April 2002 - Wine-Bugs - list.winehq.org

[Bug 572] New - starcraft crashes in recent wine builds

by wine-bugs＠winehq.com

http://bugs.codeweavers.com/show_bug.cgi?id=572 *** shadow/572 Wed Apr 10 21:57:47 2002 --- shadow/572.tmp.20975 Wed Apr 10 21:57:47 2002 *************** *** 0 **** --- 1,45 ---- + +============================================================================+ + | starcraft crashes in recent wine builds | + +----------------------------------------------------------------------------+ + | Bug #: 572 Product: Wine | + | Status: UNCONFIRMED Version: 20020228 | + | Resolution: Platform: | + | Severity: normal OS/Version: All | + | Priority: P1 Component: wine-loader | + +----------------------------------------------------------------------------+ + | Assigned To: wine-bugs(a)winehq.com | + | Reported By: mlearl(a)attbi.com | + | CC list: Cc: | + +----------------------------------------------------------------------------+ + | Milestone: TargetMilestone: --- | + | URL: | + +============================================================================+ + | DESCRIPTION | + Starcraft (video game, direct X) crashes in recent wine builds. Much + bug-chasing done base on advice from comp.emulations.ms-windows.wine: + + Trying different builds from local CVS, it appears builds prior to 31Jan2002 are + fine; the critical patch was: + + ***** + ChangeSet ID: 1012519978430168570484966 + CVSROOT: + /opt/cvs-commit + Module name: wine + Changes by: julliard(a)wine2. 02/01/31 18:32:58 + + Modified files: + loader : module.c pe_image.c pe_resource.c + + Log message: + LOAD_LIBRARY_AS_DATAFILE modules must be mapped like normal files, not + like PE images. Fixed resource loading to handle that. + ***** + + Builds which include this patch don't work for me; 'Unhandled exception' after + game setup when it starts to bring on the main playing display/GUI (works ok to + this point). + + Backtrace appears to be invalid/corrupted; if wine is run with --debugmsg +trace + it runs Very Very Slowly and works! Trying other variations to see if any + produce both the bug and a useful trace. \ No newline at end of file

April 11, 2002

[Bug 571] Changed - Read from socket hangs

by wine-bugs＠winehq.com

http://bugs.codeweavers.com/show_bug.cgi?id=571 *** shadow/571 Wed Apr 10 01:24:08 2002 --- shadow/571.tmp.3115 Wed Apr 10 16:09:14 2002 *************** *** 9,15 **** +----------------------------------------------------------------------------+ | Assigned To: wine-bugs(a)winehq.com | | Reported By: jarto(a)starsoft.fi | - | CC list: Cc: | +----------------------------------------------------------------------------+ | Milestone: TargetMilestone: --- | | URL: | --- 9,14 ---- *************** *** 48,50 **** --- 47,53 ---- I built wine from cvs yesterday before filing this bug and it had the same problem. I just forgot to mention that. So this is a real problem in the current source. + + ------- Additional Comments From tony_lambregts(a)telusplanet.net 2002-04-10 16:09 ------- + OK, I suppose we should have a version called CVS. Sorry for the noise. I have + informed Mr Wilck about this bug. Maybe he can help.

April 10, 2002

[Bug 434] Changed - Mouse button don't get released (Macromedia Flash 5.0)

by wine-bugs＠winehq.com

http://bugs.codeweavers.com/show_bug.cgi?id=434 *** shadow/434 Wed Apr 10 12:14:12 2002 --- shadow/434.tmp.27852 Wed Apr 10 14:27:55 2002 *************** *** 378,380 **** --- 378,423 ---- Hope to see this solved soon - I would love to be able to use these programs with wine. Cheers =) + + ------- Additional Comments From ivanovich(a)menta.net 2002-04-10 14:27 ------- + >I assume it was Alexandre's patch that moved functions into the wine server, + >but I wasn't able to confirm it for sure (intermediate patches wouldn't + >compile for me) + + Some time before i got a current wine cvs repository and updated and recompiled + wine and tested if the bug was present for every patch in the day in which wine + breaks, the patch http://cvs.winehq.com/patch.py?id=992978202334525382146239 + "Moved hardware message queue handling to the server." is for sure the one which + causes the bug. + + Did you talked with Alexandre about the bug? + + About what Rein Klazes asked: + "You could try to look for the WM_CAPTURECHANGED messages. This message + is sent when a window looses the capture. If I understand you correctly you + think that your problem is that this doesn't happen, while in the old situation + it did. If you have access to the spy++ + program, run it under windows and compare the messages between wine + and windows." + + the WM_CAPTURECHANGED msgs appear in user32 right? + i tried getting the debug with this: + wine --debugmsg +relay=user32-relay=GetSysColor flash.exe &> winedebugoutput.txt + + the last lines from it are: + 080714f8:Call window proc 0x406cf3c0 + (hwnd=00010024,msg=WM_MDIGETACTIVE,wp=00000000,lp=00000000) + 080714f8:Ret window proc 0x406cf3c0 + (hwnd=00010024,msg=WM_MDIGETACTIVE,wp=00000000,lp=00000000) retval=00020021 + 080714f8:Call window proc 0x759b5a + (hwnd=00020022,msg=WM_LBUTTONDOWN,wp=00000001,lp=004801ba) + fixme:imm:ImmGetContext (0x00020022): stub + -- 100% cpu , ctrl+c -- + + but i dont have spy++ (nor i know what is it)... + + maybe looking at the output of the working version of wine with this same debug + options we can locate the error? i think i can do it tonight + + btw, any idea what the last "fixme" line means? and the number 20022? (note that + retval from WM_MDIGETACTIVE is 20021) \ No newline at end of file

