wine-devel

wine-devel@winehq.org

2 participants
84469 discussions

Re: Floppy handling Fix
by Alexandre Julliard 05 Jul '02

05 Jul '02

Keith Matthews <keith_m(a)sweeney.demon.co.uk> writes: > 1 - Fix incorrect reporting of no media in floppy drive as > permissions problem. > 2 - Provide partial workaround for very long delays (c. 40 sec per > lookup, about 7 minutes to set up file browser) on lookups to > floppy drive not physically present (e.g. laptop with floppy or > CD setup) I'm not sure why you are trying to duplicate the kernel permission checks, this is clearly wrong. You should simply do an open and check errno to find out what went wrong. Also there's no point in having a DRIVE_NO_MEDIA flag since this is a temporary condition. For the case of open succeeding with no floppy drive, I think you could handle it by checking if the subsequent read() fails. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: Better fix for [bug 829] (same surface display corruption)
by Lionel Ulmer 05 Jul '02

05 Jul '02

> I took Lionel's sugestions to heart and came up with a better fix for > bug 829. This has a better check for defining when the corruption could > occur and should be a little faster Well, I had not in mind exactly what you did. > + SameSurfaceOK = (!((src == iface)&&(sbase < dbuf)&&(xdst.top < xsrc.bottom)&&(xdst.left < xsrc.right))); I do not really understand all the details of this check, so a comment would be nice, but well, I think it is unneeded (see below). > + if (SameSurfaceOK) { > + for (y = 0; y < dstheight; y++) { > + memcpy(dbuf, sbuf, width); > + sbuf += sdesc.u1.lPitch; > + dbuf += ddesc.u1.lPitch; > + } > + } else { > + > + sbuf += (sdesc.u1.lPitch*dstheight); > + dbuf += (ddesc.u1.lPitch*dstheight); > + > + for (y = 0; y < dstheight; y++) { > + sbuf -= sdesc.u1.lPitch; > + dbuf -= ddesc.u1.lPitch; > + memcpy(dbuf, sbuf, width); > + } > + } > + The problem I can see is that you use 'memcpy' even in the SameSurfaceNotOK case. And this is bad as memcpy does NOT handle overlapping memory zones. Moreover, you only have the 'descending' copy order in that case and not both cases. What I had in mind in my comments to your patch was something like that : if (src != iface) Do the standard memcpy code path else if source_Y > dest_Y do the ascending memmove copy else do the descending memmove copy This way you did not hurt the common case (ie Blt from two different surfaces). Now I agree that even in the same surface case, one could add some checks to see if the rectangles overlap or not, but I do not know if it would be worth the pain as memmove should not be THAT slower than memcpy. Lionel -- Lionel Ulmer - http://www.bbrox.org/

3 3

RE: warcraft 3
by Hans Leidekker 05 Jul '02

05 Jul '02

> After buying Warcraft 3, and reading on the mailing list that the beta had > worked under wine, I decided to try the final version under wine. It installed > fine, but won't run. Or rather, when I try to run it, it prints this out, and > just sits there: > > fixme:cdrom:CDROM_GetIdeInterface CD-ROM device with major ID 11 not supported > fixme:cdrom:CDROM_GetIdeInterface CD-ROM device with major ID 11 not supported > fixme:midi:OSS_MidiInit Synthesizer support MIDI in. Not supported yet (please report) Let me guess: you have SCSI cdrom drive. I stumbled on this one as well some time ago. I learned that 11 is the SCSI major. It looks strange to me that the function CDROM_GetIdeInterface gets called while we have a SCSI cdrom drive? Anyways, as a workaround -- I don't think this is the right solution -- I added the following line to the switch statement in cdrom:CDROM_GetIdeInterface: case SCSI_CDROM_MAJOR: *iface = 0; break; This allowed my application to proceed (don't remember which application that was though). Regards, -Hans Leidekker

2 1

GRand Prix Legends network code
by Christoph Frick 05 Jul '02

05 Jul '02

Hi! I did some debugging, to find out why the game Grand Prix Legends did not work as network client/server. The problem: Its only possible to connect to servers/let clients connect from the local network. Solution: The game set (WS_)SO_DONTROUTE on the socket - and of course this works perfect under unix - but i never had problems while i played this game under windows so this is either not set, when the game runs under windows - or it has no effect within windows. Also the game like to bind the socket for the client connection a second time, when the first connection failed or was canceled. This does not work (addr. in use) - so a (WS_)SO_REUSEADDR would have been needed. Could please someone with a deeper insight in the world of windows direct me to some docu/specs or what ever to get informations about the winsock stuff? The link in the header of the dlls/winsock/socket.c is not valid. I even dont know what winsock version is the current one (1.1,2.0,???). Help is appreciated! Thanx alot! CU!

