course shipping in the Win32 Preliminary SDK for Windows NT the RPC subset of DCE Bruce #: 16436 S12/API-RPC/WinNet 02-Nov-92 22:36:52 Sb: RAW IP Support in NT Fm: Balaji - SynOptics 76330,1047 To: David Taniguchi [MS] 72350,2054 (X) Dave, When I was refering to ICMP, I was asking about having access to the ICMP HEADER. Also, Eventhough WinSock spec doesnot require RAW sockets implemented, is Microsoft implementing it in NT?. I know that most of the stack vendors who are working on RAW Sockets for windows already have this in their stacks. -Balaji There is 1 Reply. #: 16725 S12/API-RPC/WinNet 05-Nov-92 10:46:14 Sb: #16436-RAW IP Support in NT Fm: David Taniguchi [MS] 72350,2054 To: Balaji - SynOptics 76330,1047 Hi Balaji, We are not going to support RAW sockets in product 1. I am not certain how ICMP IP packets will be intercepted. I'm checking into this. Will keep you posted. Thanks, Dave #: 16277 S13/Windows NT DDK 31-Oct-92 10:42:05 Sb: #16238-DDK Rollout Fm: Marcelo R Lopez Jr 71221,1353 To: Steve Mastrianni [PSS] 71501,1652 (X) Steve, Thank you for making me GREEN(!). I think my wife and I had OTHER MORE IMPORTANT matters at hand, and I'm sure it's something I would've enjoyed, but I wouldn't trade having been the first person to hold my son, for anything in the World !! ENVY, ENVY *~*~*~* Marcelo R. Lopez, JR. DRV Research, Inc. PS. Call you at home tonight, cool ?! #: 16379 S13/Windows NT DDK 02-Nov-92 14:50:36 Sb: Open SCSI w/CreateFile() Fm: Paul Sanders [Microsoft] 72350,2053 To: Jim Gill 70530,1771 (X) Jim, I need to retract my last answer. Miniport drivers can support IOCtls. We have no sample code that does this, so you will probably be forging new paths (not actually, it will just seem like it). Miniport IOCTLs are typically for driver developers that want to set up special adapter features such as cache or fault tolerance. Your Win32 app gets a device handle for the miniport driver by opening the file : \\.\\ScsiPortN, where N = 0, 1, ... Here are a couple of revelvant questions and their answers: | How can I determine what port my miniport driver gets mapped to? If I | have an aha154x and fd8xx in the machine, which is port0 and which is | port1? Would I have to look in the load order to determine this? There are a couple ways of doing this. One is to interrogate the registry. The information is there. The second is to open each port by name and look for your card. | What if I opened '\\.\d:' would the right thing happen and would my | ioctl get routed to the correct port? Yes it would, assuming that d: happened to map to a partition on a device on that port. SCSI drivers pass unknown IOCTls down to the next layer. Hope this helps and sorry for the previous incorrect information, Paul Sanders, Microsoft #: 16382 S13/Windows NT DDK 02-Nov-92 15:20:32 Sb: Device IO control for NT Fm: Paul Sanders [Microsoft] 72350,2053 To: Robert Strickler 71160,2462 Robert, Get the handle to you miniport driver by opening \\.\ScsiPortN, where N = 0, 1, 2... Once you have a handle, use DeviceIoControl in your Win32 app to pass the IOCtl down to the driver. See my answer to Msg #16165. Hope this helps, Paul Sanders, Microsoft #: 16338 S13/Windows NT DDK 02-Nov-92 10:17:49 Sb: Memory maped IO / Ints Fm: Fred Schempp 72677,2672 To: All Can anyone help me find a either example code or a referance that will allow me to do program memory mapped IO and write a simple interupt service rout We have scoured the Win32 docs and cant find a referance on accessing PHYSICAL memory or the handling of INT's. We are told the DDK may help but it wont be available for a month. I would appreciate any help or suggestions. Thanks There is 1 Reply. #: 16416 S13/Windows NT DDK 02-Nov-92 17:59:37 Sb: #16338-Memory maped IO / Ints Fm: Cliff Garrett 76702,1531 To: Fred Schempp 72677,2672 (X) Fred, Q) Can anyone help me find a either example code or a referance that will allow me to do program memory mapped IO and write a simple interupt service rout We have scoured the Win32 docs and cant find a referance on accessing PHYSICAL memory or the handling of INT's. We are told the DDK may help but it wont be available for a month. I would appreciate any help or suggestions. Thanks A) It sounds like you are trying to operate a board of some sort that is memory mapped, and possibly occupies some ports. This basically constitutes the need for a miniport driver, and you are right, that the DDK provides examples, and tools to do this. This DDK is available now, so you don't have to wait a month. What is the board you are trying to write code for, a scanner, fax board or something like that?. The miniport model will allow you to map memory, and register asn ISR quite easily, and allow a USER mode process to interact with it via IOCTL calls. Please let me know if you would like more information on this. Cliff Garrett #: 16381 S13/Windows NT DDK 02-Nov-92 14:55:36 Sb: Promiscuous RX Stack? Fm: Brian Lieuallen [MS] 76300,2506 To: Christian Beaumont 100034,2326 (X) Christian, This is definitly possible. You will have to write an NT device driver that will communicate with upper edge services provided by NDIS(MacAddAdapter, MacTransferData ect.) Brian There is 1 Reply. #: 16441 S13/Windows NT DDK 03-Nov-92 03:32:02 Sb: #16381-Promiscuous RX Stack? Fm: Christian Beaumont 100034,2326 To: Brian Lieuallen [MS] 76300,2506 (X) Thanks Brian, I'll get stuck into the coding just as soon as I receive my DDK, but it sounds like it may take some time. Christian #: 16469 S13/Windows NT DDK 03-Nov-92 09:51:38 Sb: NDIS3 & BusMaster DMA Fm: bruce tanner 100063,1120 To: Brian lieuallen [MS] 76300,2506 (X) Hi Brian, These are probably silly questions ...but ... here they are anyway:- q1. What HAL routines does NdisAllocateSharedMemory() use? q2. I am working on an NDIS3 driver for a Bus master network card. Consequently I am dealing with physical addresses aswell as virtual addresses. Are there any easy virtToPhys and physTovirt functions? Thanks. #: 16471 S13/Windows NT DDK 03-Nov-92 10:36:21 Sb: NDIS3 & Bus master DMA Fm: bruce tanner 100063,1120 To: Brian Lieuallen [MS] 76300,2506 (X) Another question, I need to be able to read a WORD from an area of memory described by a physical address. I was thinking of using NdisReadRegisterUshort() ...but I need the virtual address of the area, to be able to use this ..... again a physTovirt would be handy .... Similarly, if I wanted to write a WORD to an area of memory described by a physical address ..to use NdisWriteRegisterUshort() I would need a physTovirt. Can I acheive this functionality without physTovirt in some other way? Or am I being really silly. Any help would be much appreciated! Thanks #: 16480 S13/Windows NT DDK 03-Nov-92 11:55:26 Sb: Beta/Build.exe error Fm: Rick Holzer 70760,2473 To: sysop (X) I just installed the Beta copy of the NT and the SDK. I also re-installed the DDK. As instructed in the installation program, I run "\ddk\buildenv" and then "build -cef", in my development directory. Result: A "BUILD.EXE - Application Error" windows pops up on the screen. The message in the window is as follows: The instruction at "0x0001976e" referenced at "0x61206578".The memory could not be "read". Click on OK to terminate the application Click on CANCEL to debug the application The DOS windows displays the start of the build session: BUILD: Compile and Link for (but the text ends there). This didn't happen before the Beta release. Did I miss something concerning the new version's installation? Help!! Rick There is 1 Reply. #: 16502 S13/Windows NT DDK 03-Nov-92 15:14:49 Sb: #16480-Beta/Build.exe error Fm: Paul Sanders [Microsoft] 72350,2053 To: Rick Holzer 70760,2473 Rick, Have you, 1 - run \mstools\setenv.bat, or 2 - added \mstools\bin to your path set LIB=x:\mstools\lib set INCLUDE=x:\mstools\h Do one or the other (I do #2 from the System applet under Control Panel). Then execute \ddk\buildenv. Let me know if that clears up the problem. Paul Sanders, Microsoft #: 16304 S13/Windows NT DDK 01-Nov-92 13:50:43 Sb: KD viewing local symbols Fm: Jim Harford 72350,3705 To: all Have not gotten a response to my last message. Here is a repeat: I need to view local symbols via kernel debugger. I can see global variables but setting NTDEBUG=ntsd does not seem to help. I am using July SDK and DDK. Can someone at Microsoft help me with this? Thanks in advance! There are 2 Replies. #: 16333 S13/Windows NT DDK 02-Nov-92 10:02:22 Sb: #16304-KD viewing local symbols Fm: Paul Sanders [Microsoft] 72350,2053 To: Jim Harford 72350,3705 Jim, That would be my fault. I wasn't sure of the answer and with most people gone to the conference, I wasn't able to get much assistance. I have sent email around and should get some help shortly. Sorry for the inconvenience, Paul Sanders, Microsoft #: 16507 S13/Windows NT DDK 03-Nov-92 16:05:46 Sb: #16304-KD viewing local symbols Fm: Paul Sanders [Microsoft] 72350,2053 To: Jim Harford 72350,3705 Jim, I just received email from a developer who states that he could never get it to work for either NTSD or I386KD. For now, consider it broken and I will contact the appropriate people with your problem. Sorry, Paul Sanders, Microsoft Developer Support #: 16476 S13/Windows NT DDK 03-Nov-92 11:09:14 Sb: DDK update for beta 1 Fm: Muzaffer KAL [MSC] 70324,2553 To: MS is there going to be a DDK update to match October beta? If yes, will it include more driver sources ? (like a Microsoft serial mouse driver source) Muzaffer There is 1 Reply. #: 16485 S13/Windows NT DDK 03-Nov-92 12:11:23 Sb: #16476-DDK update for beta 1 Fm: Cliff Garrett 76702,1531 To: