October 2003 - Wine-devel - list.winehq.org

Bug in cxx_frame_handler
by Robert Shearman 24 Oct '03

24 Oct '03

Hi, I believe there is a bug in cxx_frame_handler which causes an infinite loop of exceptions I believe the offending lines are 359-364. BTW, original exception was a call to NULL. Wine Trace: trace:seh:EXC_RtlRaiseException code=c0000005 flags=0 addr=(nil) trace:seh:EXC_RtlRaiseException info[0]=00000000 trace:seh:EXC_RtlRaiseException info[1]=00000000 trace:seh:EXC_CallHandler calling handler at 0x1019c7d code=c0000005 flags=0 trace:seh:EXC_CallHandler handler returned 1 trace:seh:EXC_CallHandler calling handler at 0x5f492e0a code=c0000005 flags=0 trace:seh:EXC_CallHandler handler returned 1 trace:seh:EXC_CallHandler calling handler at 0x5f492d82 code=c0000005 flags=0 trace:seh:EXC_RtlRaiseException code=c0000005 flags=0 addr=0x40774fa4 trace:seh:EXC_RtlRaiseException info[0]=00000000 trace:seh:EXC_RtlRaiseException info[1]=0000001c trace:seh:EXC_CallHandler calling handler at 0x401a6120 code=c0000005 flags=0 trace:seh:EXC_CallHandler handler returned 2 trace:seh:EXC_CallHandler calling handler at 0x1019c7d code=c0000005 flags=10 trace:seh:EXC_CallHandler handler returned 1 trace:seh:EXC_CallHandler calling handler at 0x5f492e0a code=c0000005 flags=10 trace:seh:EXC_CallHandler handler returned 1 trace:seh:EXC_CallHandler calling handler at 0x5f492d82 code=c0000005 flags=10 trace:seh:EXC_RtlRaiseException code=c0000005 flags=0 addr=0x40774fa4 trace:seh:EXC_RtlRaiseException info[0]=00000000 trace:seh:EXC_RtlRaiseException info[1]=0000001c trace:seh:EXC_CallHandler calling handler at 0x401a6120 code=c0000005 flags=0 trace:seh:EXC_CallHandler handler returned 2 trace:seh:EXC_CallHandler calling handler at 0x401a6120 code=c0000005 flags=10 trace:seh:EXC_CallHandler handler returned 2 trace:seh:EXC_CallHandler calling handler at 0x1019c7d code=c0000005 flags=10 trace:seh:EXC_CallHandler handler returned 1 trace:seh:EXC_CallHandler calling handler at 0x5f492e0a code=c0000005 flags=10 trace:seh:EXC_CallHandler handler returned 1 trace:seh:EXC_CallHandler calling handler at 0x5f492d82 code=c0000005 flags=10 trace:seh:EXC_RtlRaiseException code=c0000005 flags=0 addr=0x40774fa4 ... same messages repeated again and again. gdb backtrace: #50 0x4005f518 in sigaction () from /lib/libc.so.6 #51 0x40775260 in __CxxFrameHandler (rec=0x4063ef04, frame=0x4074fc7c, exc_context=0x4063ef78, dispatch=0x4063ee94, context=0x4063eae0) at cppexcept.c:392 #52 0x401c4155 in __wine_call_from_32_regs () from /usr/local/lib/wine/ntdll.dll.so #53 0x40770d0d in __wine_regs_msvcrt_dll_77 () from /usr/local/lib/wine/msvcrt.dll.so #54 0x401a651f in EXC_RtlRaiseException (rec=0x4063ef04, context=0x4063ef78) at exception.c:216 #55 0x401bc646 in do_segv (context=0x4063ef78, trap_code=14, cr2=0x1c, err_code=4) at