1 0

Re: dlls/advapi32/registry.c patch for RegEnumValuesA
by Massimo 05 Jul '02

05 Jul '02

Resubmitting as for Tony Lambregts request :) Well, finally found and corrected a bug that made AutoCAD R14 hang... I guess some other apps could have the same problem solved, but had no time to test more :) In RegEnumValueA, when the size for requested data buffer is too small, function should return an error (as it does), but also return correct buffer size (as it doesn't). Found in AutoCAD; it asked for a reistry key value, got error, and then looped forever trying to get val with wrong buffer size. I guess, but had no time yet to check, that this problem could happen somwhere else in Registry.c, I'll check it out next days. With this patch, finally AutoCAD R14 is perfectly usable in wine, even without a native windows installed ! Change log: Change In RegEnumValueA, when the size for requested data buffer is too small, function should return an error (as it did), and also return correct buffer size (now it does). Max max(a)com.veneto [reversed :) ] ----- --- dlls/advapi32/registry.c Sat Jul 6 01:29:08 2002 +++ /home/massimo/wine-patches/registry.c Sat Jul 6 01:35:32 2002 @@ -1157,6 +1157,9 @@ RtlUnicodeToMultiByteSize( &len, info->Name, info->NameLength ); if (len >= *val_count) { + /* MUST RETURN REQUESTED CORRECT BUFFER SIZE, NOT ONLY TELL IT'S WRONG */ + *val_count = len + 1 ; status = STATUS_BUFFER_OVERFLOW; goto done; } @@ -1174,6 +1177,9 @@ { if (len > *count) { + /* MUST RETURN REQUESTED CORRECT BUFFER SIZE, NOT ONLY TELL IT'S WRONG */ + *count = len+1 ; status = STATUS_BUFFER_OVERFLOW; goto done; }

2 1

Re: Added support for APPMODE=<type>[:<mainFunc>]
by Alexandre Julliard 05 Jul '02

05 Jul '02

Gregg Mattinson <gm138242(a)scot.canada.sun.com> writes: > - .spec files can't be compiled with -g when using Forte C. This causes static > variables to have a "random" prefix attached to their name, and prevents the > assembly code from linking to them correctly. > - Forte C has linker issues, so cui executables with main() conflicted with > main() in $WINELOADER. Renaming main() to main2() solved the issue, so I added > support to winebuild to allow the main function name to be optionally specified > in the makefile. I have applied the rest of the patch, but these are bugs that need to be fixed, not simply worked around. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: dlls/advapi32/registry.c patch for RegEnumValuesA
by Tony Lambregts 04 Jul '02

04 Jul '02

Max wrote: >Here a patch that corrects a nasty bug in dll/advapi32/registry.c, causing >AutoCAD R14 (and I guess other apps too...) to hang. > >Regards > >Max > > > >------------------------------------------------------------------------ > >--- dlls/advapi32/registry.c Sat Jul 6 01:29:08 2002 >+++ /home/massimo/wine-patches/registry.c Sat Jul 6 01:35:32 2002 >@@ -1157,6 +1157,9 @@ > RtlUnicodeToMultiByteSize( &len, info->Name, info->NameLength ); > if (len >= *val_count) > { >+ /* CHANGELOG : 04.07.2002 by Max >+ MUST RETURN REQUESTED CORRECT BUFFER SIZE, NOT ONLY TELL IT'S WRONG */ > The comment is OK I suppose, but it is part of the change log. Change log entries should not be part of the patch itself. >+ *val_count = len + 1 ; > status = STATUS_BUFFER_OVERFLOW; > goto done; > } >@@ -1174,6 +1177,9 @@ > { > if (len > *count) > { >+ /* CHANGELOG : 04.07.2002 by Max >+ MUST RETURN REQUESTED CORRECT BUFFER SIZE, NOT ONLY TELL IT'S WRONG */ > Same here. >+ *count = len+1 ; > status = STATUS_BUFFER_OVERFLOW; > goto done; > } > > Sorry to nitpick could you resubmit something like this ---- This is the first part. This is where you can say what you did, and say what you wanted to say, in as long and as rambling way as you want. It is for expaining whatever you wanted to expain, why you wanted to do it the way you wanted to do it section. next comes the change log . after that your name and contact. Change log: Change In RegEnumValueA, when the size for requested data buffer is too small, function should return an error (as it did), and also return correct buffer size (now it does). Max max(a)veneto.co ----- attched patch