Muzaffer KAL [MSC] 70324,2553 (X) Muzaffer /is there going to be a DDK update to match October beta? If yes, will it /include more driver sources ? (like a Microsoft serial mouse driver source) Yes there will be a new DDK. This will have more sources as well. The sermouse is one of them, but there is also video. This should be release within the next week. cliff There is 1 Reply. #: 16511 S13/Windows NT DDK 03-Nov-92 16:28:26 Sb: #16485-DDK update for beta 1 Fm: Muzaffer KAL [MSC] 70324,2553 To: Cliff Garrett 76702,1531 (X) great, thanks for the info. Muzaffer #: 16571 S13/Windows NT DDK 04-Nov-92 05:53:12 Sb: NdisAllocateSharedMemory Fm: bruce tanner 100063,1120 To: Brian Lieuallen [MS] 76300,2506 (X) Hi, I want to allocate some memory for use with a BusMaster DMA card. From the looks of it I need to use NdisAllocateSharedMemory(). However, my card can only address upto 16M. So my question is whether Ill be ok allocating buffer space using NdisAllocateSharedMemory(). Thanks. There is 1 Reply. #: 16633 S13/Windows NT DDK 04-Nov-92 13:46:40 Sb: #16571-NdisAllocateSharedMemory Fm: Brian Lieuallen [MS] 76300,2506 To: bruce tanner 100063,1120 (X) Bruce, NdisAllocateSharedMemory() is the service that you will want to use. The system will take care of making sure that the memory is accessable. In answer to your earlier questions, NdisAllocateSharedMemory more or less translates to HalAllocateCommonBufer(). There is no service to specifically convert from a physical address to a virtual address, but NdisAllocateShared memory returns both a virtual address and a physical address. NdisStartBufferMapping can be used to get the physical address of a buffer to pass to your busmaster card. You can also use NdissGetBufferPhysicalAddres(). Brian #: 16715 S13/Windows NT DDK 05-Nov-92 08:29:35 Sb: NT - PCL Native Driver Fm: Giri Sithamraju 73157,3306 To: Dave Snipp The final version of Windows NT DDK when released, will it contains any code examples to write a native mode driver for PCL5/PCL4 printer. Are we limitied to always use the minidrivers with some sort of filter procedures with the Print Processors. These filter processors may add the performance enhancements that are unique to our printer hardware. Please advice me how one should prepare for the development of PCL printer drivers for NT. Thank You Giri #: 16638 S13/Windows NT DDK 04-Nov-92 14:27:32 Sb: Unique Number Allocation Fm: Michael Hall 72607,3422 To: Paul Sanders [MS] 72350,2053 (X) Paul, The class/port driver pair that we are writing doesn't fit into the set of standard devices, so the predefined Device Type values (used when calling IoCreateDevice) won't work for us. After talking with Bob Taniguchi, Keith Kegley, and several developers, the general feeling was that Microsoft would be (hopefully soon) providing some method of requesting a Device Type value. This would probably work like the Windows 3.1 VxD ID number allocation. A similar method of allocating numbers will also be needed for Public Device I/O Control values. I'm sure several of the (MS) people we talked with a the DDK conference are working on setting up these mechanisms, but I wanted to send in a note saying "YES, this is very important" and "please setup something soon." For our in-house development and testing we can choose any numbers for the Device Type values and Public Device I/O Control values that we need. However, when we start sending our drivers out for beta testing, I really would like to have our own, unique numbers for these values. I just wanted to emphasize this need. Thanks for listening. Mike There is 1 Reply. #: 16718 S13/Windows NT DDK 05-Nov-92 09:17:31 Sb: #16638-Unique Number Allocation Fm: Paul Sanders [Microsoft] 72350,2053 To: Michael Hall 72607,3422 (X) Mike, I am forwarding your mail on to appropriate groups who are interested in your concerns. If I receive any feedbac, I will pass it on to you. Thank you for your comments, Paul Sanders, Microsoft #: 16735 S13/Windows NT DDK 05-Nov-92 12:24:57 Sb: Ndis 3.0 and DMA Fm: Brian Lieuallen [MS] 76300,2506 To: Christian Thrysoe 100022,1401 (X) Christian, In order to setup busmaster ISA card with NDIS you need to set the Master element of the AdapterInformation structure to TRUE. You should also specify the channel number that you will be using. The AdapterInformation structure is then passed to NdisRegisterAdapter(). You should not call NdisAllocateDmaChannel. Brian #: 16345 S13/Windows NT DDK 02-Nov-92 10:56:21 Sb: WinNT multitasking Fm: Harry Burnette 75140,210 To: sysop (X) I am curious about the capabilities of WinNT to multitask DOS applications. For example, will I be able to back up to tape using CPBackup for DOS while simultaneously formatting a diskette, communicating on CompuServe and running Norton Speed Disk? (Just curious; don't know if I'd be crazy enough to really try this!!) Thanks, Dave P.S. If this question belongs in a different section or forum, please let me know. There is 1 Reply. #: 16398 S13/Windows NT DDK 02-Nov-92 16:01:00 Sb: #16345-WinNT multitasking Fm: Cliff Garrett 76702,1531 To: Harry Burnette 75140,210 (X) Harry, NT Will multitask MS-DOS apps great. I typically use 4dos in one command prompt, while running TAPCIS in another, and I can also format a floppy at the same time, albeit in an NT prompt. As far as the CPBackup, I haven't tried this. Cliff Garrett There is 1 Reply. #: 16515 S13/Windows NT DDK 03-Nov-92 16:46:40 Sb: #16398-WinNT multitasking Fm: Jeff Thomson 71460,3222 To: Cliff Garrett 76702,1531 (X) Cliff, > NT Will multitask MS-DOS apps great. I typically use 4dos in one command > prompt, while running TAPCIS in another I have run TAPCIS in a DOS VM, but had a bad problem w/ dropped chars @9600 baud when posting messages composed offline. Any advice on PIF settings that will allow this setup to work more reliably would be greatly appreciated. -- Jeff There is 1 Reply. #: 16740 S13/Windows NT DDK 05-Nov-92 12:49:51 Sb: #16515-WinNT multitasking Fm: Cliff Garrett 76702,1531 To: Jeff Thomson 71460,3222 (X) I do not use any PIF settings. sorry. cliff #: 16568 S13/Windows NT DDK 04-Nov-92 05:23:38 Sb: NTSD with display driver Fm: ray shapiro 71431,1450 To: sysop (X) Hi, When I was out in Anaheim, I asked about using NTSD to debug display drivers. The answer was "yes", but the magic incantaions went by a bit fast. Could someone please post the rules for invoking NTSD to debug display drivers. (I accidentally posted this to the non-tech forum a couple of days ago.) Thanks, Ray There is 1 Reply. #: 16755 S13/Windows NT DDK 05-Nov-92 14:48:32 Sb: #16568-NTSD with display driver Fm: Cliff Garrett 76702,1531 To: ray shapiro 71431,1450 (X) Ray, /When I was out in Anaheim, I asked about using NTSD to debug /display drivers. The answer was "yes", but the magic incantaions /went by a bit fast. Could someone please post the rules for /invoking NTSD to debug display drivers. I don't know in what detail you expect to debug your display driver with NTSD. I would urge you now to use i386KD. You can trace into the detail of what your BitBlt is going to do, ect.. Is the reason you are asking because you do not want to use two machines?? Please let me know what the intention is here. Thanks, Cliff Garrett There is 1 Reply. #: 16765 S13/Windows NT DDK 05-Nov-92 15:57:03 Sb: #16755-NTSD with display driver Fm: ray shapiro 71431,1450 To: Cliff Garrett 76702,1531 (X) There are several reasons why I want to use NTSD. One reason is because NTSD is a symbolc debugger which let's you see your code. Another reason is because it would be possible to do some debugging on a single machine, but I would like to use NTSD on a separate machine, also. Apparently, Jeff Newman and some of the other display Microsoft driver developers do this, and think it is useful. Jeff and others outlined a few ways in which NTSD could be used... one involved loading a secondary display driver and the other was using the kernal debugger to set a breakpoint in the display driver and having NTSD do it's thing while the driver was initializing. I'm writing a driver in C. The kernal debugger is not really all that helpful for the work I'm doing. It seems to be standard paractice other there is Redmond! Thanks, Ray #: 16816 S13/Windows NT DDK 06-Nov-92 09:15:42 Sb: NT Run-Time Functions Fm: Peter G. Johansson 71023,557 To: SYSOP (X) Where can I find the include file that has function prototype definitions for such things as RtlZeroMemory, RtlMoveMemory, etc? There used to be a file on either the SDK or the DDK called "ntrtl.h", but I cannot locate this on either October release CD-ROM. Have these been moved to a new include file? If not, why are they absent from the SDK/DDK? #: 16817 S13/Windows NT DDK 06-Nov-92 09:35:13 Sb: Problem with XGA Fm: Randy Threewits 71005,1373 To: sysop (X) There seems to be a problem running DOS apps when an XGA adapter exists. I installed NT (Oct release) on an IBM Model 80 (16 Meg Ram, 25 MHz) with a standard VGA monitor and everything worked fine. I then added an XGA-2 card in the system within DOS and everthing worked fine with both monitors and adapters. In NT I did NOT change the setup to indicate a new adapter. I still wanted to use the VGA as the monitor since the VGA driver appears to be more complete. I did not have a monitor attached to the XGA-2 adapter. So essentially my setup was VGA on planar, XGA-2 card and an 8513 monitor on the VGA. When I try to run 16-bit DOS application [ex. printf("Hello world") ] I get the following error: Title bar: NTVDM.EXE - Application Error Message: The instruction at "0x60615947" referenced memory at "0x30180700". The memory could not be "written". OK-terminate, Cancel to debug. This is essentially the same thing that occurred when I installed NT with the XGA-2 as the adapter to use. If I remove the card, HelloWorld works fine. It almost looks like NT is using the XGA driver when doing DOS virtualization. Is there some way that I can copy and rename drivers so that I use VGA on NT but XGA on DOS and OS/2? Removing the card is a non-desirable option. Thanks #: 16820 S13/Windows NT DDK 06-Nov-92 10:38:46 Sb: File Systems... Fm: Mike Shappell 70007,5116 To: MS Support I noticed that information necessary to create a new file system is not included in the DDK (not that I am surprised given the closed nature of the system). I was wondering where I can get this information. I am currently working on a project that will require writing a new IFS for NT. Thanks in advance for any information you can provide. Mike Shappell Intuitive Software Systems, Inc. #: 16320 S13/Windows NT DDK 01-Nov-92 20:18:38 Sb: Postscript viewer ? ? ? Fm: Jim U'Ren 72371,1235 To: sysop (X) I am looking for a Postscript viewer for the NT docs on the Beta CD-ROM. The file directory indicates a PSVIEW.ZIP file has be uploaded but it does not appear to be available in the Forum. Would appreciate any help you can avail. Jim There is 1 Reply. #: 16717 S13/Windows NT DDK 05-Nov-92 08:50:02 Sb: #16320-Postscript viewer ? ? ? Fm: Stu Wiley 70473,1351 To: Jim U'Ren 72371,1235 Jim, Microsoft doesn't produce a postscript viewer, so anything you find will be third party. Apparently WIX has one on their BBS, and there is another called Ghostscript floarting around up here. I suggest you ask ALL in the WINNT forum section 1, that way you can get some suggestions. I am not aware that one has been loaded to the libraries, but lib 1 in WINNT is your best bet. The October WIN32SDK release has docs in Windows Write File Format as well, so you can view through compatible word processors and desktop publishing software. The Windows Viewer that comes on teh MSDN reliminary CD-Rom is the best Microsoft solution for viewing the references contained on the WIN32SDK CD. Stu Wiley Developer Service Team There is 1 Reply. #: 16791 S13/Windows NT DDK 05-Nov-92 19:27:06 Sb: #16717-Postscript viewer ? ? ? Fm: Jim Gill 70530,1771 To: Stu Wiley 70473,1351 (X) Stu, Apart from the viewer question, are there docs on the October DDK disk? The release notes imply it (I *think*) but I don't see anything that looks like docs. Apologies in advance if I've stupidly overlooked them (that is, once you tell me where they are). There is 1 Reply. #: 16831 S13/Windows NT DDK 06-Nov-92 11:15:30 Sb: #16791-Postscript viewer ? ? ? Fm: Dwight Matheny/Microsoft 70750,2340 To: Jim Gill 70530,1771 (X) I'm assuming you got the DDK CD at the conference. There are no docs on this CD- they weren't done in time. All attendees to the Anaheim conference will get a new CD (with the docs in PS and .WRI) and a set of printed docs. -Dwight (MS) #: 16842 S13/Windows NT DDK 06-Nov-92 13:58:40 Sb: NDIS Protocol Stack Info Fm: Joe Erickson 71360,466 To: All I am interested in creating a protocol stack for Windows NT and would like to know where to get more info on doing so. I have the SDK and DDK, but they do not provide a lot of help on the subject. What would really be nice is an example protocol stack (built on NDIS 3.0 of course). Any help will be greatly appreciated! #: 16604 S13/Windows NT DDK 04-Nov-92 11:07:39 Sb: 3rd Party SCSI H/A's Fm: Cameron T. Brett (UltraS 72370,1130 To: Paul Sanders [Microsoft] 72350,2053 (X) To Sysop or SCSI Miniport Specialist: We are trying to install a third party SCSI host adapter that is not i not included in the boot disk of the October BETA release. We have spoken to some engineers at the porting lab and obtained some instructions for doing this. Is there any other documents available detailing installing a SCSI host adapter that is not included on the boot disk?? Thanks in advance for you help. Sincerely, Cameron T. Brett There is 1 Reply. #: 16856 S13/Windows NT DDK 06-Nov-92 16:07:46 Sb: #16604-3rd Party SCSI H/A's Fm: Paul Sanders [Microsoft] 72350,2053 To: Cameron T. Brett (UltraS 72370,1130 Cameron, Did this get resolved for you? I understand Lee Fisher provided UltraStor with a method for doing this. If not, let me know. Thanks, Paul Sanders, Microsoft #: 16618 S13/Windows NT DDK 04-Nov-92 12:43:02 Sb: net start/stop Fm: Michael Hall 72607,3422 To: Paul Sanders 72350,2053 (X) Paul Sanders: I am writing an NT device driver and would like to be able to dynamically load and unload it using "net start " and "net stop ". If is a monolithic device driver then the driver successfully loads and unloads. Unfortunately, my device driver is layered. My class driver successfully unloads but I am unable to reload it. This is the error that it gives me the first time: System error 5 has occurred. Access Denied. This is the error that it gives me every other time: System error 2 has occurred. The system cannot find the file specified. The information printed out by my class driver indicates that it was unable to get the port device driver object. I believe that this is because when the class driver is unloaded, I don't undo the effect of the IoGetDeviceObjectPointer() call that my class driver's DriverEntry() uses to establish a connection to my port driver. If I try to unload my port driver then the system seems to go through some timeout period and then claims that it has unloaded my port driver: The BRUTUS service is stopping...... The BRUTUS service was successfully stopped. However, the port driver's unload routine is never called and when I subsequently try to load my port driver I get the following error: The requested service has already been started. More help is available by typing NET HELPMSG 2182. I believe that the port driver is never really unloaded, but the net program incorrectly reports that it was. I also believe that the reason that the port driver is unable to unload is because it thinks that it is still being used because the class driver was unable to undo the effect of its IoGetDeviceObjectPointer() call. Please help me understand what is going wrong and explain how I can fix it. Thanks for your time, Clay Bean. There is 1 Reply. #: 16862 S13/Windows NT DDK 06-Nov-92 16:24:09 Sb: #16618-net start/stop Fm: Paul Sanders [Microsoft] 72350,2053 To: Michael Hall 72607,3422 Clay, I'm not really sure of an immediate answer for you. We'll probably have to go through a few iterations here to solve it. My first couple of questions are probably off the wall, but they're a start. #1 (shot in the dark) - log on as Administrator and retry loading and unloading. It might be related to access rights (but I doubt it). #2 find a few strategic points to break into the kernel debugger. Once in type : !drivers This will display all device drivers that are loaded plus their memory information. In the meantime, I will continue to research for you. If you get any pertinent information, please let me know. Thanks, Paul Sanders, Microsoft #: 16586 S13/Windows NT DDK 04-Nov-92 09:52:11 Sb: subsystems and pipes/LPC Fm: Enzo Piombo 100064,2756 To: ALL Hello, I need to find a way to pass data from programs running in any subsystem (OS/2, Posix, Dos, WOW, Win32) to a single program running in the Win32 subsystem and viceversa. Do anyone have any idea about which are the possible ways: I read that LPC is used internally by the system, is it usable by any program in any subsystem ? How ? Can named pipes do the job ? What about performances ? I've a further constraint: I need to produce asynchronous (unsolicited) notifications from the Win32 program to the programs running in the various subsystems. The effect of the notifications should be the call of a postroutine for DOS programs and the Post of a Windows message for WOW programs. If there is not any standard feature of this kind, can a driver implement these functionalities ? Every suggestion will be greatly appreciated. There is 1 Reply. #: 16863 S13/Windows NT DDK 06-Nov-92 16:24:13 Sb: #16586-subsystems and pipes/LPC Fm: Paul Sanders [Microsoft] 72350,2053 To: Enzo Piombo 100064,2756 Enzo, I don't think you'll want to use a device driver for this. I don't know the answer. Try reposting this question in some of the other sections of MSWIN32. Thanks, Paul Sanders, Microsoft #: 16637 S13/Windows NT DDK 04-Nov-92 14:27:25 Sb: Fail an Associated IRP? Fm: Michael Hall 72607,3422 To: Paul Sanders [MS] 72350,2053 (X) Paul, The DDK conference was very informative. I found out the answer to my "Dos copy command" question. Darryl wasn't sure why it failed in the July release, but it works correctly in the Oct release. We talked to Darryl about using Associated IRPs when sending IRPs from our class driver to our port driver. Associated IRPs have the benefit of the I/O Manager automatically completing the Master IRP when all of the Associated IRPs are completed by our port driver. My questions is: what happens when the port driver fails (completes with an error status) one of the Associated IRPs. Since the remaining Associated IRPs will be in the port driver's IRP queue, how will my port driver know to ignore those IRPs? The class driver might send three Associated IRPs to the port driver. If the first one fails, I don't want the port driver to act upon the other two IRPs. Will the I/O Manager cancel the remaining Associated IRPs or will my port driver need to "know" that it just failed an Associated IRP so it needs to remove the remaining Associated IRPs from the queue (and complete them with an error status) before proceeding with the next IRP? It seems like the I/O Manager should take care of this for me. Having my port driver "know" to ignore some number of IRPs after failing one seems a little dangerous. Thanks, Mike There is 1 Reply. #: 16864 S13/Windows NT DDK 06-Nov-92 16:24:19 Sb: #16637-Fail an Associated IRP? Fm: Paul Sanders [Microsoft] 72350,2053 To: Michael Hall 72607,3422 Mike, Nothing is done to track Associated IRPs and deal with them. This is because it would be a massive undertaking to try and anticipate all possible failure modes and deal with each of them. Unfortunately, this is the cost of Associated IRPs. It's up to you, the driver writer, to deal with each failure appropriately. It might be simpler to use multiple IRPs and handle them yourself. Thanks, Paul Sanders, Microsoft #: 16295 S14/Win32s Specific 01-Nov-92 09:35:44 Sb: #16231-Win32 Applications in Wi Fm: Bill Pytlovany 70642,2122 To: Lee Hart [Microsoft] 76150,2536 (X) >>It would really help if you try running your apps... After re-reading, I wanted to be sure you understood that I'm not talking about my apps. I'm trying to run other applications like those shipping with Windows NT.