June 2001 - Wine-Devel - list.winehq.org

Re: PATCH: set_thread_buffer fixed
by Gerard Patel June 5, 2001

June 5, 2001

At 10:43 AM 05/06/2001 +0200, you wrote: >> This calls fails because glibc detects that the value at (esp+28) >> is different of 0 - the real system call is never done, with your >> test app and Wine. > >Well, note that the second version is only pushing a 32bit offset, so >mmap64 accesses random stack space! Yes :-/ <snip> >Yes, check the sys/mman.h header, it should have __off64_t as argument >to the mmap ... asm("mmap64") not __off_t. Yes, it is __off_t and that is the problem; but how this header has come to be in this state is part of what I don't understand. It's marked as part of the glibc, but there is nothing like it in the glibc source that I used to compile. The mman.h in my glibc source is much shorter and the declaration of mmap64 seems correct. I guess that the file in my /usr/include/sys/mman.h is a generated file. I don't even want to count the number of 'mman.h' files on my system :-/ >So I think I can use this test program as check for mmap64 brokeness... Btw, the reference on large files http://www.suse.de/~aj/linux_lfs.html says that locking will never be supported on 64 bits files with glibc 2.1.3 so I don't see why it's necessary to support this case. Gerard

1 0

RE: Stanford Checker
by Patrik Stridvall June 5, 2001

June 5, 2001

> On Tue, 5 Jun 2001 David.Goodenough(a)dga.co.uk wrote: > > but has anyone thought of running these against Wine? > > I'm not sure if it'd do that much good. They're currently working on > checking the XFree86 tree (see the Xpert mailing list archives on > www.XFree86.org if you want), without a lot of big results > yet. From what > I gather, this tool has very limited and context-dependent > intelligence, Perhaps. But see below. > and is not conceptually very different from what a glorified > version of > Patrik's tools/winapi_check script might be, if given enough people > working on it. Not quite. winapi_check is 1. Specifically written for Wine 2. A syntax analyzer 3. Mostly ad hoc. The Stanford checker on the other hand is 1. Generalized for all C/C++ applications 2. A semantic analyzer 3. Formalized with a script language The way I choose when designing winapi_check was the one of the easiest ways and the way they choose is one of the hardest. So just calling it is glorified winapi_check gives the Stanford people too little credit. Anyway the way Stanford choose is really _very_ hard so what I fear is that they had to make some limiting decisions that makes their tool much less useful. That said regardless of how good the Stanford checker really will be I'm pretty there will be a lot of things that will be _much_ easier checked by writting a script for the Stanford checker than by extending winapi_check, because they have fundamentally different design philosofies. However don't underestimate winapi_check either. There is a lot of things that can quite easilt be checked by adding some clever perl pattern, if:s and a little state. Syntax level matching is not that bad just a little more crude. If somebody have something that he want to be automatically checked, please post it on the list and I will see what I can do.

1 0

Stanford Checker
by David.Goodenough＠dga.co.uk June 5, 2001

June 5, 2001

On the kernel list and in Alan Cox's patches for the kernel there are a lot of references to some logic checking tools being generated by Stanford. I do not know if these are generally available but has anyone thought of running these against Wine?

2 1

RE: Stanford Checker
by Patrik Stridvall June 5, 2001

June 5, 2001

> Maybe we have to ask them to check our code, rather than the other way > around? I think it far to much to ask. As Eric said, while it is AFAIK highly configurable with specific scripts, for each kind of check it is currently seems to be very much focused on the Linux kernel. I also believe, like Eric, that it is currently very buggy and that even if they have a powerful enough scripting language to be usable for Wine it will probably need some serious debugging to make the scripts we need to actually work in practise not only in theory. No, as I said, I think it is to much to ask. We will have to wait until they release it whenever that will be.