signal_i386.c:851 #56 0x401bca82 in segv_handler (__signal=11, __context= {sc_gs = 5335, __gsh = 0, sc_fs = 143, __fsh = 0, sc_es = 43, __esh = 0, sc_ds = 43, __dsh = 0, sc_edi = 1080292788, sc_esi = 0, sc_ebp = 1080292716, sc_esp = 1080292648, sc_ebx = 1081684044, sc_edx = 1075713776, sc_ecx = 0, sc_eax = 0, sc_trapno = 14, sc_err = 4, sc_eip = 1081560996, sc_cs = 35, __csh = 0, sc_eflags = 66050, esp_at_signal = 1080292648, sc_ss = 43, __ssh = 0, i387 = 0, oldmask = 0, cr2 = 28}) at signal_i386.c:1057 #57 0x4005f518 in sigaction () from /lib/libc.so.6 #58 0x40775260 in __CxxFrameHandler (rec=0x4063f9dc, frame=0x4074fc7c, exc_context=0x4063fa50, dispatch=0x4063f96c, context=0x4063f5b8) at cppexcept.c:392 #59 0x401c4155 in __wine_call_from_32_regs () from /usr/local/lib/wine/ntdll.dll.so #60 0x40770d0d in __wine_regs_msvcrt_dll_77 () from /usr/local/lib/wine/msvcrt.dll.so #61 0x401a651f in EXC_RtlRaiseException (rec=0x4063f9dc, context=0x4063fa50) at exception.c:216 #62 0x401bc646 in do_segv (context=0x4063fa50, trap_code=14, cr2=0x0, err_code=4) at signal_i386.c:851 #63 0x401bca82 in segv_handler (__signal=11, __context= {sc_gs = 5335, __gsh = 0, sc_fs = 143, __fsh = 0, sc_es = 43, __esh = 0, sc_ds = 43, __dsh = 49168, sc_edi = 1081408212, sc_esi = 1081408216, sc_ebp = 1081408220, sc_esp = 1081408180, sc_ebx = 1087173732, sc_edx = 1077335880, sc_ecx = 1077404792, sc_eax = 0, sc_trapno = 14, sc_err = 4, sc_eip = 0, sc_cs = 35, __csh = 0, sc_eflags = 66118, esp_at_signal = 1081408180, sc_ss = 43, __ssh = 0, i387 = 1080294784, oldmask = 0, cr2 = 0}) at signal_i386.c:1057 #64 0x4005f518 in sigaction () from /lib/libc.so.6 disassemble of section around second exception, sc_eip = 1081560996 (0x40774fa4): 0x40774f78 <cxx_frame_handler+180>: push %esi 0x40774f79 <cxx_frame_handler+181>: push %edx 0x40774f7a <cxx_frame_handler+182>: mov 0x18(%ebp),%edx 0x40774f7d <cxx_frame_handler+185>: push %edx 0x40774f7e <cxx_frame_handler+186>: pushl 0x14(%ebp) 0x40774f81 <cxx_frame_handler+189>: pushl 0x10(%ebp) 0x40774f84 <cxx_frame_handler+192>: mov 0xc(%ebp),%ecx 0x40774f87 <cxx_frame_handler+195>: push %ecx 0x40774f88 <cxx_frame_handler+196>: mov 0x8(%ebp),%edx 0x40774f8b <cxx_frame_handler+199>: push %edx 0x40774f8c <cxx_frame_handler+200>: call *%eax 0x40774f8e <cxx_frame_handler+202>: jmp 0x40775225 <cxx_frame_handler+865> 0x40774f93 <cxx_frame_handler+207>: cmpl $0x0,0xffffffe4(%ebp) 0x40774f97 <cxx_frame_handler+211>: jne 0x40774faa <cxx_frame_handler+230> 0x40774f99 <cxx_frame_handler+213>: call 0x40787070 <msvcrt_get_thread_data> 0x40774f9e <cxx_frame_handler+218>: mov 0x1c(%eax),%eax 0x40774fa1 <cxx_frame_handler+221>: mov %eax,0x8(%ebp) 0x40774fa4 <cxx_frame_handler+224>: mov 0x1c(%eax),%ecx 0x40774fa7 <cxx_frame_handler+227>: mov %ecx,0xffffffe4(%ebp) 0x40774faa <cxx_frame_handler+230>: mov 0x1fc(%ebx),%edx