1 0

Running half-life turns X display into garbage
by Joshua Thielen 04 Jul '02

04 Jul '02

Hello, When I run half-life, the X display gets totally corrupted and the only way to restore is to restart the X server. I'm using an NVidia G-force 3 video card running at 1024x768 with 16-bit color. I have already run a few traces, but since I have no idea at what point the display gets corrupted, it's kind of hard to debug. Has anyone encountered this problem yet? Are there any specific functions or channels I should be looking at in the trace? Thanks, Josh Thielen __________________________________________________ Do You Yahoo!? Sign up for SBC Yahoo! Dial - First Month Free http://sbc.yahoo.com

4 4

2 quick things on the Cygwin/Mingw port
by Steven Edwards 03 Jul '02

03 Jul '02

Hola 1. This is simple but as you guys know I have this lysdexic thing when it comes to configure. Libwine needs to import certain functions from the c library to compile, at present it is hard coded for -lmsvcrt. This needs to be changes to import -lcygwin if configure detects a cygwin host/target. I know its simple we just need to have the check in configure that sets CLIBRARY=(lmsvcrt or lcygwin) then Makefile.in in libwine needs to be changed like this: libwine.dll: $(OBJS) $(DLLWRAP) $(DLLWRAPFLAGS) --export-all --implib libwine.a -o libwine.dll $(OBJS) -lmsvcrt to libwine.dll: $(OBJS) $(DLLWRAP) $(DLLWRAPFLAGS) --export-all --implib libwine.a -o libwine.dll $(OBJS) $(CLIBRARY) 2. I was able to get Wineserver.exe to compile under cygwin when I stubbed getthreadcontext and setthreadcontext. Kernel32.dll implements these to functions so on the Mingw or Cygwin port can I import it for these needed functions? Thanks and have a good 4th Steven __________________________________________________ Do You Yahoo!? Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.com

1 0

Re: DOSFS_GetPathDrive patch
by Alexandre Julliard 03 Jul '02

03 Jul '02

Duane Clark <dclark(a)akamail.com> writes: > In particular, it calls GetFileAttributes with a path like: > \c\Xilinx\bin\lib\tcl8.3\init.tcl > which fails in DOSFS_GetPathDrive. > > Changelog: Check for absolute Unix path in DOSFS_GetPathDrive The problem I see with that is that the app can no longer access a directory that has the same name as a drive root (for instance if you have a /msdos drive then the app cannot use \msdos to mean c:\msdos). Where does the Xilinx path come from? That's not a valid Windows path anyway so that looks like a bug. -- Alexandre Julliard julliard(a)winehq.com

2 1

(no subject)
by Steven Edwards 03 Jul '02

03 Jul '02

Anyone planning on attending the Linux World Expo? I was going to be in the Bay area on the 12th anyway to tour Berkley so I think I'm going to extend my trip to make the show. Steven __________________________________________________ Do You Yahoo!? Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.com