(ie Clock, Calendar, the new WIn32 Help etc...) I would have thought they would work ok with Win32s, perhaps I'm jumping ahead too soon. I'll let you know if I have problems with the app I'm doing. Thanks! Bill Pytlovany #: 16357 S14/Win32s Specific 02-Nov-92 12:13:27 Sb: #16223-win32s octrel outofmem Fm: Mark Kumler 70404,261 To: Lee Hart [Microsoft] 76150,2536 (X) Lee: Thanks I will give that a try. Mark #: 16366 S14/Win32s Specific 02-Nov-92 14:09:27 Sb: WIN32s Fm: Lee Hart [Microsoft] 76150,2536 To: Eric M. Guli 70272,3627 (X) Do you get these error messages during the install, or after install and once you have rebooted? Do you have a kernel debugger for your Windows 3.1 machine set up? Lee #: 16367 S14/Win32s Specific 02-Nov-92 14:09:32 Sb: INSTALL PROBLEM Fm: Lee Hart [Microsoft] 76150,2536 To: Steve Woolley 75570,1620 (X) Have you tried making a fresh floppy and re-installing? (2) Can you try installing on a non-stac'd partition? Thanks, Lee #: 16406 S14/Win32s Specific 02-Nov-92 16:35:04 Sb: #16248-Unhandled exception Fm: Lee Hart [Microsoft] 76150,2536 To: Robert Zeff 70323,1251 (X) >>>I am using the debugger, but can't figure out where the problem is. Any hints would be appreciated. Which Win32sDebug flags have you set in system.ini? You may want to set the verbose bit, as documented on Pages 19 & 20 of the Win32s Programmer's Reference. The SEH Info bit may also be of use. Lee #: 16284 S14/Win32s Specific 31-Oct-92 17:41:56 Sb: MFC / DLL Fm: Jim Conger 73220,324 To: ALL I'm having trouble converting a DLL based on the MFC classes from Windows 3.1 to Windows NT. Although both the DLL and EXE compile fine, the linker can't find the exported function from the DLL. I'm using the following linker command line to create the DLL and export library: coff -link -DLL -base:0x1C000000 \ -subsystem:windows -entry:WinMainCRTStartup -machine:i386 \ -out:MYDLL.dll MYDLL.obj nafxcw.lib libc.lib ntdll.lib \ winspool.lib kernel32.lib user32.lib gdi32.lib comdlg32.lib \ version.lib lib -out:MYDLL.lib -def:MYDLL.def -machine:i386 (I'm suspicious that the entry point should be something other than WinMainCRTStartup. I tried LibMain, but that resulted in an unresolved external.) The real problem surfaces when I try to link the EXE that uses the DLL. The linker does not find the exported function in MYDLL.LIB. Here is the linker call for the EXE: coff -link -subsystem:windows -entry:WinMainCRTStartup \ -out:MYEXE.exe MYEXE.obj nafxcw.lib libc.lib ntdll.lib \ winspool.lib kernel32.lib user32.lib gdi32.lib comdlg32.lib \ version.lib MYDLL.lib I verified that the MYDLL.lib export library exists and contains the correct name for the exported function (declared _extern "C" to avoid name mangling). Clearly the linker is not finding that exported function name, but I cannot figure out why. Any ideas? Jim Conger There is 1 Reply. #: 16330 S14/Win32s Specific 02-Nov-92 09:17:22 Sb: #16284-MFC / DLL Fm: Lee Hart [Microsoft] 76150,2536 To: Jim Conger 73220,324 (X) (1) this question belongs in Section 7 (Tools-Microsoft). Please direct any followup to there. (2) It appears that you are calling LIB incorectly - you must specify all of the .OBJ files on the LIB command line. This is a common problem. Lee Microsoft Developer Support There is 1 Reply. #: 16425 S14/Win32s Specific 02-Nov-92 19:11:06 Sb: #16330-MFC / DLL Fm: Jim Conger 73220,324 To: Lee Hart [Microsoft] 76150,2536 (X) OK, I'll send to the request to section 7 - thanks. #: 16283 S14/Win32s Specific 31-Oct-92 16:42:21 Sb: WM_SETHANDLE Fm: KENNETH R SCHROCK 70621,1521 To: Bruce Ramsey/Microsoft 70324,2742 (X) My NT editor has a problem when running on 3.1 using Win32s. It will load the file, but it doesn't show in the edit window. Same for save. It will save a file of 0 bytes. Seems to be a problem with the WM_SETHANDLE and WM_GETHANDLE messages or memory. There is 1 Reply. #: 16404 S14/Win32s Specific 02-Nov-92 16:34:54 Sb: #16283-WM_SETHANDLE Fm: Lee Hart [Microsoft] 76150,2536 To: KENNETH R SCHROCK 70621,1521 (X) >>>My NT editor has a problem when running on 3.1 using Win32s. It will load the file, but it doesn't show in the edit window. Same for save. It will save a file of 0 bytes. Seems to be a problem with the WM_SETHANDLE and WM_GETHANDLE messages or memory. Page 12 of the Win32s Programmer's Reference clearly states that WM_SETHANDLE and WM_GETHANDLE are not supported. Lee There is 1 Reply. #: 16440 S14/Win32s Specific 03-Nov-92 01:56:08 Sb: #16404-WM_SETHANDLE Fm: KENNETH R SCHROCK 70621,1521 To: Lee Hart [Microsoft] 76150,2536 (X) I keep saying I'm going to read things before I start. Thanks. #: 16470 S14/Win32s Specific 03-Nov-92 09:53:26 Sb: Unverisal Thunk Fm: Craig S. Ellis 70444,50 To: ALL To All, In the back of the October WIN32 release notes was a proposal for a universal Thunk (UT). I would just like to take this time out to agree with this proposal and would very much welcome this enhancement in the next release of Win32s. In short, I believe the UT will provide developer's a bridge to more to the 32 bit API without being tied down by the device driver suppler's time schedule. When does Microsoft plan to make UTs available to the public? Craig S. Ellis Reuters Information Technology #: 16443 S14/Win32s Specific 03-Nov-92 04:56:57 Sb: WINDOWS NT AND NEC-CD Fm: RICHARD VEZINA 75140,1243 To: SysOp (X) Problem installing NT, i have a NEC CD-ROM model CDR-74 with a NEC SCSI controller, unfortunately NT does not recognize the NEC SCSI. Is there a way to make it work or do I have to find a supported SCSI controller ? Thanks. Richard There is 1 Reply. #: 16456 S14/Win32s Specific 03-Nov-92 09:00:54 Sb: #16443-WINDOWS NT AND NEC-CD Fm: John Oellrich [AT&T] 72611,1452 To: RICHARD VEZINA 75140,1243 (X) Richard, The NEC host adaptor (which is a Trantor unit) is not supported by this NT release (Trantor is working on updating their driver). Presuminng your setup works under DOS, you can install NT. Boot DOS, log in to the CD drive. CD in to the i386 directory and execute WINNT.EXE. Or of course you can buy a supported host adaptor. John There is 1 Reply. #: 16472 S14/Win32s Specific 03-Nov-92 10:41:16 Sb: #16456-WINDOWS NT AND NEC-CD Fm: RICHARD VEZINA 75140,1243 To: John Oellrich [AT&T] 72611,1452 (X) Thank you John for the info, i'll give it a try. Thanks Richard #: 16315 S14/Win32s Specific 01-Nov-92 17:25:50 Sb: #16202-Win32s fails Fm: Mark Gibbons 76216,1032 To: Robert Zeff 70323,1251 (X) There's a workshop on Windows for Workgroups tomorrow that conflicts with major (read mandatory) staff meeting at my place of employment. It will be going on less than one hundred yards from me! What exactly is Windows for Workgroups, and where can I find out more? (And WfW has meant Word for Windows to a lot of people up to now). Thanks, Mark There are 2 Replies. #: 16316 S14/Win32s Specific 01-Nov-92 17:38:43 Sb: #16315-Win32s fails Fm: Robert Zeff 70323,1251 To: Mark Gibbons 76216,1032 (X) WFW is a networking version of Win3.1, but with Email, network DDE, scheduling, + much more. -Robert #: 16563 S14/Win32s Specific 04-Nov-92 02:23:50 Sb: #16315-Win32s fails Fm: Andreas Klein 100063,1637 To: Mark Gibbons 76216,1032 (X) Windows for Workgroups is Windows 3.1 plus peer networking capabilities (you don't need a "big" server) featuring directory and printer sharing (and using) as well as clipboard sharing and Network DDE. For more information you may want to contact your dealer. He should have a overview sheet fot this product. >> (And WfW has meant Word for Windows to a lot of people up to now).