http://bugs.winehq.org/show_bug.cgi?id=13420
Summary: Hang on first keypress Product: Wine Version: 1.0-rc2 Platform: PC OS/Version: Linux Status: UNCONFIRMED Severity: enhancement Priority: P2 Component: -unknown AssignedTo: wine-bugs@winehq.org ReportedBy: flobber@mailinator.com
Start a simple app (i.e. winecfg or regedit). Everything seems to work fine.
Press any key (even shift is ok): the app hangs completely - no redraw, nothing.
Can be kill with CTRL-C though.
First noticed with .rc1, still present in .rc2. Last version tested without that bug: .58
Ciao.
http://bugs.winehq.org/show_bug.cgi?id=13420
Flob Flobber flobber@mailinator.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Severity|enhancement |blocker
http://bugs.winehq.org/show_bug.cgi?id=13420
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Severity|blocker |major
--- Comment #1 from Vitaliy Margolen vitaliy@kievinfo.com 2008-05-25 11:15:58 --- Not critical. Please run a regression test as described here: http://wiki.winehq.org/RegressionTesting
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #2 from Flob Flobber flobber@mailinator.com 2008-05-25 11:54:42 --- Well, not critical for you but for me, Wine is completely unusable - in other words, a blocker.
For the record, I'm using old-style xmodmap keyboard settings - xkb is disabled.
My keyboard layout is non-standard but even with xorg's standard layout the app hangs.
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #3 from Flob Flobber flobber@mailinator.com 2008-05-25 12:20:24 --- Regarding the "regression testing":
I don't have the necessary tools to compile wine. I'm using prebuild rpm's for suse.
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #4 from Dmitry Timoshkov dmitry@codeweavers.com 2008-05-25 22:20:48 --- Please generate a log WINEDEBUG=+x11drv,+key,+keyboard,+event wine notepad &>~/kbd_hang.log
Compress kbd_hang.log with 'bzip2 -9' and attach it here.
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #5 from Felix Saphir felix.saphir@presswatch.de 2008-05-26 08:55:09 --- Created an attachment (id=13350) --> (http://bugs.winehq.org/attachment.cgi?id=13350) Created with "WINEDEBUG=+x11drv,+key,+keyboard,+event wine notepad.exe"
I've encountered the same problem with notepad dying "during" the first keypress, the char is not even displayed. Log created like instructed and attached.
http://bugs.winehq.org/show_bug.cgi?id=13420
Felix Saphir felix.saphir@presswatch.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |felix.saphir@presswatch.de
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #6 from Flob Flobber flobber@mailinator.com 2008-05-26 17:29:39 --- Created an attachment (id=13375) --> (http://bugs.winehq.org/attachment.cgi?id=13375) Output of "WINEDEBUG=+x11drv,+key,+keyboard,+event wine notepad"
As requested, the log of starting notepad with the given WINEDEBUG settings and pressing the right shift key. The cursor stops blink at the moment I press the shift key - I don't have to release it. Then killed notepad with CTRL-C.
http://bugs.winehq.org/show_bug.cgi?id=13420
Flob Flobber flobber@mailinator.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #13375|text/plain |application/octet-stream mime type| |
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #7 from Vitaliy Margolen vitaliy@kievinfo.com 2008-05-26 22:10:45 --- What distros are you using? What xorg versions (X -version)? Wine not even receiving the event - should see something like this:
trace:event:call_event_handler KeyPress for hwnd/window 0x10024/4600005 trace:key:X11DRV_KeyEvent type 2, window 4600005, state 0x0010, keycode 0x0032 trace:key:X11DRV_KeyEvent nbyte = 0, status 0x3 trace:key:X11DRV_KeyEvent KeyPress : keysym=FFE1 (Shift_L), # of chars=0 / ""
Are you using any extra input modules (xim, xcim)?
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #8 from Flob Flobber flobber@mailinator.com 2008-05-27 01:16:45 --- SuSE 9.3 with Xorg 7.1.99.2.
No input methods. No Xkb. No deadkeys. German keyboard but non-standard layout.