1 0

Bug in RegEnumValueA ?
by Massimo 03 Jul '02

03 Jul '02

Again investigating AutoCAD r14 app under wine, I found a bug that make app hang when some registry values are present. At first I've made a script that cleans up registry before launching AutoCAD,but this was only a way-around to the problem. Looking at relay output, I found a loop in calling RegEnumValueA : // @@@@ START OF LOOP BLOCK HERE......... :Call advapi32.dll.RegEnumValueA (000000cc,00000001,40486844,404868ec,00000000,404868f4,41e62020, 404868fc) ret=008a3850 :Call ntdll.dll.NtEnumerateValueKey (000000cc,00000001,00000001,40486544,00000100,40486540) ret=406afb24 :Ret ntdll.dll.NtEnumerateValueKey() retval=00000000 ret=406afb24 :Call ntdll.dll.RtlUnicodeToMultiByteSize(4048653c,40486558 L"REGPATH\\\\HKEY_LOCAL_MACHINE\\Software\\Autodesk\\AutoCAD\\R1 4.0\\ACAD-2452448:80143967\\AutodeskApps\\LsObj",0000000e) ret=406afc03 :Ret ntdll.dll.RtlUnicodeToMultiByteSize() retval=00000000 ret=406afc03 :Call ntdll.dll.RtlUnicodeToMultiByteN (40486844,00000007,00000000,40486558,0000000e) ret=406afc24 :Ret ntdll.dll.RtlUnicodeToMultiByteN() retval=00000000 ret=406afc24 :Call ntdll.dll.RtlUnicodeToMultiByteSize(40486538,40486566 L"\\\\HKEY_LOCAL_MACHINE\\Software\\Autodesk\\AutoCAD\\R14.0\\AC AD-2452448:80143967\\AutodeskApps\\LsObj",000000bc) ret=406afc77 :Ret ntdll.dll.RtlUnicodeToMultiByteSize() retval=00000000 ret=406afc77 :Call ntdll.dll.RtlNtStatusToDosError(80000005) <--ERROR HERE ! ret=406afd5d :Ret ntdll.dll.RtlNtStatusToDosError() retval=000000ea ret=406afd5d :Ret advapi32.dll.RegEnumValueA() retval=000000ea ret=008a3850 <-- MORE DATA AVAILABLE RETURN (AUTOCAD KEEPS TRYING TO GET DATA) // @@@@ END OF LOOP BLOCK HERE Here RegEnumValueA returns an error of invalid buffer; Looking into code it seems to me that the NtEnumerateValueKey() used to find the largest name and value sizes of keys to be enumerated returns a wrong size; next use for reading the key gives an arror of buffer overflow, being it allocated with wrong (too small) size. AutoCAD keeps trying to get the keys/values, in and endless loop. Looking more with some debug print statements, it appears that Autocad allocates 4 bytes of buffer for the value, but RegEnumValueA needs much more, 94 in this case, giving an Out-Of-Buffer error back. It seems to me that the 94 bytes are for the name of the key, not for the value.... Regards Max

3 2

Re: Fix for [bug 829] DIB_DirectDrawSurface_Blt same src./dest. surface = screen corruption
by Lionel Ulmer 03 Jul '02

03 Jul '02

The patch looks fine to me. There is just one comment opn these lines : > + if (SameSurface) > + { > + static BOOL WarnSurface = FALSE; > + if (!WarnSurface) > + { > + FIXME("\tSame source and destination could cause display problems\n"); > + WarnSurface = TRUE; > + } > + } Why did you add this ? If you verified your patch and it's working, why add additionnal debug informations that will only confuse people :-) ? Moreover, I do not know the performance hit between a 'memcpy' and a 'memmove', but well, one could have put the old code in a if (!SameSurface) to have the fastest path available to the most common case. A part from these nitpicks, the patch looks fine and I will test it this evening on FalloutTactics to see if it fixes the screen corruption :-) Lionel -- Lionel Ulmer - http://www.bbrox.org/

2 1

Re: Fix for [bug 829] DIB_DirectDrawSurface_Blt same src./dest. surface = screen corruption
by Ian D. Stewart 03 Jul '02

03 Jul '02

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Tuesday 02 July 2002 18:57, Tony Lambregts wrote: > Change log: Fix screen corruption when source and destination are the > same surface and areas are the same size. I've applied this patch. While the on-screen garbage has not entirely gone away, it does appear to be better (but then, it could just be my imagination...). No change to hang-up after playing movies. Ian -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.7 (GNU/Linux) iD8DBQE9Iv8YAjpHQQmBAUMRAtjMAJ0eSyS5Z06Kr51uLloq/Nn9CuKQXwCglPrD Ff3Vfg2m2tcqpA1VkeQI9+0= =IPbG -----END PGP SIGNATURE-----