<< No, that's W4W . ciao, Andreas #: 16564 S14/Win32s Specific 04-Nov-92 02:27:06 Sb: #16202-Win32s fails Fm: Andreas Klein 100063,1637 To: Robert Zeff 70323,1251 (X) No Robert unfortunately there will some problems persist when you try to use RC2. If you already checked that Win32s _installed_ ok and the device= line is in system.ini you can only try to disable the Server VxD's and try out if Win32s then runs ok. Otherwise I think you're out of luck for now. Sorry. ciao, Andreas #: 16407 S14/Win32s Specific 02-Nov-92 16:35:09 Sb: Win32s - C7 link problem Fm: Lee Hart [Microsoft] 76150,2536 To: Eric V. Smith 70621,3367 (X) I am looking into this. I take it that this happens every time you run the C7 link? Lee There is 1 Reply. #: 16430 S14/Win32s Specific 02-Nov-92 19:56:05 Sb: #16407-Win32s - C7 link problem Fm: Eric V. Smith 70621,3367 To: Lee Hart [Microsoft] 76150,2536 (X) Lee, Yes, this happens every time I run C7's link. Eric. There is 1 Reply. #: 16612 S14/Win32s Specific 04-Nov-92 11:29:45 Sb: #16430-Win32s - C7 link problem Fm: Lee Hart [Microsoft] 76150,2536 To: Eric V. Smith 70621,3367 Thanks for the bug report. The problem has been reproduced here (sorry for the delay, but those of us who are supporting Win32 tend to have Win32 development tools installed and not C7) and our developers are looking into the problem. As I get more information on possible fixes, etc. I will post them here. Lee #: 16405 S14/Win32s Specific 02-Nov-92 16:34:59 Sb: #16259-Win32s: Exception ? Fm: Lee Hart [Microsoft] 76150,2536 To: Robert Zeff 70323,1251 (X) >>>Also, I can get the same message (Different IP, of course) with freecell if I select "select game" twice. I am looking into this. Thanks. >>>Another question: Should a console program work with Win32s? Since Windows 3.1 does not have a Console API for Windows Apps, no, Win32s does not currently support this. Lee There is 1 Reply. #: 16619 S14/Win32s Specific 04-Nov-92 13:14:36 Sb: #16405-Win32s: Exception ? Fm: Lee Hart [Microsoft] 76150,2536 To: Lee Hart [Microsoft] 76150,2536 (X) To: Everyone who has been having problems with Unhandled Kernel mode exceptions in Win32s. We have identified the problem and we are currently internally testing a fix to this problem. We hope to be able to make this fix available in the near future (the problem is that try/except has problems in Win32s in the October Release). I do not have an exact timetable available at this time. Lee Microsoft Developer Support #: 16634 S14/Win32s Specific 04-Nov-92 14:08:47 Sb: Win32s Doc Errors Fm: Lee Hart [Microsoft] 76150,2536 To: Al Longyear 70165,725 >>>1. On page 7, it states "* Win32 applications see a flat 32-bit address space (CS == DS == SS == ES)". This is technically impossible for the 80386/80486 processor. The CS register can *NOT* be the same as the DS/SS/ES/FS/GS registers. The CS must be loaded with a segement of type "code" while the "DS/SS, etc" register must be loaded with segments of type "data". (They may have the same limits and base addresses and would threfore have the the same phyical byte location at the same offset. But they can *NOT* be identical.) Good point. I have raised a bug against the documentation to try to make this as clear, yet accurate, as possible. >>>2. On page 15, it states "* Local variables (stack variables) are faster to pass by reference between the Win32 applciation and the system". Faster than what? Is it faster than a GlobalAlloc'ed variable? Is it faster than a Win16 stack variable? My understanding is that because Win32s must convert between the flat address space of the Win32 app and the segmented addess space of Windows 3.1, there is an overhead associated with this. I believe the reason local variables are faster is that this conversion has already been done for the stack and so less overhead is required. Memory space GlobalAlloc's by a Win32 application under Win32s would need to have the tiled selectors set up in order to be passed to any 16 bit app (or any 16 bit portion of the Operating System.) I am gettting more information on this. Thanks for the feedback (and the interesting questions.) Lee Microsoft Developer Support #: 16493 S14/Win32s Specific 03-Nov-92 14:52:03 Sb: Installation Crash Fm: Sandy Bernstein 72477,1132 To: Dwight Matheny I am having trouble installing Windows NT. During installation, I get the following message before my computer crashes: Microsoft Windows NT - Beta October 1992 [16064 Kb Available Memory] *** Fatal System Error: 0x00000069 *** Phase 1 I/O Initialization Failed Kernel Debugger using Com2 What does this mean? How do I fix the problem? I am also getting a message about a problem with my EISA motherboard. AMI told me to download 2 files from their BBS: !AMI16B1.CFG & !AMI28A1.CFG. Where should I put them? Is there anything I need to do to get Windows NT to recognize them? Thanks for your help. There is 1 Reply. #: 16644 S14/Win32s Specific 04-Nov-92 15:55:59 Sb: #16493-Installation Crash Fm: Lee Hart [Microsoft] 76150,2536 To: Sandy Bernstein 72477,1132 (X) >>>I am having trouble installing Windows NT. During installation, I get the following message before my computer crashes: Error 0x69 - you have an unsupported disk controller, or your partitions are confusing Windows NT. For more assistance please go to the WINNT forum, section 3 (Windows NT Setup). The .CFG files are just EISA config files - once you install them using the EISA config program that came with your machine you should be OK. If your EISA problem is that EISA memory and ISA memory do not match, then you need to run your EISA config program to update your memory settings, and run your BIOS setup to update the memory settings there. Again, please post any followup to WINT Section 3. Lee Microsoft Developer Support #: 16566 S14/Win32s Specific 04-Nov-92 03:44:11 Sb: CD ROM Compatibility Fm: Douglas Badin 71332,1573 To: SYSOP (X) I am interested in buying TEXEL's CD ROM model DM3024 with a Trantor T130. Is this going to be added to the NT Hardware Compatibility List soon? There is 1 Reply. #: 16645 S14/Win32s Specific 04-Nov-92 15:56:03 Sb: #16566-CD ROM Compatibility Fm: Lee Hart [Microsoft] 76150,2536 To: Douglas Badin 71332,1573 (X) >>>I am interested in buying TEXEL's CD ROM model DM3024 with a Trantor T130. Is this going to be added to the NT Hardware Compatibility List soon? Questions about Hardware Support and the HCL belong in the WINNT forum, Section 8, H/W Compatibility. You should also contact the Hardware vendor for information. Lee Microsoft Developer Support #: 16630 S14/Win32s Specific 04-Nov-92 13:40:41 Sb: CD-ROM Compatibility Fm: Craig Salyers 70643,23 To: SYSOP (X) I have a Sony CDU6205-10 CD-ROM. Is this going to be added to the Hardware Compatibility List in the near (or distant) furture? There is 1 Reply. #: 16646 S14/Win32s Specific 04-Nov-92 15:56:09 Sb: #16630-CD-ROM Compatibility Fm: Lee Hart [Microsoft] 76150,2536 To: Craig Salyers 70643,23 >>>I have a Sony CDU6205-10 CD-ROM. Is this going to be added to the Hardware Compatibility List in the near (or distant) furture? I don't know. If it's SCSI, maybe - ask in WINNT Section 8 (I deal with Win32s here, not hardware - I don't know why I'm getting all of these HCL questions - can you tell me why?). If it's not SCSI, ask Sony. Lee Microsoft Developer Support #: 16499 S14/Win32s Specific 03-Nov-92 15:09:46 Sb: NDIS support in Win32s Fm: Joe Erickson 71360,466 To: All Will Win32s support the Ndis calls documented in the Win32 DDK? If this is not in the current plan is there any way to get Microsoft to reconsider? It's quite important in deciding whether to do a new network stack for Win32s or Windows 3.1. Thanks Joe Erickson There is 1 Reply. #: 16658 S14/Win32s Specific 04-Nov-92 16:17:54 Sb: #16499-NDIS support in Win32s Fm: Lee Hart [Microsoft] 76150,2536 To: Joe Erickson 71360,466 (X) >>>Will Win32s support the Ndis calls documented in the Win32 DDK? I believe you misunderstand Win32s. Win32s is to allow user mode (aka ring 3 code on x86 machines) Win32 programs to run under Windows 3.1. As a result, the APIs, etc. from the Windows NT DDK do not apply. If you wish to do a network stack for a Win32 application running under the Win32s extensions for Windows 3.1, you really want to do a Windows 3.1 (normal or Windows for Workgroups) network stack, and let the Win32 app call the normal WNet APIs. If I misunderstood your question, I apologize and I look forward to more information. Lee Microsoft Developer Support #: 16788 S14/Win32s Specific 05-Nov-92 19:06:40 Sb: INSTALL PROBLEMS Fm: Steve Woolley 75570,1620 To: ALL HELP AGAIN, THE LAST TIME I WROTE WAS MESSAGE #16128 SINCE THEN, I MADE A NEW FLOPPY FROM THE CD AND TRIED TO REINSTALL WIN32S ON TOP OF WINDOWS 3.1 I REMOVED STACKER, I HAVE 4M RAM AND PLENTY OF HARD DISK SPACE I GOT FARTHER IN THE INSTALLATION BUT NOW I AM RECEIVING THE FOLLOWING ERROR: MICROSOFT TEST DRIVER(DS) ERROR PARSETIME ERROR: DUPLICATE DEFINITION SETUPAPI.INC(350) WHAT SHOULD I TRY NEXT ? THANKS FOR YOUR HELP STEVE #: 16811 S14/Win32s Specific 06-Nov-92 05:34:03 Sb: WIN32s Fails Fm: Eric M. Guli 70272,3627 To: Lee Hart/ 76150,2536 (X) Thanks for your reply. Yes, the error occurs on install after what seems to be the point where setup is copying all the files to the hard disk. This is on the 486DX/2 50 Eisa machine that this error occurs. No, I do not have a kernal debugger for Windows 3.1 installed I am only working from the new October release of NT. Would Nortons Desktop cause any problems with install? However, I did strip Windows down to just program manager and still received the same error. I also have a WINJET 800 installed for my printer driver running through a HPIII, but I also replaced that with the straight HPIII driver and the error still happens. What can I do to get WIN32s running? I don't want to at this point run NT as my operating system because this is a work machine that I cannot totally experiment with. I guess the question is do I need NT to run WIN32s? Thanks, Eric Guli #: 16785 S14/Win32s Specific 05-Nov-92 18:11:49 Sb: Win32s release Fm: John Lutz 73730,3577 To: MS Will the Win32s DLLs be released separately before NT or do we have to wait for the NT release to ship Win32s apps. John There is 1 Reply. #: 16854 S14/Win32s Specific 06-Nov-92 16:06:20 Sb: #16785-Win32s release Fm: Lee Hart [Microsoft] 76150,2536 To: John Lutz 73730,3577 (X) >>>Will the Win32s DLLs be released separately before NT or do