April 10, 2002

[Bug 434] Changed - Mouse button don't get released (Macromedia Flash 5.0)

by wine-bugs＠winehq.com

http://bugs.codeweavers.com/show_bug.cgi?id=434 *** shadow/434 Wed Apr 10 08:46:02 2002 --- shadow/434.tmp.19569 Wed Apr 10 12:14:12 2002 *************** *** 364,366 **** --- 364,380 ---- Since our apps seem to be responding so similarly, see if you can do what Rein Klazes is asking (I will try also). + + + ------- Additional Comments From anemo(a)silencegreys.com 2002-04-10 12:14 ------- + I get this same bug under Adobe Illustrator 8.0 - When using a tool or moving + any of the toolbars / tool windows around inside the program, mouse releases are + never recognized. + + I get the same gdb messages as reported by ivanovich(a)menta.net. There is a + freely available tryout download of Illustrator available at the Adobe website ( + http://www.adobe.com/support/downloads/product.jsp?product=27&platform=Wind… + ), so this might be a useful free test program for locating this bug. + + Hope to see this solved soon - I would love to be able to use these programs + with wine. Cheers =) \ No newline at end of file

April 10, 2002

[Bug 434] Changed - Mouse button don't get released (Macromedia Flash 5.0)

by wine-bugs＠winehq.com

http://bugs.codeweavers.com/show_bug.cgi?id=434 *** shadow/434 Tue Apr 9 15:55:06 2002 --- shadow/434.tmp.4110 Wed Apr 10 08:46:02 2002 *************** *** 339,341 **** --- 339,366 ---- note: someone in #wine-devel at irc.opensource.net told me maybe it was an error in the communication with the server :-? + + ------- Additional Comments From rick(a)valeoinc.com 2002-04-10 08:46 ------- + Strange the debug info hack didn't work.. I just reformatted, and installed + Mandrake 8.2, with KDE 3, and the debug info thing still worked. + + I don't know how to get line number info on debugging.. If you want to scan the + wine-users list archives 2/02, I attempted to trace back the wine calls to try and + see what wasn't returning.. There's a thread (Help tracing WINE function) + between myself and Lawson Whitney that basically ends with "do regression + testing, the app itself is doing MSVC++ calls, and ask the author of the patch to + fix it". + + I assume it was Alexandre's patch that moved functions into the wine server, but + I wasn't able to confirm it for sure (intermediate patches wouldn't compile for me), + and after stating my assumptions, it never went anywhere. + + I think at that point is when the same issue in Adobe Illustrator came up in the + wine-users list, and you added on 'me too for Flash' :) + + Scanning that wine-users archive thread, I think I may have missed this last + message: http://www.winehq.com/hypermail/wine-users/2002/02/0216.html + + Since our apps seem to be responding so similarly, see if you can do what Rein + Klazes is asking (I will try also). +

April 10, 2002

[Bug 571] Changed - Read from socket hangs

by wine-bugs＠winehq.com

http://bugs.codeweavers.com/show_bug.cgi?id=571 *** shadow/571 Tue Apr 9 21:16:54 2002 --- shadow/571.tmp.2848 Wed Apr 10 01:24:08 2002 *************** *** 42,44 **** --- 42,50 ---- Your choice pick one <hint> *Pick 1!* </hint> If you do choose the join the cvs crowd and need assitance I will be glad to help. + + + ------- Additional Comments From jarto(a)starsoft.fi 2002-04-10 01:24 ------- + I built wine from cvs yesterday before filing this bug and it had the same + problem. I just forgot to mention that. So this is a real problem in the current + source. \ No newline at end of file

April 10, 2002

[Bug 571] Changed - Read from socket hangs

by wine-bugs＠winehq.com

http://bugs.codeweavers.com/show_bug.cgi?id=571 *** shadow/571 Tue Apr 9 15:59:57 2002 --- shadow/571.tmp.25614 Tue Apr 9 21:16:54 2002 *************** *** 26,28 **** --- 26,44 ---- address, port and a command to send. The received reply is shown in a memo. In this testprog I read from TSocketStream one char at a time. It's definitely not a fast way but easy enough for this demonstration. + + ------- Additional Comments From tony_lambregts(a)telusplanet.net 2002-04-09 21:16 ------- + Martin Wilck has just finished doing some work in winsock and between the looks + of your trace and looking at what martin did this problem may already be solved. + Unfortunately I do not have Delphi so I cannot test your program. You have these + options to see if I am correct. + + 1.Install wine from CVS and test it yourself. + + 2.Wait for the next release of wine and test it yourself. + + 3.Cross your fingers and hope that someone with Delphi and Wine CVS will test + + Your choice pick one <hint> *Pick 1!* </hint> + + If you do choose the join the cvs crowd and need assitance I will be glad to help.