2 1

Arts, DSound, and Media Player
by Jeremy Shaw 03 Jul '02

03 Jul '02

Hello, I am trying to get media player to work with the winearts audio driver. I compiled the latest source from CVS, added Drivers="winearts.drv" to the config file, and installed media player. When I try to play a file (like an mp3), I do not get any audio. If I watch the elapse time, it will increment normally for a few seconds and then start playing super fast. Also when it starts playing super-fast I get the following error a bunch of times: fixme:dsound:DSOUND_MixOne problem with underrun detection (mixlen=17640 < primary_done=49392) At this point in time I know very little about what is wrong. Do you guys (we are all guys ???) think it is a problem with the wine source, my build of KDE, or just some messed up config files. Or perhaps I just installed too many native DLLs trying to make media player work? I will keep hacking away, but any guidance will be highly valued. I have not gotten ever gotten sound using the winearts driver, but so far I have only tried media player. Thanks in advance, Jeremy Shaw.

2 1

Re: [help] GDI text fonts
by Huw D M Davies 03 Jul '02

03 Jul '02

On Wed, Jul 03, 2002 at 04:54:18PM +0800, leanne wrote: > Huw D M Davies wrote: > > > > >It's still not clear to me whether you're using XServer rendered fonts > >or client side (Wine) rendered fonts. Could you send me the > >--debugmsg +font,+xrender log ? > > > >Huw. > > > Hello, > > I have checked cvs today and found that it supports to download tt > fonts to printer but the position of the text messes up. > Maybe it's a bug. > > I still find the font of the interface is very ugly, what do you think? > Attached please find the log and screenshot. > Very thanks. > > trace:font:DumpFontList Family: L"AR PL Mingti2L Big5" > trace:font:DumpFontList L"Reguler" > trace:font:DumpFontList Family: L"AR PL KaitiM Big5" > trace:font:DumpFontList L"Regular" There's a problem with Type42 font downloads, a workaround for this to edit generic.ppd and change the line *TTRasterizer: Type42 to *TTRasterizer: None this will tell Wine to send download the font as a Type 1 outline which should work. Do you mean that the roman (English) font is ugly (since I have no idea whether the Chinese characters look ok maybe you let me know how they look)? You only have two fonts installed (see the above fragment from the log), so the english text has to come from one of these fonts. What you may want to do is to install some fonts like Arial, Times New Roman etc. You could either copy these from a Windows installation (or add a line to the [FontDirs] section of config to point at an existing Windows partition font directory), or you can download a selection of the core Windows fonts from Microsoft's website. Huw. -- Huw Davies huw(a)codeweavers.com

1 0

Still getting this error
by Steven Edwards 03 Jul '02

03 Jul '02

With the last release I can build and use wine but with the current CVS I still get the following errors. Warning: loading builtin gdi32.dll, but native version already present. Expect trouble. Warning: loading builtin user32.dll, but native version already present. Expect trouble. wine: Unhandled exception, starting debugger... I am still not able to track down the patch that causes this. I have a clean WINE with no windows on any of my drives in my slackware box. Thanks Steven __________________________________________________ Do You Yahoo!? Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.com

1 1

Re: Private Profile Regression Test
by Dmitry Timoshkov 03 Jul '02

03 Jul '02

"Joshua Thielen" <thielen(a)netprince.net> wrote: > Anyhoo, please let me know what you think - are the tests sufficient, > too much, not written properly? etc. First of all, thank you very much Joshua! Really great work! A couple of comments though: /* * Special case: NULL lpReturnedString with size > 0 * This causes an application error in win2k. Let's make wine better :) */ IMO it's not very good idea to make some tests succeed under Wine, which clearly crashes under win2k. That would play some bad role in the case of the back porting Wine code to Windows, since some code, internal to Wine, could rely on that forbidden API behaviour. I also have the following failures under win2k: Number of bytes read do not match FAIL: c:\result.ini does not match compare1.ini Number of bytes read do not match FAIL: c:\result.ini does not match compare2.ini -- Dmitry.