we have to wait for the NT release to ship Win32s apps. I believe you can, but let me double check. One fact you must keep in mind is that it is quite likely that there will be some change in the Win32 system DLLs anytime before we ship that would require that you re-build your app in order to work. For example, we have been moving system APIs around in the DLLs to improve performance. This means that your older apps that try to dynalink with them in the wrong place will not load. As a result, as soon as a newer version of Win32s is released, and the Win32s install program updates your customer's installation (because of the versioning info in the resources), your Win32 app will no longer load and you will need to update your customers with a newer version of your app. (Persoanlly I think that's a reasonable price to pay, as long as you are aware of it.) Lee Microsoft Developer Support #: 16786 S14/Win32s Specific 05-Nov-92 18:12:50 Sb: DLLs Fm: John Lutz 73730,3577 To: MS Can you create DLLs containing 32-bit code and use them under Windows 3.1 via Win32s? John There is 1 Reply. #: 16855 S14/Win32s Specific 06-Nov-92 16:06:25 Sb: #16786-DLLs Fm: Lee Hart [Microsoft] 76150,2536 To: John Lutz 73730,3577 (X) >>>Can you create DLLs containing 32-bit code and use them under Windows 3.1 via Win32s? Yes, if the calling app is a Win32 app. Yes, if the DLL is a standard Windows 3.1 DLL (32 bit code in 16 bit segments). You can't attach a Win32 DLL to a Win16 app (discounting using a stub Win32 app and some form of IPC to talk to the Win16 app.) So I think the answer to your question is generally no. (If I misunderstood your question please be more detailed.) Lee Microsoft Developer Support There is 1 Reply. #: 16872 S14/Win32s Specific 06-Nov-92 18:14:59 Sb: #16855-DLLs Fm: John Lutz 73730,3577 To: Lee Hart [Microsoft] 76150,2536 (X) Lee, What we would like to do is develop our application completely on NT then use Win32S to run on Windows 3.1. The reason I asked about DLLs is because we would like to make our application extensible. For example, we would like to be able to put data readers for different file formats in different DLLs. Each DLL would contain the same set of routines, but would work on different file formats. The application needs to be able to sense which DLLs are installed at runtime and load them dynamically. Our goal is to be able to extend the application with new DLLs (readers) after it has been shipped. Given that our application will be completely 32-bit and that we need to run on both Windows 3.1 and NT... 1) is this possible and 2) are there any limitations that I need to be aware of? Thanks, John #: 16274 S15/Unicode/NLS 31-Oct-92 09:17:30 Sb: Complete unicode font? Fm: GMS 100063,2012 To: Sheldon Fox 70162,3422 (X) One more STRONG vote for full language support (minimum English/German/Frensh/Spanish/Japanese/Chinese/Thai) in one NT version. G. M. Sontheimer There is 1 Reply. #: 16285 S15/Unicode/NLS 31-Oct-92 18:04:03 Sb: #16274-Complete unicode font? Fm: Paul Watson [XLsoft] 76056,1751 To: GMS 100063,2012 (X) The difficulty of doing this depends on what you mean by "full" support. Franky, I don't expect the base system to have elegant input methods for every language. However, I do expect to be able to buy and add input methods for any language from Microsoft and third party vendors. Yes, the base system should have -some- way of entering any script, even if we have to resort to entering hex values. Yes, the base system should include one font that can be render all characters of the defined scripts. Regards...Paul There is 1 Reply. #: 16754 S15/Unicode/NLS 05-Nov-92 14:43:44 Sb: #16285-Complete unicode font? Fm: GMS 100063,2012 To: Paul Watson [XLsoft] 76056,1751 My "dream support" is: One OS with a Unicode Font for all languages I mentioned included. I would like to use the same source code for all language versions and put all text strings in string resources. If key combinations are needed for entering them this would be bearable, but preferrably I would youst like to plug in my German (or Japanese) keyboard, change the keyboard driver, and add the text strings needed. The application would detect the country settings of Win NT and choose the string resources for the appropriate language... #: 16751 S15/Unicode/NLS 05-Nov-92 14:27:22 Sb: Why no WinMainW? Fm: David Van Camp 70740,366 To: Microsoft Why is there no unicode and transparent versions of WinMain? How utterly annoying! dvc There is 1 Reply. #: 16826 S15/Unicode/NLS 06-Nov-92 11:08:50 Sb: #16751-Why no WinMainW? Fm: Steve Firebaugh [MS] 75430,412 To: David Van Camp 70740,366 David Van Camp, >Why is there no unicode and transparent versions of WinMain? Certain run time objects (e.g. window classes) are marked as unicode or ansi. When the system passes them parameters, it checks what type of string they expect, and then converts the strings appropriately. An executable which has just been loaded has no way of specifying whether it expects unicode or ansi strings. For this reason, the system always passes ansi parameters to WinMain(). This may change at some point in the distant future, but it will not change before Windows/NT release 1.0. The suggested work around is to call GetCommandLine() if you want the command line parameters in unicode form. Steve Firebaugh #: 16750 S15/Unicode/NLS 05-Nov-92 14:27:20 Sb: PROC_XXX macros & Unicod Fm: David Van Camp 70740,366 To: Microsoft The PROC_EXTDEVICEMODE, etc., macros in drivinit.h are define as MAKEINTRESOURCE(xx) which is incorrect for unicode because GetProcAddress has no unicode version and MAKEINTRESOURCE uses a (LPTSTR) cast. The best solution (in my opinion) would be to make a unicode version of GetProcAddress, but, barring that, an alternate ASCII-compatible MAKEINTRESOURCEA() macro is needed: #: 16827 S15/Unicode/NLS 06-Nov-92 11:08:57 Sb: #16750-PROC_XXX macros & Unicod Fm: Steve Firebaugh [MS] 75430,412 To: David Van Camp 70740,366 David Van Camp, You are correct, there is an error in the header files with the definition of MAKEINTRESOURCE. It ought to be transparent such that it is mapped to either an ansi or a unicode cast. I have reported this bug, and we will try to get it fixed by the next release. Thank you for pointing it out. Steve Firebaugh #: 16696 S15/Unicode/NLS 04-Nov-92 22:42:36 Sb: Unicode notepad font Fm: James D. Howard 71172,2010 To: SECTION 15 The default font used by the Unicode notepad appears to have some problems. The glyphs for "i", "I", and "l" (lower-case L) are indistinguishable. The problem with "i" and "I" is, I'd guess, a bug, but the problem with "I" and "l" is a font design issue. Could the default font be selected or modified, please to assure that the glyphs are always distinguishable. (For example IBM's RS/6000 spoils a very nice on-line help facility by likewise selecting a Helvetica-like font that makes certain command line options undecipherable.) There is 1 Reply. #: 16828 S15/Unicode/NLS 06-Nov-92 11:09:12 Sb: #16696-Unicode notepad font Fm: Steve Firebaugh [MS] 75430,412 To: James D. Howard 71172,2010 James Howard, There is a font on the compact disk named uclucida.ttf. If you install this font, I believe that you will see the problem you have reported go away. The 'unipad' accessory expects to use this font. When the font is not installed, it still creates a logical font with the face name "Lucida Sans Unicode." The font mapper ends up creating an Arial Bold font with a size and weight set just right to screw up the i, I, and l. (I find that on my VRAM 7 display, the logfont equivalent to the font shown in unipad is: {16, 7, 0,0, 600, 0,0,0, UNICODE_CHARSET, 0,0, 2,2, "Arial"}. Almost any other size/weight for the Arial font produces glyphs for i & I at least which are distinguishable. You are correct that from the user's perspective, this is a bug. I have reported it, but I can not say for certain how it will be fixed. In any case, thank you for bringing it to our attention. Steve Firebaugh #: 16599 S15/Unicode/NLS 04-Nov-92 10:39:32 Sb: Unicode printf Fm: Carl W. Brown 71250,1322 To: David Manns 100112,2773 David, Yes there is a Unicode equivalent of "printf" it is "wsprintf". However "wsprintf" is a generic function the handles both ANSI and Unicode. The Unicode only version is "wsprintfw" Carl There is 1 Reply. #: 16752 S15/Unicode/NLS 05-Nov-92 14:27:29 Sb: #16599-Unicode printf Fm: David Van Camp 70740,366 To: Carl W. Brown 71250,1322 (X) <> Carl, Where did you get this info? There is no 'wsprintfw'. wsprintf takes unicode strings for both the output buffer and the fmt string. For *ANY* of the NT printf functions - unicode or otherwise - the fmt string expects the following parameters: %s expects an ANSI string, %ws expects a Unicode (wide) string %ts expects an ANSI string in sprintf, Unicode in wsprintf %c expects a ANSI char %wc expects a Unicode (wide) char %tc expects an ANSI char in sprintf, Unicode in wsprintf tsprintf is the 'transparent' version of sprintf. See TCHAR.H for all available 'transparent' mapping macros. Since tsprintf will map to sprintf or wsprint depending on the UNICODE flag, when this option is available, you can use transparently defined strings (TCHAR or LPTSTR) with it. Also, the Unicode equivelent of "printf" is "wprintf" not "swprintf". ------- DOC BUG NOTE: Under the WINAPI32.HLP String Functions section it says for %ts and %tc "... if wsprintf was compilied with the #define UNICODE compile flag..." Since MICROSOFT - not us - compiles this function, this info is useless. I just reported this to David McBride. ------- BUG NOTE: there