Just checked: there is no KeyEvent. The only message that still comes when the app is stuck is this when I enter/leave the window with the mouse:
| trace:event:call_event_handler PropertyNotify for hwnd/window 0x10020/4d | trace:event:call_event_handler PropertyNotify for hwnd/window 0x10020/4d | trace:event:process_events processed 2 events
No expose or motion events any more nor a ClientMessage for the window-close button.
Ciao.
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #9 from Dmitry Timoshkov dmitry@codeweavers.com 2008-05-27 01:29:53 --- Does 'xev' show KeyPress events?
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #10 from Flob Flobber flobber@mailinator.com 2008-05-27 01:33:39 ---
Does 'xev' show KeyPress events?
Sure.
KeyPress event, serial 28, synthetic NO, window 0x2200001, root 0x4d, subw 0x2200002, time 689206011, (38,63), root:(1296,756), state 0x0, keycode 62 (keysym 0xffe2, Shift_R), same_screen YES, XLookupString gives 0 bytes: XmbLookupString gives 0 bytes: XFilterEvent returns: False
KeyRelease event, serial 28, synthetic NO, window 0x2200001, root 0x4d, subw 0x2200002, time 689206147, (38,63), root:(1296,756), state 0x1, keycode 62 (keysym 0xffe2, Shift_R), same_screen YES, XLookupString gives 0 bytes:
http://bugs.winehq.org/show_bug.cgi?id=13420
Dmitry Timoshkov dmitry@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Severity|major |normal
--- Comment #11 from Dmitry Timoshkov dmitry@codeweavers.com 2008-05-27 01:55:49 --- Can you try an older Wine release, say 0.9.55 or even older? There there a lot of changes recently in XIM support, that may has broken something for you.
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #12 from Felix Saphir felix.saphir@presswatch.de 2008-05-27 02:10:16 --- I've got two setups on identical servers: one is working fine with 0.9.57, the other one had 0.9.48 running and was upgraded to 1.0. Now applications hang on first keypress (didn't touch the first server). I tried .57 on the second server as well, but .48 was the last working version. Both servers run CentOS 4 with kernel 2.6.9-67 and KDE in a VNC session.
http://bugs.winehq.org/show_bug.cgi?id=13420
Dmitry Timoshkov dmitry@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression
--- Comment #13 from Dmitry Timoshkov dmitry@codeweavers.com 2008-05-27 02:18:45 --- Could you please run the regression test as requested in the Comment #1 ?
http://wiki.winehq.org/RegressionTesting
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #14 from Felix Saphir felix.saphir@presswatch.de 2008-05-27 08:52:08 --- Phew, regression testing is quite lengthy :) However, I ran into a compile error: | wrap.c: In function `ME_WrapTextParagraph': | wrap.c:356: error: syntax error before "wc"
The last version successfully compiled was 0.9.52.88-something, now the first lines from "git log" are: | commit 28cc0f9e690a675aa00b0ccc02b442fcf8a819ea | Author: Eric Pouech eric.pouech@orange.fr | Date: Tue Jan 1 22:04:40 2008 +0100 | richedit: Fix the para computation when zoom is used.
"git describe" gives: | wine-0.9.52-102-g28cc0f9
Should I mark this bisection as bad and continue?
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #15 from Dmitry Timoshkov dmitry@codeweavers.com 2008-05-27 09:08:44 --- Just copy wrap.c from a working version.
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #16 from Lei Zhang thestig@google.com 2008-05-27 15:11:48 --- Are you running wine inside vnc? If so, this is probably a duplicate of bug 11822.
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #17 from Felix Saphir felix.saphir@presswatch.de 2008-05-27 16:40:03 --- Could be a duplicate for what I have here (two ident. servers over VNC, and PS/2 keyboard on the client, no kbd on the servers, also see #12). Flob, how about you? I'll continue the regresssion tomorrow, and try what's suggested in #26 bug #11822. Little side note: Is it okay to take part, although I'm not the reporter? I'm willing to help as best as I can.
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #18 from Flob Flobber flobber@mailinator.com 2008-05-27 17:23:43 --- Indeed, the vnc bug looks similar.
But... here wine inside Xvnc (v6.8.2) works fine.