2 2

Re: Small 'let's prevent a crash' fix
by Lionel Ulmer 02 Jul '02

02 Jul '02

> This does not fix the real problem (ie why does this DC not have a font even > if the application called a 'SelectObject' on it), but well, it cannot harm > to be more verbose and to prevent crashes if possible :-) OK, found the 'error' : the problem comes from the fact that I am using client-side fonts (aka XRender) : so the 'font' part of the physDev structure is always '0'. And that also means that we cannot use as is the glXUseXFont function. So basically, I need to implement it myself using 'GetGlyphOutline' (should not be that hard to do :-) ). Now my question : is 'GetGlyphOutline' always working even when not using client-side fonts ? If yes, I will completely remove the call to glXUseXFonts (I know, in that case, fonts displayed by OpenGL will be different by the one displayed by X, but well, ...) and always use the manual 'GetGlyphOutline' function. And I will try at the same time to remove all GLX calls from wgl.c and put them in special X11DRV escapes to have a real X11 free wgl.c file (ie for example, creating an OpenGL context would be a special escape for the X11DRV as it's basically something that is in GLX). Lionel -- Lionel Ulmer - http://www.bbrox.org/

2 1

Re: Removed extra , from enum declarations
by Andreas Mohr 02 Jul '02

02 Jul '02

On Tue, Jul 02, 2002 at 03:39:04PM -0400, Gregg Mattinson wrote: > Resubmitting because I forgot to attach the file. Sorry. I'm also sorry about > the large number of posts today. You feel sorry about WHAT !? Heck, I guess you might have a minimalistic reason to feel sorry once you manage to crash the mailing list server due to sending way too many patches, but until that happens... ;-) Have a nice day ! -- "My attitude is, everybody should try competing with Microsoft once in their life. Once." - Marc Andreessen, former Netscape lead employee, in a "browser wars" interview

1 0

Re: Added support for APPMODE=<type>[:<mainFunc>]
by Joerg Mayer 02 Jul '02

02 Jul '02

On Tue, Jul 02, 2002 at 04:06:56PM -0400, Gregg Mattinson wrote: > - .spec files can't be compiled with -g when using Forte C. This causes static > variables to have a "random" prefix attached to their name, and prevents the > assembly code from linking to them correctly. > - Forte C has linker issues, so cui executables with main() conflicted with > main() in $WINELOADER. Renaming main() to main2() solved the issue, so I added > support to winebuild to allow the main function name to be optionally specified > in the makefile. > - Forte C does not support .previous It was replaced with a .section statement > for the appropriate section > - '.section .text' is incorrect. It should be '.section ".text"' You omitted: - Replace constants by numeric values (aka magic numbers). > - fprintf( outfile, " rec.code = 0x%08x;\n", EXCEPTION_WINE_STUB ); > + fprintf( outfile, " rec.code = 0x%08x;\n", 0x80000100 ); - Inserted compiler specific compilation (as opposed to feature specific compilation) > +#ifndef __SUNPRO_C > fprintf( outfile, "\n#include \"pshpack1.h\"\n" ); > +#endif - Removed some unneeded includes. Ciao Jörg -- Joerg Mayer <jmayer(a)loplof.de> I found out that "pro" means "instead of" (as in proconsul). Now I know what proactive means.

1 0

Re: Void functions can't return a value
by Joerg Mayer 02 Jul '02

02 Jul '02

On Tue, Jul 02, 2002 at 03:34:44PM -0400, Gregg Mattinson wrote: > - void functions can't return a value, even if it is the return value from > another void function. Just looking at the patch a bit. This looks wrong to me. If you want to make the changes, you should convert return xxx() into { xxx(); return } > void WINAPI PathStripPathAW(LPVOID lpszPath) > { > if (SHELL_OsIsUnicode()) > - return PathStripPathW(lpszPath); > - return PathStripPathA(lpszPath); > + PathStripPathW(lpszPath); > + PathStripPathA(lpszPath); <----- should not be executed if condition is met > } This pattern is repeated several times. Ciao Jörg -- Joerg Mayer <jmayer(a)loplof.de> I found out that "pro" means "instead of" (as in proconsul). Now I know what proactive means.

1 0

Re: heads up: wine 20020605 broken on FreeBSD, workaround included
by Gerald Pfeifer 02 Jul '02