is a bug in wsprintf, et. al, where if %s is specified, the resulting Unicode output string will have ASCII chars inbedded within it. E.g.: wprintf ( wcs, L"abc%s", "def" ); will return with wcs containing (in bytes): 0,'a',0,'b',0,'c','d','e',0,0 This has already been reported to MS's C-Lib development team. The work arround for now is to use only wide char strings as fmt parameters. Use mbstowcs() to convert an ANSI string before passing it to wsprintf. There are 2 Replies. #: 16792 S15/Unicode/NLS 05-Nov-92 19:32:20 Sb: #16752-Unicode printf Fm: Carl W. Brown 71250,1322 To: David Van Camp 70740,366 (X) David, I got the info from the July notes. Obviously things have changed since then. All the generic calls were supposed to test the user define of UNICODE with an #: 16830 S15/Unicode/NLS 06-Nov-92 11:09:30 Sb: #16792-Unicode printf Fm: Steve Firebaugh [MS] 75430,412 To: Carl W. Brown 71250,1322 Carl, The information you got from the July release notes was correct, even though the implementation of wsprintf() in that release was not handling unicode correctly. It is working correctly in the October release. >The other important API is the RegisterWindowClassW/RegisterWindowClassA which >does not seem to be in the Win32 API. There is a RegisterClassW and a RegisterClassA in the Win32 API. They are both working correctly. Steve Firebaugh #: 16829 S15/Unicode/NLS 06-Nov-92 11:09:21 Sb: #16752-Unicode printf Fm: Steve Firebaugh [MS] 75430,412 To: David Van Camp 70740,366 David Van Camp, >There is no 'wsprintfw'. There is in fact a wsprintfW() as part of the Win32 API. In the October beta it is working correctly. Steve Firebaugh #: 16671 S15/Unicode/NLS 04-Nov-92 18:39:22 Sb: Unicode Addendum 1.0.1 Fm: Steve Firebaugh [MS] 75430,412 To: All I just uploaded a zipped copy of a text file which contains the "Unicode Addendum 1.0.1." (library 15, addend.zip) I have included excerpts of the file below to give people an idea of its contents. Steve Firebaugh ====================================================================== Unicode 1.0.1 Addendum 92.11.03 8:52 UNICODE 1.0.1 The following document is an ASCII version of the Unicode 1.0.1 addendum, which has been added to Volumes 1 and 2 of The Unicode Standard. Because the formatting has been lost and the original text contains nonASCII characters, a dollar sign is used as a placeholder instead, and the text has been modified slightly for readability. .. 1. Introduction As discussed in Volumes 1 and 2, small changes have been made to Unicode 1.0 in order to incorporate it into the international character encoding standard, ISO 10646, which was approved by ISO as an International Standard in June, 1992. The Unicode Consortium plans to issue Unicode 1.1 in early 1993. The character content and encoding will be identical to that of ISO 10646. To that end, Unicode 1.1 will include approximately 5,400 additional characters from ISO 10646 that are not already in Unicode 1.0. In order to expedite use of Unicode in the interim, the Unicode Consortium is issuing an intermediate version, Unicode 1.0.1, which consists of Unicode 1.0 modified by the changes necessary to make the character codes a proper subset of ISO 10646. This paper describes the differences between Unicode 1.0.1 and Unicode 1.0 (for more information, see Volume 1, pp. xix-xx and Volume 2, pp. 4-9 and 427-431). Implementations that use Unicode 1.0.1 as thus defined will be completely compatible with Unicode 1.1, and therefore fully compatible with ISO 10646. There is 1 Reply. #: 16722 S15/Unicode/NLS 05-Nov-92 10:13:15 Sb: #16671-Unicode Addendum 1.0.1 Fm: Steven Olson 70313,2246 To: Steve Firebaugh [MS] 75430,412 (X) Steve: Just a short note to say "Thanks" for uniput.zip. I realize this is not a final project but it does work. Any idea on when uclucida.ttf will be complete (or at least expanded)? Thanks Again! Steve Olson CRA There is 1 Reply. #: 16866 S15/Unicode/NLS 06-Nov-92 16:58:49 Sb: #16722-Unicode Addendum 1.0.1 Fm: Steve Firebaugh [MS] 75430,412 To: Steven Olson 70313,2246 Steve, Thank you for the feedback. I have been trying to nail down a firm schedule for uclucida.ttf, and no one will commit to one. The current (soft) goal is to have 1400 glyphs covered by the first release. As I understand it, this particular font will never be expanded to cover all of the unicode glyphs. Someone pointed out to me that the final NT o.s. will be distributed on floppy disks. A font large enough to cover all unicode glyphs will certainly not fit on one floppy disk. For this reason, people are reconsidering the font model, such that multiple font files, which cover disjoint unicode code point ranges, may be combined into a single logical run time font. Steve Firebaugh #: 16282 S16/Tools-3rd Party 31-Oct-92 15:38:54 Sb: SlickEdit won't install Fm: Len Gray 71630,1703 To: All I ordered SlickEdit yesterday from MicroEdge - requested Saturday delivery so I could get familiar over the weekend. It arrived bright and early this morning! When I go to the B: prompt and type NTINST, the disk hums and it looks like something is happening, but soon the B: prompt returns. :-( Has anyone been here before? I sure would appreciate any help in getting me up and running before Monday! Thanks in advance! Len There are 2 Replies. #: 16286 S16/Tools-3rd Party 31-Oct-92 19:15:29 Sb: #16282-SlickEdit won't install Fm: Peter Handsman[Inmark] 70550,2570 To: Len Gray 71630,1703 (X) Len, I got the october update of Slick on friday, it works and installs fine for me. ntinst.cmd is a batch file, maybe you should try to run the command inside by hand. Peter #: 16313 S16/Tools-3rd Party 01-Nov-92 17:10:31 Sb: #16282-SlickEdit won't install Fm: Robert Reichel [MS] 72360,3504 To: Len Gray 71630,1703 (X) Call MicroEdge tech support. I've frequently gotten ahold of them on weekends and late at night. The guy you will talk to is the author of SlickEdit, so he will definitely be able to help you sort things out. #: 16561 S16/Tools-3rd Party 04-Nov-92 00:05:44 Sb: MEWN11.ZIP Fm: Pierre Perret 73757,2337 To: Sysop (X) I am uploading MEWN11.ZIP to LIB 1. This file replaces the older MEWN11.ZIP that was available since mid-October. The old file contained update 1.1 of MicroEMACS for Windows, which had been compiled for the July PDK. The new file contains update 1.1a, which is compiled for the Windows NT October BETA. (in addition, update 1.1a contains a minor bug fix and makefile changes to support DEC Alpha machines (only i386 and MIPS were supported by update 1.1) - Pierre Perret - #: 16769 S16/Tools-3rd Party 05-Nov-92 16:39:52 Sb: 3rd party NFS on NT? Fm: Alex Bronstein 75070,2452 To: all Has anybody found a usable NFS client that can be run on NT? We're changing the source storage strategy in our shop in a way which would require daily access to an NFS server with some shell scripts. Right now I have client NFS on DOS (via Sun's PC-NFS) but no great way to do the scripts. Or I can have Hamilton C-shell on NT + associated tools, but no NFS client... I guess pointers to good shell (Bourne or C) on DOS would also be appreciated. Thanks, Alex (aka alex@gain.com) #: 16841 S16/Tools-3rd Party 06-Nov-92 13:46:02 Sb: MicroEmacs Fm: Judy Polonsky 70544,2411 To: Pierre Perret I had been using MicroEmacs successfully with the July SDK. I have now installed the October SDK and DDK and reloaded the system with Microemacs (memacs.zip , 196507 bytes, 8-22-92 - containing 7 files). When I click on the Microemacs icon, I get a message saying "The Dynamic Link Library USERRTL.dll could not be found in the specified path. C:\emac;c:;c:\winnt\system;c:winnt;c:\winnt\system" The file userrtl.dll is nowhere on the disk. Any suggestions on what might be wrong? I hate using MEP! Judy Polonsky, BusLogic #: 16360 S17/Unmonitored Chat 02-Nov-92 12:35:53 Sb: CDROM (UK) Fm: Christian Beaumont 100034,2326 To: All Has anyone in the UK received the October release of the CDROM yet? Or should I start pestering people? Christian There is 1 Reply. #: 16364 S17/Unmonitored Chat 02-Nov-92 13:45:50 Sb: #16360-CDROM (UK) Fm: Michael Webb 100014,1424 To: Christian Beaumont 100034,2326 (X) I called Microsoft UK - they said that they expect arrival of shipment in UK this week and will ship to UK users within 24 hours! Mike Webb There is 1 Reply. #: 16442 S17/Unmonitored Chat 03-Nov-92 03:34:06 Sb: #16364-CDROM (UK) Fm: Christian Beaumont 100034,2326 To: Michael Webb 100014,1424 Great, I'll look forward to receiving that but I'll probably be at the DDK Conference so I won't get it until next week (boo) Christian #: 16533 S17/Unmonitored Chat 03-Nov-92 18:56:02 Sb: CDROM->SCSI-2? Fm: Bruce Ramsey/Microsoft 70324,2742 To: Ben Laurie 100014,1235 (X) Hi Ben - >Ben> Wow! Bugfix notifications! Will wonders never cease? So long as the release notes mention it, or someone reminds me, I'll happily test it. Its only 1 jumper, after all >Bruce> Thanks again for the offer to test - I'll try to remind you :-) We're testing here too, yet are interested in whether you find it's fixed in your environment... I don't know if you have your copy of the October build yet, but in any case the October build would be the build to try this on again to see if the spin-up problem is fixed in your environment I don't think the release notes comment on this, so this is your reminder :) Bruce #: 16339 S17/Unmonitored Chat 02-Nov-92 10:25:57 Sb: Bad Int. Support Fm: Sigurdur St. Hjalmarsson 100063,1446 To: Ulrich Paul 100022,3315 My sentiments exactly - they have what is commonly known as an ethno-centric view of the world: there is the USA, and then there is the rest of the world! There is 1 Reply. #: 16565 S17/Unmonitored Chat 04-Nov-92 02:28:36 Sb: #16339-Bad Int. Support Fm: Peter A Winskill 70323,2547 To: Sigurdur St. Hjalmarsson 100063,1446 The rest of the world...? Oh, I guess you mean Virginia! #: 16291 S17/Unmonitored Chat 01-Nov-92 06:16:33 Sb: CDROM delays Fm: Mike Walsh (Helsinki) 72557,3170 To: ALL Here is a list of promised developer CDROMs with dates promised and dates received. It is intended to show the time delays faced by some developers. 