# wine notepad press any key in notepad - hang (btw, not a CPU hog hang - system is 100% idle).
but
# Xvnc :1 & # vncviewer :1 & # DISPLAY=:1 wine notepad
works. The only noticable effect is that terminating notepad takes exceptionally long. After 3-4 seconds this message appears:
| fixme:event:wait_for_withdrawn_state window 0x20028/a00013 wait timed out | fixme:event:wait_for_withdrawn_state window 0x10024/a00005 wait timed out
Hmm... I think I found something. After these two lines I thought "maybe this is due to a missing window manager..." So I started twm and bang, the problem occurs - notepad hangs on first keypress. When I terminate twm, notepad runs fine again. Same with windowmaker.
Summing up: wine notepad within an Xvnc runs fine but takes long to terminate (with error message) as long as no window manager is running. Either with twm or windowmaker the hang occures.
Btw, my regular window manager is icewm.
More info: if I enable "Emulate a virtual desktop" the problem vanishes - notepad works.
Maybe related: If I disable all options in "winecfg.Graphics. Window_Options" the wine window never gets focus and thus no key event at all.
Ciao.
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #19 from Alexandre Julliard julliard@winehq.org 2008-05-28 03:33:27 --- When notepad is hung please start winedbg and do a "bt all", and attach the whole output. Thanks.
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #20 from Felix Saphir felix.saphir@presswatch.de 2008-05-28 04:54:52 --- Regression finished, same bad commit as in bug #11822: | e14261dbb737e6177be95df6cea787401c1f7ad1 is first bad commit | commit e14261dbb737e6177be95df6cea787401c1f7ad1 | Author: Alexandre Julliard julliard@winehq.org | Date: Wed Feb 6 20:37:41 2008 +0100 | winex11: Use the gdi display when initializing the keyboard to avoid creating a second display connection. | :040000 040000 2249ad98256bc020aea00cdd47b869d5f2f43974 6165e5d183c78ba56618be362c5a8f79563bc3f2 M dlls
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #21 from Felix Saphir felix.saphir@presswatch.de 2008-05-28 04:56:33 --- Created an attachment (id=13428) --> (http://bugs.winehq.org/attachment.cgi?id=13428) backtrace from notepad in wine-1.0
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #22 from Alexandre Julliard julliard@winehq.org 2008-05-28 05:01:19 --- Thanks; most likely that's yet another XIM deadlock bug. You can try to comment out the XInitThreads call in x11drv_main.c to confirm this.
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #23 from Felix Saphir felix.saphir@presswatch.de 2008-05-28 05:57:26 --- Okay, works in 0.9.57, if the call is commented out: | if (!(env = getenv("XMODIFIERS")) || !*env) /* ... */ | /*if (!XInitThreads())*/ ERR( "XInitThreads failed, trouble ahead\n" );
Can use it as a switch between working/non-working :)
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #24 from Flob Flobber flobber@mailinator.com 2008-05-28 07:51:32 --- Yeah, with XMODIFIERS I can toggle between working and non-working.
Without XMODIFIERS: wine notepad -> hang With XMODIFIERS: XMODIFIERS=1 wine notepad -> no hang
But something's still fishy:
no window-manager: wine notepad -> no hang with twm: wine notepad -> hang
both times without XMODIFIERS.
http://bugs.winehq.org/show_bug.cgi?id=13420
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |FIXED
--- Comment #25 from Alexandre Julliard julliard@winehq.org 2008-05-28 14:52:20 --- I removed XInitThreads again.
http://bugs.winehq.org/show_bug.cgi?id=13420
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #26 from Alexandre Julliard julliard@winehq.org 2008-05-31 04:19:36 --- Closing bugs fixed in 1.0-rc3.
http://bugs.winehq.org/show_bug.cgi?id=13420
--- Comment #27 from Flob Flobber flobber@mailinator.com 2008-05-31 11:05:53 --- Fixed in .64.
Thanks, Flob
http://bugs.winehq.org/show_bug.cgi?id=13420
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |r000n@yandex.ru
--- Comment #28 from Alexandre Julliard julliard@winehq.org 2008-06-13 05:34:04 --- *** Bug 12758 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=13420
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |bugzilla@tut.by
--- Comment #29 from Alexandre Julliard julliard@winehq.org 2008-06-13 05:44:18 --- *** Bug 13143 has been marked as a duplicate of this bug. ***