1 0

Re:Stanford Checker
by Eric Pouech June 5, 2001

June 5, 2001

last time I checked (a month ago), the tool wasn't freely available (nor finished), so nothing can be done for now when the tool is finished (<put here your date estimation>), one could be able to configure the tool for use in Wine (read: write the scripts for the things we want to check, like windows locks...) but, the tool is (as of today) configured for Linux, and check can only be done agains the Linux tree (AFAIU, they check for specific constructs like getting a spin lock, enabling / disabling IRQ... so highly dependent on Linux source...) however, I don't know (yet) the power of the script language they use for configuring, and whether it could be used for Wine... you'll have to wait until the tool is released... (I suspect, the tool is still buggy *AND* still requires lots of processing by hand - you get a 100 warnings from where only a few are real issues). A+ > On the kernel list and in Alan Cox's patches for the kernel there are a lot > of references to some logic checking tools being generated by Stanford. I > do not know if these are generally available but has anyone thought of > running these against Wine? --------------- Eric Pouech (http://perso.wanadoo.fr/eric.pouech/) The future will be better tomorrow, Vice President Dan Quayle __________________________________________________ Voila vous propose une boite aux lettres gratuite sur Voila Mail: http://mail.voila.fr

1 0

RE: Stanford Checker
by David.Goodenough＠dga.co.uk June 5, 2001

June 5, 2001

Maybe we have to ask them to check our code, rather than the other way around? Patrik Stridvall <ps(a)leissner.se> on 05-06-2001 12:31:25 PM To: David Goodenough/DGA/GB cc: wine-devel(a)winehq.com Subject: RE: Stanford Checker [Forgot to include list. Trying again.] > On the kernel list and in Alan Cox's patches for the kernel > there are a lot > of references to some logic checking tools being generated by > Stanford. I > do not know if these are generally available but has anyone thought of > running these against Wine? Yes, I have. However the problem is that it is not publicly available (yet) and I read somewhere (don't remember where) that it is no use trying to get copy even if you promise not to distribute it further. I will probably look into it when it is released. I have no idea when that will be though.

1 0

RE: Stanford Checker
by Patrik Stridvall June 5, 2001

June 5, 2001

[Forgot to include list. Trying again.] > On the kernel list and in Alan Cox's patches for the kernel > there are a lot > of references to some logic checking tools being generated by > Stanford. I > do not know if these are generally available but has anyone thought of > running these against Wine? Yes, I have. However the problem is that it is not publicly available (yet) and I read somewhere (don't remember where) that it is no use trying to get copy even if you promise not to distribute it further. I will probably look into it when it is released. I have no idea when that will be though.

1 0

CVS compilation problem: XvSelectNotify
by Heiko Nardmann June 5, 2001

June 5, 2001

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I just compiled CVS wine using ./configure --prefix=/opt/wine-20010605 --enable-dll --with-x and get the following error message starting Word: err:module:BUILTIN32_dlopen failed to load .so lib for builtin x11drv.dll: /opt/wine-20010510/lib/libwine_tsx11.so: undefined symbol: XvSelectPortNotify Could not load graphics driver 'x11drv' - -- Heiko Nardmann (Dipl.-Ing.), h.nardmann(a)secunet.de, Software Development secunet Security Networks AG - Sicherheit in Netzwerken (www.secunet.de) Weidenauer Str. 223-225, D-57076 Siegen Tel. : +49 271 48950-13, Fax : +49 271 48950-50 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.4 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE7HJ2Opm53PRScYygRAkZsAKCXRyt3ncR7sFRNatc7/6Bls3+eywCg9sO5 7oombEYBU9llG4o8Gp7LeyA= =a8Qa -----END PGP SIGNATURE-----

