- Wine-Bugs - list.winehq.org

[Bug 571] Changed - Read from socket hangs
by wine-bugs＠winehq.com April 11, 2002

April 11, 2002

http://bugs.codeweavers.com/show_bug.cgi?id=571 *** shadow/571 Thu Apr 11 02:26:29 2002 --- shadow/571.tmp.5513 Thu Apr 11 04:37:04 2002 *************** *** 66,68 **** --- 66,89 ---- ------- Additional Comments From jarto(a)starsoft.fi 2002-04-11 02:26 ------- The bug occurs with 20020310 too. I can test an earlier version if you just tell me which one to test. + + ------- Additional Comments From martin.wilck(a)fujitsu-siemens.com 2002-04-11 04:37 ------- + Please try if the following patch resolves your problem: + If not, please send me another trace (Btw how do you generate these beautiful + traces with server and client messages interleaved?) + + Patch URL: + http://www.winehq.com/hypermail/wine-patches/2002/04/0021.html + + Rationale: The created socket uses overlapped mode by default. The code calls + ReadFile() + with an OVERLAPPED structure to receive its answer. The ReadFile() request + succeeds immediately, and with the current CVS implementation the event of the + OVERLAPPED structure + is not set in the immediate completion case. The above patch changes the way + immediate completion is handled (especially, the event handling) and might + therefore be the solution. + + Please inform me of either success or failure. This patch is waiting for + acceptance, and your case + may be strong one if it helps you out :)

1 0

[Bug 571] Changed - Read from socket hangs
by wine-bugs＠winehq.com April 11, 2002

April 11, 2002

http://bugs.codeweavers.com/show_bug.cgi?id=571 *** shadow/571 Thu Apr 11 02:21:49 2002 --- shadow/571.tmp.31519 Thu Apr 11 02:26:29 2002 *************** *** 61,63 **** --- 61,68 ---- ------- Additional Comments From martin.wilck(a)fujitsu-siemens.com 2002-04-11 02:21 ------- Can you tell whether the problem also occurs with wine versions _before_ CVS 2002/04/05 ? + + + ------- Additional Comments From jarto(a)starsoft.fi 2002-04-11 02:26 ------- + The bug occurs with 20020310 too. I can test an earlier version if you just tell + me which one to test. \ No newline at end of file

1 0

[Bug 571] Changed - Read from socket hangs
by wine-bugs＠winehq.com April 11, 2002

April 11, 2002

http://bugs.codeweavers.com/show_bug.cgi?id=571 *** shadow/571 Thu Apr 11 02:16:17 2002 --- shadow/571.tmp.31357 Thu Apr 11 02:21:49 2002 *************** *** 57,59 **** --- 57,63 ---- ------- Additional Comments From martin.wilck(a)fujitsu-siemens.com 2002-04-11 02:14 ------- Well it's not right that I have "finished" my work on Winsock as Tony writes - but I'll take a look at your problem. Bug may be introduced by myself. + + ------- Additional Comments From martin.wilck(a)fujitsu-siemens.com 2002-04-11 02:21 ------- + Can you tell whether the problem also occurs with wine versions _before_ CVS + 2002/04/05 ?

1 0

[Bug 571] Changed - Read from socket hangs
by wine-bugs＠winehq.com April 11, 2002

April 11, 2002

http://bugs.codeweavers.com/show_bug.cgi?id=571 *** shadow/571 Thu Apr 11 02:14:49 2002 --- shadow/571.tmp.31191 Thu Apr 11 02:16:17 2002 *************** *** 9,14 **** --- 9,15 ---- +----------------------------------------------------------------------------+ | Assigned To: wine-bugs(a)winehq.com | | Reported By: jarto(a)starsoft.fi | + | CC list: Cc: martin.wilck(a)fujitsu- | +----------------------------------------------------------------------------+ | Milestone: TargetMilestone: --- | | URL: |

1 0

[Bug 571] Changed - Read from socket hangs
by wine-bugs＠winehq.com April 11, 2002

April 11, 2002

http://bugs.codeweavers.com/show_bug.cgi?id=571 *** shadow/571 Wed Apr 10 16:09:14 2002 --- shadow/571.tmp.31081 Thu Apr 11 02:14:49 2002 *************** *** 51,53 **** --- 51,58 ---- ------- 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. + + + ------- Additional Comments From martin.wilck(a)fujitsu-siemens.com 2002-04-11 02:14 ------- + Well it's not right that I have "finished" my work on Winsock as Tony writes - + but I'll take a look at your problem. Bug may be introduced by myself. \ No newline at end of file

1 0

[Bug 572] Changed - starcraft crashes in recent wine builds
by wine-bugs＠winehq.com April 10, 2002

April 10, 2002

http://bugs.codeweavers.com/show_bug.cgi?id=572 *** shadow/572 Wed Apr 10 22:09:19 2002 --- shadow/572.tmp.21846 Wed Apr 10 22:11:33 2002 *************** *** 47,49 **** --- 47,59 ---- ------- Additional Comments From mlearl(a)attbi.com 2002-04-10 22:09 ------- Created an attachment (id=61) gzip of last 100k lines of trace from --debugmsg +relay=kernel32 + + + ------- Additional Comments From mlearl(a)attbi.com 2002-04-10 22:11 ------- + Well, attachment doesn't display properly; probably I got the mime type wrong. + + It's also at + + http://mlearl.home.attbi.com/public/winedebugend.txt.gz + + for at least the near future. \ No newline at end of file

1 0

[Bug 572] Changed - starcraft crashes in recent wine builds
by wine-bugs＠winehq.com April 10, 2002

April 10, 2002

http://bugs.codeweavers.com/show_bug.cgi?id=572 *** shadow/572 Wed Apr 10 21:57:47 2002 --- shadow/572.tmp.21612 Wed Apr 10 22:09:19 2002 *************** *** 43,45 **** --- 43,49 ---- 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. + + ------- Additional Comments From mlearl(a)attbi.com 2002-04-10 22:09 ------- + Created an attachment (id=61) + gzip of last 100k lines of trace from --debugmsg +relay=kernel32

1 0

[Bug 572] New - starcraft crashes in recent wine builds
by wine-bugs＠winehq.com April 10, 2002

April 10, 2002

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

1 0

[Bug 571] Changed - Read from socket hangs
by wine-bugs＠winehq.com April 10, 2002

April 10, 2002

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.

1 0

[Bug 434] Changed - Mouse button don't get released (Macromedia Flash 5.0)
by wine-bugs＠winehq.com April 10, 2002

April 10, 2002

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

1 0