02 Jul '02

[ I'm now also Cc:ing this to the Wine developers, thus the full-quote. In fact, all non-GLIBC platforms seem to be affected by this bug! ] Based on your pointer, I believe I found the problem. In aclocal.m4 we have: AC_DEFUN([WINE_CHECK_ERRNO], [AC_CACHE_CHECK([for reentrant libc: $1],[wine_cv_libc_r_$1], [AC_TRY_RUN([int myerrno = 0; char buf[256]; int *$1(){return &myerrno;} main(){connect(0,buf,255); exit(!myerrno);}], wine_cv_libc_r_$1=yes, wine_cv_libc_r_$1=no, wine_cv_libc_r_$1=yes)]) AS_IF([test "$wine_cv_libc_r_$1" = "yes"],[$2],[$3])]) which, in configure, becomes: int myerrno = 0; char buf256; int *__error(){return &myerrno;} main(){connect(0,buf,255); exit(!myerrno);} That is, instead of an array of 256 characters, we have one character, and the invocation of connect() is definitely incorrect. Could someone more familiar with autoconf please have a look at this? Gerald On Mon, 1 Jul 2002, Scott Bolte wrote: > Over the weekend I tried to get wine 20020605 to run on a > new installation of FreeBSD 4.6. It took a while, I haven't > tried wine since 1997 after all, but I determined that my > problems were due to configure's reentrant libc test failing > when building the port. > > The wine_cv_libc_reentrant test in configure.ac first checks > for Linux (__errno_location) and then tests for FreeBSD > (__error). I don't have more time to figure out why, but > nesting the FreeBSD test that way results in a false > negative. The configuration will continue, as will the > build, but the resulting non-reentrant code goes into an > endless loop when trying to spawn threads. > > A crude workaround, for FreeBSD users at least, is to > replace the Linux test with the FreeBSD one. This needs to > be done before configure is run. I did it by hand, but I > expect the following code would do the trick. > > % cd /usr/ports/emulators/wine > % make patch > % perl -i -pe 's/__errno_location/__error/' work/wine-20020605/configure.ac > % make > % make install > > Below is the autoconf test script that demonstrates the > problem. It needs to be run in > /usr/ports/emulators/wine/work/wine-20020605/configure.ac > > Scott > > P.S. I'm not on the freebsd-emulators mailing list, but I > see from freebsd-questions that at least one other > person got wine to compile on 4.6. I can't explain > why it worked for him and not for me. However, I > installed a brand new copy of 4.6 from the standard > ISO images just to run wine. > > ---- testconf.ac > dnl Process this file with autoconf to produce a configure script. > dnl Original author: Michael Patra > dnl See ChangeLog file for detailed change history. > > m4_define(WINE_VERSION,regexp(m4_include(VERSION),[version $[-.0-9A-Za-z]+$],[\1])) > > AC_PREREQ(2.53) > AC_INIT([Wine],WINE_VERSION) > AC_CONFIG_SRCDIR(server/atom.c) > AC_CONFIG_AUX_DIR(tools) > > dnl **** Check for some programs **** > > AC_PROG_CC > > dnl **** Check for reentrant libc **** > > wine_cv_libc_reentrant=no > > dnl ##### > dnl ##### There is a problem if the FreeBSD test is nested > dnl ##### after the Linux test. In that instance, it fails. > dnl ##### If, on the other hand, it is tested first, it works. > dnl ##### > dnl ##### Swap the order of two following test case to see > dnl ##### how they differ. (Ignore the second case since it > dnl ##### will use cached data.) > dnl ##### > > dnl FreeBSD style errno location > WINE_CHECK_ERRNO([__error], [wine_cv_libc_reentrant=__error]) > > dnl Linux style errno location > WINE_CHECK_ERRNO([__errno_location], [wine_cv_libc_reentrant=__errno_location], > dnl FreeBSD style errno location > WINE_CHECK_ERRNO([__error], [wine_cv_libc_reentrant=__error] > )) > -- Gerald "Jerry" pfeifer(a)dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/

3 4

alsa disable configure?
by Chris Thielen 02 Jul '02

02 Jul '02

Are there any patches in waiting to disable the use of ALSA in the configure script ... or is this not needed?

1 0

IsDialogMessageA/W question
by Sander van Leeuwen 01 Jul '02

01 Jul '02

Hi, Is there a specific reason why unprocessed messages in IsDialogMessage don't get dispatched? Experiments in Windows 2000 show this is not (always) correct. It seems all messages passed on to IsDialogMessage are dispatched and it always returns 1. IsDialogMessage doesn't call DispatchMessage, but the function called internally by DispatchMessage. Changing the behaviour to match Windows 2000 fixes at least one problem I noticed in an MFC application (button getting stuck when using the spacebar). There's not much information in MSDN about this function, so I can't say if IsDialogMessage always behaves like this. Sander

3 2

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

wine-devel