1 0

Re: PATCH: set_thread_buffer fixed
by Gerard Patel June 5, 2001

June 5, 2001

At 07:20 AM 05/06/2001 +0200, you wrote: >Not really, as Alexandre said. > >Could you please test the attached small test program? It fails (of course). Here is the call with my working computer (glibc2.2, linux 2.4) 0x8048668 <main+72>: call 0x8048464 <write> 0x804866d <main+77>: add $0xc,%esp 0x8048670 <main+80>: push $0x0 0x8048672 <main+82>: push $0x0 0x8048674 <main+84>: push %ebx 0x8048675 <main+85>: push $0x1 0x8048677 <main+87>: push $0x3 0x8048679 <main+89>: push $0x4 0x804867b <main+91>: push $0x0 0x804867d <main+93>: call 0x80484a4 <mmap64>p For the failing one : 0x80486c8 <main+72>: call 0x80484dc <write> 0x80486cd <main+77>: addl $0x20,%esp 0x80486d0 <main+80>: addl $0xfffffff8,%esp 0x80486d3 <main+83>: pushl $0x0 0x80486d5 <main+85>: pushl %ebx 0x80486d6 <main+86>: pushl $0x1 0x80486d8 <main+88>: pushl $0x3 0x80486da <main+90>: pushl $0x4 0x80486dc <main+92>: pushl $0x0 0x80486de <main+94>: call 0x804851c <mmap64> This calls fails because glibc detects that the value at (esp+28) is different of 0 - the real system call is never done, with your test app and Wine. In the glibc 2.1.3 source, this test is not explained; in the 2.2.2 source, there is a comment saying that the offset is 'too large'. I understand that the app is calling mmap64 on an operating system that is not supporting it (it's perfectly correct in this case, the failing box having 2.2.5 kernel), and glibc tests if the call can be mapped to standard mmap. I think that the problem is in the compilation; pushing only 6 dwords on the stack to call mmap64 is wrong. I have not yet figured out why, though (I have never been any good with C header problems) Gerard

2 1

Re: PATCH: set_thread_buffer fixed
by Gerard Patel June 5, 2001

June 5, 2001

At 08:12 AM 03/06/2001 +0200, you wrote: >> After applying your patch to current cvs, rm config.cache, ./configure, >> make depend && make Wine crashes as before. > >You need to run 'autoheader -l include' if any change to acconfig.h >is submitted. Thanks, it worked. Problem solved IMO. Gerard

1 0

mIRC crash when processing large files.
by rpbrown＠xtra.co.nz June 5, 2001

June 5, 2001

Details how to replicate the crash. (yes its time consuming, but it shows the crash) Ok, this is what I do to crash mIRC in wine. I install the mp3-script for mIRC. http://welcome.to/autoget load the script according to the readme..and join IRC Undernet, any Undernet server will do. /join #mpthree the person we're looking for is SuperServ (an Op) right mouse on the nick, and grab there mp3 list.. you can try other ppls lists, but SuperServ has a rather big list (8Mb) which at the end of the download the script keeps on loading (processing the txt file) ..and then crashes mIRC. Loading _this_ list itself into mIRC under windows normally, mIRC is locked for a 3-4 minutes (processing the list).. maybe wine gives up too soon, since its not getting a response within X amount of time. Or the list is too big for wine to handle? thats my guesses.. I hope this helps. -mdew

1 0

uname(1) on UnixWare (Solaris)
by Bang Jun-Young June 4, 2001

June 4, 2001

Hi, I've nearly completed new DLL configuration code in configure.in. Before making it available, I need to know what uname(1) command prints out on UnixWare (and Solaris?). As in NetBSD, it's: $ uname NetBSD I guess it would be $ uname UnixWare on UnixWare. I also wonder if UnixWare and Solaris refer to the same. Jun-Young -- Bang Jun-Young <bjy(a)mogua.org>

1 0

Re: FreeType autoconf checks
by Alexandre Julliard June 4, 2001

June 4, 2001

Ian Pilcher <ian.pilcher(a)home.com> writes: > CPPFLAGS works. Where the heck is that documented? (I can't find it in > the info pages for gcc or cpp.) It's not a gcc option, it's a configure internal variable (from $ac_cpp definition). This is the whole point: it will work independently of the compiler, or at least break the same way during configure and during the compile, which is what we want. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: PATCH: set_thread_buffer fixed
by lawson_whitney＠juno.com June 4, 2001

June 4, 2001

On Sun, 3 Jun 2001, Marcus Meissner wrote: > Is this the case where set_thread_buffer fails or not? > Okay, it is not. It crashes in debug_printf or dump_exports (pe_image.c). However, it is so too a glibc-2.1.3 problem. I think. Sort of. The same Wine that crashes with kernel 2.4.4 and 2.4.5 works fine with kernel 2.2.14, and I bet later 2.2.x kernels when (if) I get around to building them. I think I might not be smart enough to debug it in detail. Maybe I should just send www4mail after glibc-2.2 instead. > I don't see a failure message here. > > Also, could you try to run the attached test program and see if it fails? success with either kernel version. Hmmm, do I have 2.0.30 still lying around? 1.3.18? > > Ciao, Marcus > Lawson Constants aren't, and variables won't. ---cut--- ________________________________________________________________ GET INTERNET ACCESS FROM JUNO! Juno offers FREE or PREMIUM Internet access for less! Join Juno today! For your FREE software, visit: http://dl.www.juno.com/get/tagj.