1 0

Re: wine/include setupapi.h
by Dimitrie O. Paun 23 Oct '03

23 Oct '03

On October 21, 2003 11:07 pm, Alexandre Julliard wrote: > Patch: > http://cvs.winehq.com/patch.py?root=/home/winehq/opt/cvs-commit&id=9836 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ We've changed the cvsroot, and it reared it's ugly head. Can we either (1) create a symlink from /opt/cvs-commit to the new root, and switch back to the old root (2) apply this patch to patch.py Index: patch.py =================================================================== RCS file: /home/wine/tools/patch.py,v retrieving revision 1.13 diff -u -r1.13 patch.py --- patch.py 27 Sep 2002 22:14:53 -0000 1.13 +++ patch.py 23 Oct 2003 13:34:38 -0000 @@ -49,7 +49,7 @@ # # The CVS root directory -root = '/opt/cvs-commit' +root = '/home/winehq/opt/cvs-commit' # The commit log file logs = 'CVSROOT/commitlog' -- Dimi.

2 1

re: Proposal to resolve the issue of the broken wine-users gateway
by Dan Kegel 23 Oct '03

23 Oct '03

Jer wrote: > Say, Just chatting on IRC today, and ShadowHawk pointed out > that a host of other sites (Franks corner, etc) have sprung up > almost entirely because there are no user based forums > on winehq.org. > > This is, in part, due to the fact that the gateway between > the newsgroup and wine-users is busted. And it is that way, > because the gateway we were using is a dead project, and Jer > hasn't been able to wrestle it back into working > (and doesn't really want to). > > Further, I think it may be time to recognize that cool looking > forums are what users expect; that mailing lists and newsgroups > are really becoming the 'old' way of doing things. Nah. Forums blow donkey dong because they aren't easily google-searchable. I think we can get what users want by asking http://www.gmane.org/ to handle the gateway; then users that want a web interface can trivially use a gateway like http://groups.google.com. And that'd be *even less* work for Jer than setting up and maintaining Phorum. - Dan -- Dan Kegel http://www.kegel.com http://counter.li.org/cgi-bin/runscript/display-person.cgi?user=78045

1 0

getting started with Winelib
by Ralf Juengling 23 Oct '03

23 Oct '03

Hi, I just installed wine-20031016. Trying to compile the test drive application as explained in the winelib user's guide 1.3.2, I end up getting the error message below. Any help appreciated, Ralf [juenglin(a)alpspitze winemine]$ ./configure --with-wine=/local checking whether make sets $(MAKE)... yes checking for gcc... gcc ... checking "for g++ -fno-for-scope option"... yes checking for windef.h... /local/include/wine/windows checking for windef.h header... /local/include/wine/windows checking for libwine.so... /local/lib checking for -lwine... /local/lib checking for libwine_unicode.so... /local/lib checking for -lwine_unicode... /local/lib checking for libwine_uuid.a... /local/lib checking for -lwine_uuid... /local/lib checking for libntdll.dll.so... configure: error: Could not find the Wine dlls (libntdll.dll.so) -- Ralf Juengling <juenglin(a)cse.ogi.edu>

1 0

Winesetuptk 0.7 rc1
by Ivan Leo Murray-Smith 23 Oct '03

23 Oct '03

Hi list, The rc1 of winesetuptk 0.7 is at http://www003.portalis.it/115/download/prerelease/winesetuptk0.7rc1.tar.bz2 This version is up to date with wine current, there are a few interesting changes, for example you can find the about box without reading the source, for details see the changelog.

3 2

Regression patch found - now what?
by Shachar 23 Oct '03

23 Oct '03

Hi all, I have identified that the commit at http://www.winehq.org/hypermail/wine-cvs/2003/09/0204.html is causing a regression to my application. This is the same application that kept generating the sigsegv and catching it not so long ago (http://www.winehq.org/hypermail/wine-devel/2003/10/0044.html) after applying this patch, the program will no longer run. I get a dialog saying "unexpected error". I suspect that moving the instruction emulation ruined, in some way, the application's ability to call cli and sti. Can anyone who knows what this patch was all about comment on what is going here? Shachar -- Shachar Shemesh Open Source integration consultant Home page & resume - http://www.shemesh.biz/