April 10, 2002

[Fwd: bug in newfns.c]

by Jeremy White

-----Forwarded Message----- > From: iMil <imil(a)home.imil.net> > To: bugs(a)codeweavers.com > Subject: bug in newfns.c > Date: 06 Apr 2002 04:25:32 +0200 > > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > > Hi wine enthusiasts, > > I think I found a bug in newfns.c. The QUERYPERF_Init() function causes a > very nasty crash on certain apps, just after calculation of the CPU speed. > Here I have a 1.1GHz AMD processor, just after printing the trace : > > using frequency: 1100700000Hz > > wine instantly crashes. Some friends of mine that have lower-speed > processors do not have this error. I find out this w/ the -debugmsg +win32 > debug mode. > > Thanks for the great work, > Truely yours, > > - ------------------------- > iMil <imil(a)home.imil.net> > http://gcu-squad.org > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.0.6 (NetBSD) > Comment: For info see http://www.gnupg.org > > iD8DBQE8rlyfFG3BlGWyzUIRAtAzAJ0WurhM33Of21PXozUf+BEwyfK8BgCeLsZq > dTN063lHvyeI/qQY8Qlv858= > =/Hun > -----END PGP SIGNATURE-----

April 9, 2002

[Bug 571] New - Read from socket hangs

by wine-bugs＠winehq.com

http://bugs.codeweavers.com/show_bug.cgi?id=571 *** shadow/571 Tue Apr 9 15:59:57 2002 --- shadow/571.tmp.10470 Tue Apr 9 15:59:57 2002 *************** *** 0 **** --- 1,28 ---- + +============================================================================+ + | Read from socket hangs | + +----------------------------------------------------------------------------+ + | Bug #: 571 Product: Wine | + | Status: UNCONFIRMED Version: 20020310 | + | Resolution: Platform: | + | Severity: normal OS/Version: All | + | Priority: P1 Component: wine-net | + +----------------------------------------------------------------------------+ + | Assigned To: wine-bugs(a)winehq.com | + | Reported By: jarto(a)starsoft.fi | + | CC list: Cc: | + +----------------------------------------------------------------------------+ + | Milestone: TargetMilestone: --- | + | URL: | + +============================================================================+ + | DESCRIPTION | + Wine hangs when a Delphi 5 app reads from a socket using TWinSocketStream.Read. + + I've made a small test program that demonstrates the problem. An executable + + Delphi source is available in http://www.starsoft.fi/jarto/SocketTest.tar.gz + + Last 500 lines of --debugmsg +all: http://www.starsoft.fi/jarto/socketlog.txt + + The test app is not very elegant but gets the job done. You can define an ip + address, port and a command to send. The received reply is shown in a memo. In + this testprog I read from TSocketStream one char at a time. It's definitely not + a fast way but easy enough for this demonstration. \ No newline at end of file

April 9, 2002

[Bug 434] Changed - Mouse button don't get released (Macromedia Flash 5.0)

by wine-bugs＠winehq.com

http://bugs.codeweavers.com/show_bug.cgi?id=434 *** shadow/434 Tue Apr 9 11:29:46 2002 --- shadow/434.tmp.10102 Tue Apr 9 15:55:06 2002 *************** *** 309,311 **** --- 309,341 ---- It may very well be the same issue, with the same fix, just rearing it's ugly head differently. + + + ------- Additional Comments From ivanovich(a)menta.net 2002-04-09 15:55 ------- + I "installed" your program (i used vmware, lazy me :) + + I can reproduce the bug just clicking into the text windows, same sympthoms as + in Flash. + + I get some read errors from strace but only when i move the mouse over the + window after the bug have made your program use 100%cpu (maybe it only tries to + search for a mouse release event when it's moved?) + + Your debug info hack didnt worked here (i put my P3-866 to 733 like yours and + neither worked with flash nor with your program, also tried with 433 and 910 mhz + in flash only) + + the gdb output from your program while it's in the bug is the same (or very + similar) last lines are this: + + Single stepping until exit from function X11DRV_GetCursorPos, + which has no line number information. + 0x406cbb19 in GetCursorPos () from /usr/local/lib/wine/user32.dll.so + + btw do you know how to enable line number information for debugging? maybe i + have to ./configure with a special option? + + ah wine version is cvs from yesterday or 2 days ago + + note: someone in #wine-devel at irc.opensource.net told me maybe it was an error + in the communication with the server :-? \ No newline at end of file

April 9, 2002