1 0

Re: FreeType autoconf checks
by Alexandre Julliard June 4, 2001

June 4, 2001

Ian Pilcher <ian.pilcher(a)home.com> writes: > One thing to keep in mind, however. The FreeType headers themselves > use relative paths to include one another, so this is an issue that > we're not going to be able to completely eliminate. This is why I > had to mess around with the C_INCLUDE_PATH variable; AC_CHECK_HEADER > tries to preprocess the header file, so it has to be able to find any > subsidiary headers. (Is this what you mean by finding a way to check > that #include <freetype/freetype.h> works?) Yes; we should be using only relative paths in AC_CHECK_HEADER, since this is what we will use in the code. I don't think C_INCLUDE_PATH is portable either. Did you try setting CPPFLAGS instead? -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: FreeType autoconf checks
by Ian Pilcher June 4, 2001

June 4, 2001

Alexandre Julliard wrote: > > Well, I don't like it much, but I haven't had time to look into a > better solution yet. Basically I think it's dangerous to check for the > file name with an absolute path and then include it with a relative > path, since there is no guarantee we'll get the same file. I think we > should really find a way to check that #include <freetype/freetype.h> > works given the right C flags. Also I'm not sure about the portability > of the ${FREETYPEINCL#-I} construct. > I see your point. Now that I think about it, I probably would have been bitten by it if anyone will. I run Red Hat 7.1 which has: /usr/include/freetype/freetype.h (from FreeType 1.x) /usr/include/freetype2/freetype/freetype.h (from FreeType 2.x) I just checked the gcc documentation, and directories specified with '-I' options on the command line are checked first. I have no idea, however, if other compilers behave in a similar manner. Using absolute paths for the FreeType headers would be non-trivial, but probably doable. We would have to use autoconf to define macros that include the complete path name of the headers we want to include; then #include those macros instead of the file names. Somehow, I don't think you're going to want to do that. One thing to keep in mind, however. The FreeType headers themselves use relative paths to include one another, so this is an issue that we're not going to be able to completely eliminate. This is why I had to mess around with the C_INCLUDE_PATH variable; AC_CHECK_HEADER tries to preprocess the header file, so it has to be able to find any subsidiary headers. (Is this what you mean by finding a way to check that #include <freetype/freetype.h> works?) I also had the same thought about the ${FREETYPEINCL#-I} construct. According to http://www.opengroup.org/onlinepubs/7908799/xcu/chap2.html#tag_001_006_002 it seems to be OK. Frankly, I share your distaste for this whole mess. -- ======================================================================== Ian Pilcher ian.pilcher(a)home.com ========================================================================

1 0

Re: fixup bogus SetLastError()
by Alexandre Julliard June 4, 2001

June 4, 2001

Andreas Mohr <a.mohr(a)mailto.de> writes: > objects/gdiobj.c really has no business dealing with SetLastError(). > SetLastError() is a mechanism to communicate errors to external programs, > so it should definitely not be used for highly internal functions. You didn't include the gdiobj.c patch, but I disagree with it anyway... There can be many good reasons for setting last error in internal functions; for instance setting it inside GDI_GetObjPtr avoids having to set it in every single caller. And in the very rare case where this is not what we want, the caller can always set it to another value. -- Alexandre Julliard julliard(a)winehq.com

3 2

stack limit
by Stephen Crowley June 4, 2001

June 4, 2001

What is the correct way to increase the default stack size for winelib applications? Also, are there any drawbacks to making the value fairly large? -- Stephen Crowley Software Engineer Placemark Investments 972.404.8100 x28

