- Wine-Devel - list.winehq.org

[Alsa] A simple test for IDirectSound.GetCaps.
by Sylvain Petreolle Dec. 6, 2002

Dec. 6, 2002

Running the dsound test using winealsa.drv, the program hangs and try to call the debugger. [syl(a)snoop dsound]$ make test make[1]: Entre dans le répertoire `/home/wine/dlls/dsound/tests' ../../../tools/runtest -q -P wine -M dsound.dll -T ../../.. -p dsound_test.exe.so dsound.c && touch dsound.ok err:ntdll:RtlpWaitForCriticalSection section 0x400ec230 "../../loader/module.c: loader_section" wait timed out, retrying (60 sec) tid=0876cdc8 wine: Unhandled exception, starting debugger... err:ntdll:RtlpWaitForCriticalSection section 0x400ec230 "../../loader/module.c: loader_section" wait timed out, retrying (60 sec) tid=0876cdc8 err:seh:EXC_DefaultHandling Unhandled exception code c0000194 flags 0 addr 0x400bb478 make[1]: Quitte le répertoire `/home/wine/dlls/dsound/tests' According to http://www.delphifaq.com/fq/q1061.shtml, exception c0000194 is STATUS_POSSIBLE_DEADLOCK. How could I debug the tests ? ===== Sylvain Petreolle spetreolle(a)users.sourceforge.net Fight against Spam ! http://www.euro.cauce.org/en/index.html ICQ #170597259 "Don't think you are. Know you are." Morpheus in Matrix, chapter 15. ___________________________________________________________ Soyez solidaire soutenez laction du Téléthon avec Yahoo! France. http://www1.telethon.fr/030-Espace-Relais-Dons/webtirelire1.asp?hebergeur_i…

1 0

WineBuild/WineLib <-> MFC
by jaymz＠free.fr Dec. 5, 2002

Dec. 5, 2002

Going further into the MFC dll handling, I raised the following issue: I have two dynamic libs (mfc.dll.so and myapp.dll.so), both loaded by the wrapper (myapp.exe.so). The problem is that global or static class variables are badly exported from mfc.dll.so to myapp.dll.so. Here's an example: In mfc.dll.so, let assume that I export a global integer variable named myVar. I add @ extern myVar myVar in mfc.dll.spec. When I do a GetProcAddress of myVar from the wrapper, the returned address is good, but when I link my application, myVar is undefined (WineBuild did not include it in myapp.dll.spec.c). If I add instead in the spec file something like @ stdcall myVar() myVar, it's still good when using GetProcAdress, I can now link my application, but at runtime, the address of the variable is not the good one, so I'm using corrupted data ! Therefore, when the app runs, all static data are in fact not or badly initialized and it crashes. What's the exact syntact to correctly export class static variable from a dll ? Thanks a lot for your attention.

2 1

[RFC] wrc & windres
by Dimitrie O. Paun Dec. 5, 2002

Dec. 5, 2002

Hi folks, As we discussed here a few weeks ago, there is good reason to have a windres compatible mode for wrc. Needless to say, this will greatly ease porting apps that use windres to Winelib. More information about the issues at hand can be found in this thread: http://www.winehq.com/hypermail/wine-devel/2002/11/index.html#1518 These messages in particular deal with this problem: http://www.winehq.com/hypermail/wine-devel/2002/11/1546.html http://www.winehq.com/hypermail/wine-devel/2002/11/1550.html http://www.winehq.com/hypermail/wine-devel/2002/11/1551.html http://www.winehq.com/hypermail/wine-devel/2002/11/1552.html http://www.winehq.com/hypermail/wine-devel/2002/11/1555.html That being said, what would it take to (1) keep the current wrc behavior, and (2) make it compatible with windres? First, windres supports any conversions between .rc, .res, .o files, whereas wrc supports only .rc -> .res. It would be interesting (from the Winelib point of view) to also support .rc and .res -> .o. The other transformations supported by windres (.o -> .res -> .rc) are not as interesting, as they are not used in the build process. With that in mind, what do we need to do to have said compatibility? Well there are two things: 1. Support windres command line options 2. Add conversion of .rc and .res to .o I will treat each of these below. Command line compatibility -------------------------- Luckily, there is only one (seldom used) option that conflicts: -I. Moreover, it has a long format name in windres as well, which can be used instead. In other words, it should be very simple to add support for the windres options to wrc (most of them will just get ignored). In other words we can support windres's options while keeping full backwords compatibility. Having this compatibility makes sense as long as we support the conversion to .o files as well. Assuming we can do that, I would say this is something we should have. Conversion to .o files ---------------------- As Alexandre pointed out in here: http://www.winehq.com/hypermail/wine-devel/2002/11/1550.html generating .o files in proper format is non-trivial. And since I'm willing to do the work, there's no chance I'm going to propose something difficult to implement! :) What I have in mind is quite simple: store the .res file, as a byte array, in a .wine.resources ELF section, and have winebuild pick it up from there (as an alternative to getting it directly from the .res file, so current behavior is conserved). With these two changes, we eliminate the biggest nastiness of getting apps compiled with Winelib. The modification will typically be: -WINDRES=windres +WINDRES=wrc Moreover, the solution is: o Simple to implement o Fully backwards compatible with current Wine practice o Fully compatible with Win32 apps using the MinGW toolchain o Supports multiple resource files, etc. So now the big question: Alexandre, are patches that implement these ideas acceptable? Sorry for the long message, and thanks for reading this far. -- Dimi.