4 4

Re: winetests: new addition
by Rolf Kalbermatter 23 Oct '03

23 Oct '03

Message: 1 Date: Wed, 22 Oct 2003 10:10:58 +0200 From: Jakob Eriksson <jakov(a)vmlinux.org> To: dpaun(a)rogers.com CC: Ferenc Wagner <wferi(a)afavant.elte.hu>, wine-devel(a)winehq.org Subject: Re: winetests: new addition >My thought exactly. I also tried to write a recursive delete using >plain Win32 but failed miserably. :-) You mean something like this: (Stolen shamelessly and a little modified fror SHELL_DeleteDirectory in shlfileop.c) #define IsDotDir(x) ((x[0] == '.') && ((x[1] == 0) || ((x[1] == '.') && (x[2] == 0)))) static BOOL removeDir(LPCSTR dir) { BOOL ret = TRUE; HANDLE hFind; WIN32_FIND_DATAA wfd; CHAR path[MAX_PATH]; /* Make sure the directory exists before going further */ PathCombineA(path, dir, "*"); hFind = FindFirstFileA(path, &wfd); if (hFind == INVALID_HANDLE_VALUE) return ret; do { LPSTR lp = wfd.cFileName; if (!lp[0]) lp = wfd.cAlternateFileName; if (IsDotDir(lp)) continue; PathCombineA(path, dir, lp); if (FILE_ATTRIBUTE_DIRECTORY & wfd.dwFileAttributes) { ret = removeDir(path); } else ret = DeleteFileA(path); } while (ret && FindNextFileA(hFind, &wfd)); FindClose(hFind); RemoveDirectoryA(dir); return ret; } Rolf Kalbermatter

2 1

printing broke?
by Kevin DeKorte 23 Oct '03

23 Oct '03

All, I have a question about cups printing. I have had wine apps be able to print in the past, but after a recent reinstall of wine I can't seem to print anymore. I can print from Linux just fine and Wine sees my cups printers. But when I try to print from Notepad it does nothing and when I try to print from Notes I get an insuffiencent memory error. I have tried both a source (CVS current as of 10/21) install and the RPM (2003/10/16) from source forge and both give me the same results. I admit I have not tried the 2003/9/11 build yet. Anyone else seeing this? Kevin

1 0

(vstserver/winelib) What happened to pthread_create and pthread_join?
by Kjetil S. Matheussen 23 Oct '03

23 Oct '03

Hi, I when trying to use vstserver with the latest cvs-version of wine (23.10.2003), wine cant load the winelib program because pthread_create and pthread_join functions are missing. Of course, I do solve this problem just by linking with the glibc pthread library. But, what has happened? When reading scheduler/pthread.c, some radical change seems to have happened since the last time I read the file. In short, my question is: How do I configure/make/link my winelib program to get proper support for the pthreads commands with the new system? (And a related question, any plans to support pthread_setschedparam in the future?) --

1 0

Re: NTFS driver (was: ReactOS GPL vs. proprietary drivers)
by Szakacsits Szabolcs 23 Oct '03

23 Oct '03

Jakob Eriksson wrote: > Nevertheless, I too wonder what ntfsresize -i /device would do on a > Longhorn... is the command "ntfsresize" still there BTW? Hmmm, ntfsresize is an open source Linux util what I wrote over a year ago and today used by several distros to non-destructively resize NTFS safely, if feasible (it was written a "do it or leave it" way). There are several ways to try it out. If one has already Linux then just install ntfsprogs and try it on the Longhorn partition. If one has floppy then boot a floppy Linux (but not Tom's RTBT or Leka, they have 32 bit file size and other limits) and one might use the statically linked ntfsresize from the below page. If one has Knoppix or one of the other rescue, Live CD or distro CD that includes it then just boot it and run ntfsresize from a terminal, etc. I list here what I was told or know about who uses it and how, etc http://mlf.linux.rulez.org/mlf/ezaz/ntfsresize.html It's zero risk because the -i (info) options mounts the partition read-only. There can be 3 outcomes, 1. segfault -> NTFS changed significantly 2. change will be caught by libntfs or ntfsresize 3. no changes and it will work. I expect 2. Unfortunately I don't know any easier way how to figure out if there were changes or not. Szaka

1 0