2 1

Re: stack limit
by Alexandre Julliard June 4, 2001

June 4, 2001

Stephen Crowley <stephenc(a)placemark.com> writes: > What is the proper way to increase the stack size for winelib > applications? There seems to be a limit of around 2^20 bytes right now. There is no way to do this at the moment; but it shouldn't be too hard to add support for a 'stack' statement in the application spec file. -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: fixup bogus SetLastError()
by Alexandre Julliard June 4, 2001

June 4, 2001

520053692817-0001(a)t-online.de (Andreas Mohr) writes: > Now that's where I disagree ! > The caller can *not* always set it to another value. > If it does and the program intends to read the LastError of a *previously* > executed API that failed, then the LastError will be reset even though > it shouldn't have been ! Then the caller needs to save/restore it across the GDI_GetObjPtr call. That's trivial to do. > In short: the SetLastError behaviour needs to match Windows behaviour very, > very closely. Definitely; but it's not clear to me that not setting last error is anywhere closer to Windows than setting it to INVALID_HANDLE. GDI_GetObjPtr should do what most callers need, and IMO this means setting it. I may be wrong, I haven't checked the error returns of all GDI functions (and of course this is nowhere documented). > This would be SelectObject() in this case, BTW. Which case? Do you have evidence that a SelectObject() with an invalid handle must not set last error? If so then of course it needs fixing. -- Alexandre Julliard julliard(a)winehq.com

2 1

stack limit
by Stephen Crowley June 4, 2001

June 4, 2001

What is the proper way to increase the stack size for winelib applications? There seems to be a limit of around 2^20 bytes right now. -- Stephen Crowley Software Engineer Placemark Investments 972.404.8100 x28

1 0

tan not exported by msvcrt, ntdll
by Francois Gouget June 3, 2001

June 3, 2001

I tried compiling Winelib applications without '-lm' (btw, the comment in configure.in is incorrect, it concerns Linux too, not just BeOS). But I had undefined symbol errors for applications that use tan and a couple other math functions. The reason is that both ntdll and msvcrt declare them with '-noimport': * msvcrt.spec: @ forward -noimport tan ntdll.tan * ntdll.spec: @ cdecl -noimport tan(double) tan So a Winelib application that imports msvcrt and ntdll does not get tan and thus must link with libm in the Unix sense. sin, cos and most others were ok though! Changelog: * dlls/msvcrt/msvcrt.spec Remove some '-noimport' for math functions -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ 1 + e ^ ( i * pi ) = 0

1 0

AnimateWindow() calls SetWindowPos() on a timer???
by Bobby Bingham June 3, 2001

June 3, 2001

Forgot a detail ... Every time it send the messages it tells it not to move, size, redraw, activate, or change z-order. I could see the reason for these if if didn't tell it not to do anything, but ... ?? Bobby Bingham uhmmmm(a)ameritech.net

1 0

AnimateWindow() calls SetWindowPos() on a timer???
by Bobby Bingham June 3, 2001

June 3, 2001

I am implementing AnimateWindow(), and I recorded all the messages win98 sent to my window when I called AnimateWindow(). It send a lot of groups of WM_WINDOWPOSCHANGING, WM_NCCALCSIZE, and WM_WINDOWPOSCHANGED. The number of times it sends these seems to correspond to how long I tell it to make the animation, almost as if it were on a timer. It looks like it calls SetWindowPos() on the window a lot, but I don't see why they would do this. Does anyone know?? I can supply my test app once I clean it up a little. Bobby Bingham uhmmmm(a)ameritech.net

1 0

off topic: Windows 98 and binutils
by Brian Craft June 3, 2001

June 3, 2001

Hey -- Sorry for the OT post, but I suspect someone here will know the answer to this. I'm trying to objdump some Windows 98 device drivers & dlls. Last time I did something like this it was user-mode code from 95 or NT (not sure which), and I just configured binutils for i386-pe. Only this same configuration won't read the Windows 98 files -- it says the file format isn't recognised. I can still read the old Windows binaries with this config. I also tried i386-msdos, for good measure. What's the scoop? 98 has a new format? Is there a binutils target that will do the job? b.c.

2 1