3 6

Re: [RFC] wrc & windres
by Jeff Smith Dec. 5, 2002

Dec. 5, 2002

>From: "Dimitrie O. Paun" <dpaun(a)rogers.com> >Hi folks, > >As we discussed here a few weeks ago, there is good reason >to have a windres compatible mode for wrc. <snip> >That being said, what would it take to (1) keep the current wrc >behavior, and (2) make it compatible with windres? Here is an alternative idea for dealing with command-line incompatibility: create a link 'windres' pointing to 'wrc'. In the program, we could determine how the program was invoked and handle the command-line options accordingly. This should technically work, unless we find some reason to 'mix' command line options, in which case we we would have to start 'blending' the two syntaxes. Please note that I am not familiar with the command line options of either program. <snip> >The modification will typically be: >-WINDRES=windres >+WINDRES=wrc WINDRES would not have to be changed using this method. -- Jeff S. _________________________________________________________________ The new MSN 8: smart spam protection and 2 months FREE* http://join.msn.com/?page=features/junkmail

3 2

Re: possible memory corruption
by steve.lustbader＠philips.com Dec. 5, 2002

Dec. 5, 2002

>>>>>> "steve" == steve lustbader <steve.lustbader(a)philips.com> writes: > > steve> When I try to load a window in my app that has a treeview in it, > steve> it crashes immediately, before the window even gets displayed. > steve> This is the end of the backtrace: > > Try running with native commctrl to see if the corruption comes from our > commctrl implementation. Then run with --debugmsg +relay,+heap. Wine will > check on every heap manipulation to see if the heap is still okay. Then >look what corrupted the heap. It works fine with a native (WinXP) comctl32.dll. I'll post the +relay,+heap output shortly. -Steve

1 0

possible memory corruption
by steve.lustbader＠philips.com Dec. 5, 2002

Dec. 5, 2002

When I try to load a window in my app that has a treeview in it, it crashes immediately, before the window even gets displayed. This is the end of the backtrace: 0x400d6377 (HEAP_FindFreeBlock+0x3b [heap.c:608] in libntdll.dll.so): movl 0x0(%edx),%eax 608 DWORD arena_size = (pArena->size & ARENA_SIZE_MASK) + Wine-dbg>bt Backtrace: =>0 0x400d6377 (HEAP_FindFreeBlock+0x3b(heap=0x41230000, size=0x108, ppSubHeap=0x406e1af4) [heap.c:608] in libntdll.dll.so) (ebp=406e1adc) 1 0x400d6df5 (RtlAllocateHeap+0x71(heap=0x41230000, flags=0xa, size=0x104) [heap.c:1028] in libntdll.dll.so) (ebp=406e1b04) 2 0x40b283e2 (COMCTL32_Alloc+0x4a(dwSize=0x104) [comctl32undoc.c:352] in comctl32.dll.so) (ebp=406e1b24) 3 0x40b732bb (TREEVIEW_UpdateDispInfo+0x227(infoPtr=0x41230838, wineItem=0x41232060, mask=0x1) [treeview.c:762] in comctl32.dll.so) (ebp=406e1ba4) 4 0x40b73d94 (TREEVIEW_InsertItemA+0x324(infoPtr=0x41230838, lParam=0x406e1be8) [treeview.c:1318] in comctl32.dll.so) (ebp=406e1bd8) 5 0x40b73eb6 (TREEVIEW_InsertItemW+0xb2(infoPtr=0x41230838, lParam=0x406e1f08) [treeview.c:1389] in comctl32.dll.so) (ebp=406e1c34) 6 0x40b79881 (TREEVIEW_WindowProc+0x6a5(hwnd=0x2006a, uMsg=0x1132, wParam=0x0, lParam=0x406e1f08) [treeview.c:5237] in comctl32.dll.so) (ebp=406e1c50) 7 0x408ec923 (WINPROC_wrapper+0x17 in user32.dll.so) (ebp=406e1c74) 8 0x408ec9ae (WINPROC_CallWndProc+0x82(proc=0x40b791dc, hwnd=0x2006a, msg=0x1132, wParam=0x0, lParam=0x406e1f08) [winproc.c:183] in user32.dll.so) (ebp=406e1ca4) I had thought originally that it might be a treeview problem, since the crash happens based on calls from treeview, but Dimi says that it's probably memory getting corrupted somewhere. What trace would be helpful to debug this? Thanks, Steve, PS - I had previously posted a +treeview trace a couple of weeks ago. Here it is again, just in case it's helpful: <App starting up> trace:treeview:TREEVIEW_Register trace:treeview:TREEVIEW_Register trace:treeview:TREEVIEW_Register <I press button to open dialog containing treeview> trace:treeview:TREEVIEW_Create wnd 0x2006a, style 40010027 trace:treeview:TREEVIEW_SetFirstVisible (nil): <null item> trace:treeview:TREEVIEW_SetFont 0x3a6a 0 trace:treeview:TREEVIEW_Create wnd 0x20068, style 40010027 trace:treeview:TREEVIEW_SetFirstVisible (nil): <null item> trace:treeview:TREEVIEW_SetFont 0x3a6a 0 trace:treeview:TREEVIEW_DeleteItem TVI_ROOT trace:treeview:TREEVIEW_ValidItem invalid item (nil) trace:treeview:TREEVIEW_ValidItem invalid item (nil) trace:treeview:TREEVIEW_ValidItem invalid item (nil) trace:treeview:TREEVIEW_ValidItem invalid item (nil) trace:treeview:TREEVIEW_SetFirstVisible (nil): <null item> trace:treeview:TREEVIEW_DeleteItem TVI_ROOT trace:treeview:TREEVIEW_ValidItem invalid item (nil) trace:treeview:TREEVIEW_ValidItem invalid item (nil) trace:treeview:TREEVIEW_ValidItem invalid item (nil) trace:treeview:TREEVIEW_ValidItem invalid item (nil) trace:treeview:TREEVIEW_SetFirstVisible (nil): <null item> trace:treeview:TREEVIEW_InsertItemA parent 0x41230a98 position 0xffff0002: '<callback>' trace:treeview:TREEVIEW_DoSetItem setting callback, item 0x41232060 trace:treeview:TREEVIEW_InsertItemA new item 0x41232060; parent 0x41230a98, mask 5 trace:treeview:TREEVIEW_GetNextItem flags:3, item 0x41232060;returns (nil) wine: Unhandled exception, starting debugger...