1. July CDROM: promised San Francisco conf. July 1992 received San Francisco conf. July 1992 2. July CDROM (Finland): promised Helsinki conf. 14th Sept 1992 received October 7th 1992 3. Dev. Network CDROM beta 1 (ex-US to San F. participants) promised September 1992 not arrived Nov 1 1992 (also the letter asking for address changes has not yet arrived) 4. Dev. Network CDROM beta 1 Scandinavia questions on this emailed to MS Dev Net rep 10th October - no answer received by 1 Nov 1992. (email to Scan. rep not US) 5. European Dev. CDROM (as issued at Brussels NT conf. in August (?)) promised by MS Finland on receipt of a completed questionaire questionaire sent 9 Oct. 1992; no CDROM received by 1 Nov 1992. To be fair here is the present state of the latest OS/2 Developer's CDROM a. European DAP: CDROM "sent to members" 20 Oct. 1992; no CDROM received by 1 Nov 1992. b. Worldwide OS/2 DAP (via Compuserve): initial application made on 18th October (first day); later found out that all applications made then were being ignored and had to re-apply; re-applied 25th October; no confirmation of acceptance made by 1 Nov 1992. Mike (who hopes this is of interest to someone out there) There is 1 Reply. #: 16616 S17/Unmonitored Chat 04-Nov-92 12:23:01 Sb: #16291-CDROM delays Fm: George Meng 76711,421 To: Mike Walsh (Helsinki) 72557,3170 (X) Mike, Thanks for the data. I want you to know that we hear your comments, and we're definitely trying to improve the process of delivering materials to customers both in the U.S. and, especially, to customers in other countries. George Meng MS Developer Network #: 16329 S17/Unmonitored Chat 02-Nov-92 09:17:18 Sb: ISV Dev'pment Lab Fm: Lee Hart [Microsoft] 76150,2536 To: Dag Baardsen [PS Norway] 100010,2257 (X) I can give you some tips from the other side - I recently finished working in the ISV Porting Lab here and here are some common mistakes that reduce the effectiveness of the lab: (1) ISVs assume that all equipment will be provided. We do plan on having one machine per office, but even this is not always possible. If you can bring your own equipment you will be better off. (2) ISVs have not done a lot of work prior to visiting the lab. Give the fact that muchof Win16->Win32 porting is mechanical, you are not getting the maximum benefit of the lab if you arrive with lots of mechanical porting to be done. If you can at least compile fairly cleanly when you arrive it will give you a better opportunity to port your application to Win32 and then on to the more strict MIPS compiler. If you are interested in porting to MIPS, it is to your benefit to become familiar with some of the more common problems experienced when portont to MIPS (for example use of either old-style or ANSIfunction declarations, and corresponding comment format (/* .. */ or //) Lee P.S. It doesn't hurt to sharpen your Darts skills either ;-) #: 16344 S17/Unmonitored Chat 02-Nov-92 10:53:31 Sb: ISV Dev'pment Lab Fm: John Hall [MS SDE] 70750,2341 To: Dag Baardsen [PS Norway] 100010,2257 (X) To Lee's comments, I would add: Since you are an OS/2 -> Windows translation, it will help a great deal if you have run the basic OS/2 -> Windows translation tools on all of your source (sed script on .C and .H files, ms_ssed on .C files) Next thing: create .def files for all of your dll's and exe's modeled after the appropriate template. Making 'stub functions' for dll's can be useful. For each dll, create a notimp.c which has each function in the dll, with a MessageBox() that puts a "This funtion not implemented" on the screen. This makes it easy to iteratively add functionality. Things go much quicker if you can just get the main window up, then start adding things. This approach helps that. If you use shared memory, study and understand the shrmem example you have been given. Finally, if you aren't sure about any of this contact your MS rep Magnus. Thanks, John Hall, The OS/2 -> Windows guy and your host while in the Porting Lab. I'd be happy to take questions before you get here in the OS/2 -> Windows forum. There are 2 Replies. #: 16479 S17/Unmonitored Chat 03-Nov-92 11:38:29 Sb: #16344-ISV Dev'pment Lab Fm: Alex Bronstein 75070,2452 To: John Hall [MS SDE] 70750,2341 (X) Have you had any ISV in your lab porting X/Unix apps? Can you comment (without company names of course) on the experience? Thanks, Alex There is 1 Reply. #: 16690 S17/Unmonitored Chat 04-Nov-92 21:46:03 Sb: #16479-ISV Dev'pment Lab Fm: John Hall [MS SDE] 70750,2341 To: Alex Bronstein 75070,2452 (X) >> Have you had any ISV in your lab porting X/Unix apps? Can you >> comment (without company names of course) on the experience? Yes, we have, though many of them have been non-graphical server type applications. In fact, it is still an open question whether an ISV with both a Unix and an OS/2 1.x product should start from the Unix base or the Os/2 base. For non-graphical types, I lean toward the Unix base at the moment. Many other Unix ports are happening that have not come through the lab. It is really exciting -- the same people who told us to 'go have congress with ourselves' a few years ago when we wanted them to port to that *other* system are now quite busy on Windows NT. However, due to the politics of the Unix market many of these people have been adamant about *not* releaseing their names -- a deviation from most other ISV's. The major questions I have for Unix ports on a base level is: Do you always follow a fork with an exec or exit (vfork)? Yes is good. Where/why do you use Signals? Just keeping track of children dying is OK. Other problems crop up but they don't bother me at night. This discussion should move to the 'Porting From Unix' section. #: 16562 S17/Unmonitored Chat 04-Nov-92 01:53:14 Sb: #16344-ISV Dev'pment Lab Fm: Dag Baardsen [PS Norway] 100010,2257 To: John Hall [MS SDE] 70750,2341 (X) Lots of thanks to you both, I'll come back to you later. Most of our questions will go through Magnus, thougmight (just *might*) have others... Dag There is 1 Reply. #: 16691 S17/Unmonitored Chat 04-Nov-92 21:46:42 Sb: #16562-ISV Dev'pment Lab Fm: John Hall [MS SDE] 70750,2341 To: Dag Baardsen [PS Norway] 100010,2257 (X) I hang out on 'Porting from OS/2'. See you there. #: 16822 S17/Unmonitored Chat 06-Nov-92 10:53:45 Sb: shareware under nt Fm: Scott Alexander 76556,557 To: Len Gray 71630,1703 (X) Instead of checksumming your dialog resource, you should consider reviewing the registry API's for NT. You can store your copyright and registration data in your registry entry, along with date installed, etc. The registry data is free form , so you could encrypt or checksum your area to prevent tampering. You could even encrypt your key when registering if you were worried about people finding the data in the registry. MS is encouraging the use of the registry for data like this. There is 1 Reply. #: 16837 S17/Unmonitored Chat 06-Nov-92 12:36:36 Sb: #16822-shareware under nt Fm: Len Gray 71630,1703 To: Scott Alexander 76556,557 (X) I agree, the registry is the best place for sensitive strings like copyright and registration data. A solicitation dialog, however, can be changed in many ways. Controls can be removed, or they can just be moved off into a non-visible area of the dialog. Or their size can be changed so the text doesn't show. All in all, the easiest way to confirm that the user is seeing and operating what you wanted them to is to have a means to checksum the dialog. #: 16746 S17/Unmonitored Chat 05-Nov-92 13:30:31 Sb: Seeking NT apps Fm: Robert Lauriston 75300,3205 To: all I'm writing another magazine article previewing NT and would like to acquire beta copies of any major-category Win32 apps (spreadsheets, word processors, graphics, etc.--not utilities or programming tools). Nondisclosure basis is fine, particularly since I'm not yet writing about apps per se but rather just trying to see how Win32 NT programs differ from the usual DOS and Windows apps. Please reply by e-mail if you have an app you'd like to show me. There is 1 Reply. #: 16851 S17/Unmonitored Chat 06-Nov-92 15:33:33 Sb: #16746-Seeking NT apps Fm: Seattle Lab 71202,560 To: Robert Lauriston 75300,3205 (X) Dear Robert, We are porting a multi-user business operating system to NT from it's current UNIX environment. This port of the VMark operating system, UniVerse, could not have been considered to a single-threaded environment. NT and OS/2 are the first PC environments that can truly support substantial multi-user based applications and environments. At the DDK conference, I spoke with several others who were also looking at bringing traditional mini and main-frame apps and environs into NT. NT will probably bring a wave of seasoned and sophisticated business applications into the Microsoft fold. Hope this commentary helps, Dennis Krueger Seattle Lab There is 1 Reply. #: 16861 S17/Unmonitored Chat 06-Nov-92 16:17:46 Sb: #16851-Seeking NT apps Fm: Robert Lauriston 75300,3205 To: Seattle Lab 71202,560 (X) NT certainly seems suited for multiuser and client-server development. What I'm looking for is more end-user stuff, though. Thanks.