| Summary: | XMMS segfaults upon exit after failed assertion in gtkwidget.c (xmms-eq plugin) | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Sybren Stüvel <sybren> |
| Component: | Current packages | Assignee: | Luis Medinas (RETIRED) <metalgod> |
| Status: | RESOLVED FIXED | ||
| Severity: | minor | CC: | as.gentoo, stian, tcort |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | x86 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: | eq-xmms-0.6-exit_fix.patch | ||
|
Description
Sybren Stüvel
2004-12-18 06:33:29 UTC
What plugins do you have in xmms. Might be a plugin that makes that error as-well. (I use the same version without problemes). What output do you use? etc. Input plugins, all enabled: - libcdread.so - libcdaudio.so - libxmmsmad.so - libmikmod.so - libmp4.so - libmpg123.so - libvorbis.so - libxmms-flac.so - libsmpeg_xmms.so - libmid.so - libtonegen.so - libwav.so Output plugins: - libALSA.so (in use) - libcrossfade.so - libdisk_writer.so - libesdout.so - libOSS.so Effect plugins, none enabled: - libecho.so - libstereo.so - libcompress.so - libnormvol.so - libvra.so - libvoice.so General plugins, none enlabled: - libir.so - libitouch.so - libjoy.so - libsong_change.so Visualization plugins, none enabled: - libbscope.so - libblursk.so - libdflowers.so - libdscope.so - libdspectogram.so - gdancer.so - libinfiinte.so - libiris.so - libjess.so - libnebulus.so - libogl_spectrum.so - libsanalyzer.so - libgoom.so I use ALSA output, but the same error occurs when using OSS output. uninstall your plugins until youu find the one that causes this please. no response for 3 weeks... closing. Ok, removing all possible plugins fixed it. Thanks. I'd rather you tell me which plugin is causing this problem. Please reopen this bug and update the summary! The plugin wich caused this problem is eq-xmms. ---START---output before unemerge----------------------- magic xmms # xmms ** CRITICAL **: file configfile.c: line 266 (xmms_cfg_write_string): assertion `value != NULL' failed. ** CRITICAL **: file configfile.c: line 266 (xmms_cfg_write_string): assertion `value != NULL' failed. ** CRITICAL **: file configfile.c: line 266 (xmms_cfg_write_string): assertion `value != NULL' failed. Gtk-WARNING **: invalid cast from (NULL) pointer to `GtkWidget' Gtk-CRITICAL **: file gtkwidget.c: line 1387 (gtk_widget_destroy): assertion `widget != NULL' failed. Gtk-WARNING **: invalid cast from (NULL) pointer to `GtkWidget' Gtk-CRITICAL **: file gtkwidget.c: line 1387 (gtk_widget_destroy): assertion `widget != NULL' failed. GLib-CRITICAL **: file gmain.c: line 500 (g_source_remove): assertion `tag > 0' failed. magic xmms # emerge unmerge xmms-eq -vDp ---STOP------------------------------------------------- LOL, I just noticed the bug was closed exactly 1 year ago. xD And please alter Hardware to "AMD64". ok reopening the bug and researching about it reassign bug to me since Jeremy is no longer xmms maintainer. thx =) buggy plugin try to write a patch to fix this issue :). I grabbed eq-xmms from CVS today and the bug is still present. There is an open upstream bug report[1] from 11-Nov-2004. I changed the bug report summary from xmms-eq to eq-xmms as the name of the package is eq-xmms[2] not xmms-eq, xmms-eq doesn't exist in portage. Now people who click on the "Bugs" link for this package on packages.gentoo.org will be able to find this bug :) [1] http://sourceforge.net/tracker/index.php?func=detail&atid=496208&aid=1064985&group_id=61119 [2] http://packages.gentoo.org/search/?sstring=eq-xmms "I changed the bug report summary from xmms-eq to eq-xmms" *cough* you did not ;) btw: Isn't this a good occasion to alter the packages name from eq-xmms to xmms-eq? (In reply to comment #15) > "I changed the bug report summary from xmms-eq to eq-xmms" > *cough* you did not ;) bugs.gentoo.org outputs an HTML text field for the summary which makes it appear as though it is editable by me, but it doesn't apply my changes or give any warnings/errors after I submit it. I know that kind of bug :-( Only gentoo ppl (andmins?) and the bugs reporter can alter the summary. That's kind of logic, anyway you're right ... there should be _no_ input field for data you can not alter. Btw, my comment was no criticism, just a note. :) Created attachment 77051 [details, diff] eq-xmms-0.6-exit_fix.patch The following 3 warnings happened because the plugin was attempting to write a configuration file when the user hadn't configured the plugin. Fixed to use the default values if the user hasn't configured the plugin. ** CRITICAL **: file configfile.c: line 266 (xmms_cfg_write_string): assertion `value != NULL' failed. ** CRITICAL **: file configfile.c: line 266 (xmms_cfg_write_string): assertion `value != NULL' failed. ** CRITICAL **: file configfile.c: line 266 (xmms_cfg_write_string): assertion `value != NULL' failed. The following 4 warnings happened because the plugin's cleanup function always gets called, even if the init function doesn't get called. The init function doesn't get called when the plugin is disabled. The cleanup function was trying to remove things that hadn't been initialized. Fixed to check if widget == NULL before trying to remove the widget. Gtk-WARNING **: invalid cast from (NULL) pointer to `GtkWidget' Gtk-CRITICAL **: file gtkwidget.c: line 1387 (gtk_widget_destroy): assertion `widget != NULL' failed. Gtk-WARNING **: invalid cast from (NULL) pointer to `GtkWidget' Gtk-CRITICAL **: file gtkwidget.c: line 1387 (gtk_widget_destroy): assertion `widget != NULL' failed. The cleanup function was trying to remove a timeout that hadn't been added (because init wasn't called for the same reason as above). Fixed to check that the timeout has a tag > 0 before attempting to remove the timeout. GLib-CRITICAL **: file gmain.c: line 500 (g_source_remove): assertion `tag > 0' failed. I tested the patch on Alpha and it works as expected. While testing other functionality I noticed that if you try random things like configuring the plugin before enabling it, you will get a lot of errors. Just clicking the Apply button causes some console messages. When I contact the upstream dev(s) about the fix for this bug I'll ask if/when there will be a future release[1] and if/when the plugin will come out of beta[2]. [1] last release May 15, 2004 http://sourceforge.net/project/showfiles.php?group_id=61119 [2] Project page says: "Development Status : 4 - Beta" http://sourceforge.net/projects/equ patch commited Thanks Thomas. |