2 1

valgrind and wine
by Dan Kegel Dec. 5, 2002

Dec. 5, 2002

valgrind wine reports: ==6476== Valgrind detected that your program requires ==6476== the following unimplemented functionality: ==6476== modify_ldt(): I (JRS) haven't investigated this yet; sorry. ==6476== This may be because the functionality is hard to implement, ==6476== or because no reasonable program would behave this way, ==6476== or because nobody has yet needed it. In any case, let me know ==6476== (jseward(a)acm.org) and/or try to work around the problem, if you can. Dang. - Dan

3 4

RE: Remnants of -DSTRICT Conversion
by Patrik Stridvall Dec. 5, 2002

Dec. 5, 2002

> Just a small request. Could someone write a script that would > remove one of > the last remnants of the -DSTRICT conversion. It is a little > issue with lots > of 0x%p's everywhere (mainly in NTDLL registry functions) > which need to be > replaced by just %p's. Sure here is the script: find . -name \*.c | xargs perl -i -pe 's/0x%p/%p/g' and as a bonus to Alexandre here is the patch. --8<--- *** strict * dlls/commdlg/filedlg95.c, dlls/ntdll/file.c, dlls/ntdll/heap.c, dlls/ntdll/misc.c, dlls/ntdll/nt.c, dlls/ntdll/om.c, dlls/ntdll/reg.c, dlls/ntdll/sec.c, dlls/ntdll/sync.c, loader/module.c, scheduler/process.c, scheduler/thread.c, win32/device.c: Replace 0x%p with %p.

1 0

Remnants of -DSTRICT Conversion
by Robert Shearman Dec. 5, 2002

Dec. 5, 2002

Hi all, Just a small request. Could someone write a script that would remove one of the last remnants of the -DSTRICT conversion. It is a little issue with lots of 0x%p's everywhere (mainly in NTDLL registry functions) which need to be replaced by just %p's. Thanks, Rob

1 0

Use COMSPEC command interpreter to run .bat files.
by Sylvain Petreolle Dec. 5, 2002

Dec. 5, 2002

Is your patch allowing the run of .bat files by wine itself, or only a W-process ? When I try it on command line I have this : [syl(a)snoop cygwin]$ wine cygwin.bat --debugmsg +process trace:process:PROCESS_InitWine starting process name="" file=(nil) argv[0]="cygwin.bat" trace:process:find_exe_file looking for "cygwin.bat" trace:process:find_exe_file Trying built-in exe "C:\\cygwin\\cygwin.bat" trace:process:find_exe_file Trying native exe "C:\\cygwin\\cygwin.bat" wine: cannot determine executable type for 'C:\cygwin\cygwin.bat' [syl(a)snoop cygwin]$ cat cygwin.bat @echo off C: chdir \cygwin\bin bash --login -i ===== Sylvain Petreolle spetreolle(a)users.sourceforge.net Fight against Spam ! http://www.euro.cauce.org/en/index.html ICQ #170597259 "Don't think you are. Know you are." Morpheus in Matrix, chapter 15. ___________________________________________________________ Soyez solidaire soutenez laction du Téléthon avec Yahoo! France. http://www1.telethon.fr/030-Espace-Relais-Dons/webtirelire1.asp?hebergeur_i…

5 14