[gutsy] GSlice: g_thread_init() Warning messages

saw the same error when upgrading feisty to gutsy. seems to affect all programs running on mono.

e.g.:

joachim@madkeunnen:~$ muine
***MEMORY-WARNING***: /usr/lib/muine/muine.exe[16864]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...

(muine:16864): Gtk-CRITICAL **: gtk_widget_grab_default: assertion `GTK_WIDGET_CAN_DEFAULT (widget)' failed
Error grabbing key 144, 0x8571800
Error grabbing key 153, 0x8571800
Error grabbing key 162, 0x8571800
Error grabbing key 162, 0x8571800
Error grabbing key 164, 0x8571800

possibly related to bug #116687

it is not a mono problem, but a glibc problem:
this error appears in other apps like firefox, beryl, rythmbox...

assigning to glib, since the errors are not mono-specific.

Nicolò, how is that a glib bug? The warning comes from it but it's to signal an application bug

Which signal?

I also got this with gvim and many other apps that use glib in one way or another...

it is impossible that suddenly almost all applications have this bug...
firefox, nautilus, banshee, rhythmbox, tomboy, pidgin, seahorse, glchess, gnome-terminal...

it's not suddent, glib only started displaying warning about those incorrect usages recently though

yes, with suddenly I intended that all the warnings came up in the
same moment (what I upgraded from feisty to gutsy 2-3 weeks after the
official release of feisty)

that's because glib used to display no warning about this error, now it does and you know that applications have to be updated

do you mean that all there application had this bug but the bug was
ignored by the developers?

the bug was not ignored, they most likely didn't know that was a wrong use

Ok, thank you!

ok, so we must add every package that has this problems into this bug!
Am I right?

no, wait for now, it might be easier to change some library

I suspect that this bug might be causing both vlc and gxine to crash. It certainly is an ominous error message.

Disregard my earlier comment about vlc and gxine -- the vlc crashes at least were due to something else.

-----------------------------------------------
Docboard Outliner - program loads but gives this error message:
-----------------------------------------------
***MEMORY-WARNING***: dockboard-outliner[5971]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...
susan@susan:/usr/bin$

--------------------------------------
Gramps - program does not load at all
--------------------------------------
***MEMORY-WARNING***: gramps.py[6038]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...
Segmentation fault (core dumped)

***MEMORY-WARNING***: thunderbird-bin[6160]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...
*** glibc detected *** /usr/lib/thunderbird/thunderbird-bin: free(): invalid next size (fast): 0x08cccf28 ***

Many programs seems to have this message...rhythmbox, gnome-dictionary (and many gnome-* I think) ... Can we mark all bugs with this message as duplicate of this one ?
Maybe similair bugs :
https://bugs.launchpad.net/ubuntu/+source/listen/+bug/116687
https://bugs.launchpad.net/ubuntu/+source/muine/+bug/117726
https://bugs.launchpad.net/ubuntu/+source/gnome-terminal/+bug/117962
https://bugs.launchpad.net/ubuntu/+source/gconf2/+bug/119196
https://bugs.launchpad.net/ubuntu/+source/gdesklets/+bug/118718
https://bugs.launchpad.net/ubuntu/+source/gnome-compiz-manager/+bug/118948

***MEMORY-WARNING***: desktop-effects[5748]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...

julien: no, we can't. the first two bugs are entirely unrelated to this issue. the others, I don't know about.

I can confirm this bug. See bug duplicate: https://bugs.launchpad.net/ubuntu/+source/gnome-compiz-manager/+bug/118948

can we upgrade this bug to a higher importance, I think this is a serious issue and a basic programme as thunderbird should work and all of the other above mentioned programmes.

how is that a serious issue? Those software had already this bugs previous cycle and are working mostly correctly, that's not because there is a warning now on the wrong usage that things are sudently breaking or something

I repeat my previous message:

***MEMORY-WARNING***: thunderbird-bin[6160]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...
*** glibc detected *** /usr/lib/thunderbird/thunderbird-bin: free(): invalid next size (fast): 0x08cccf28 ***

this is not just a warning, thunderbird is in effect crashing after 2 or 3 minutes so unusable and in my case not working correctly.

the crash looks like a thunderbird bug and there is no indication it's due to glib warning, could you rather open a bug on thunderbird where the corresponding maintainers will have a look?

Audacity will not load because of the above error.
***MEMORY-WARNING***: [27964]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...
*** glibc detected *** audacity: double free or corruption (out): 0x086eb860 ***
Backtrace attached.

I cannot believe that all of sudden all the above mentioned application do have a similar bug, susan has to file a bug at audacity, I at mozilla, Juan Pablo at desktop-effects, Julian at rhythmbox, gnome-dictionary etc, etc. I think there must be some underlining problem in glibc. Or??

another one:

***MEMORY-WARNING***: [19535]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...
*** glibc detected *** filezilla: double free or corruption (out): 0x085ae040 ***

file a bug at filezilla?

I support the happe's opinion. You must be kidding to report EVERY
similiar/same bug in every single application/package.

ptomes@ptomes-laptop:~$ gnome-compiz-preferences

***MEMORY-WARNING***: gtk-window-decorator[7432]: GSlice:
g_thread_init() must be called before all other GLib functions; memory
corruption due to late invocation of g_thread_init() has been
detected; this program is likely to crash, leak or unexpectedly abort
soon...
Chyba GConf: Chybný typ: Očekával jsem 'float', ale obdržel jsem 'int'
pro klíč /apps/compiz/plugins/decoration/allscreens/options/shadow_radius

ptomes@ptomes-laptop:~$

https://bugs.launchpad.net/ubuntu/+source/gnome-compiz-manager/+bug/118948

ptomes@ptomes-laptop:~$ desktop-effects
nvidia hardware not available

***MEMORY-WARNING***: desktop-effects[19130]: GSlice: g_thread_init()
must be called before all other GLib functions; memory corruption due
to late invocation of g_thread_init() has been detected; this program
is likely to crash, leak or unexpectedly abort soon...
gtk-window-decorator: Screen 0 on display ":0.0" already has a
decoration manager; try using the --replace option to replace the
current decoration manager.
GLX_EXT_texture_from_pixmap is available with direct rendering.
/usr/bin/compiz.real (core) - Fatal: No GLXFBConfig for default depth,
this isn't going to work.
/usr/bin/compiz.real (core) - Error: Failed to manage screen: 0
/usr/bin/compiz.real (core) - Fatal: No manageable screens found on display :0.0

***MEMORY-WARNING***: metacity[19177]: GSlice: g_thread_init() must be
called before all other GLib functions; memory corruption due to late
invocation of g_thread_init() has been detected; this program is
likely to crash, leak or unexpectedly abort soon...

***MEMORY-WARNING***: metacity[19190]: GSlice: g_thread_init() must be
called before all other GLib functions; memory corruption due to late
invocation of g_thread_init() has been detected; this program is
likely to crash, leak or unexpectedly abort soon...
ptomes@ptomes-laptop:~$

https://bugs.launchpad.net/ubuntu/+source/desktop-effects/+bug/120344

2007/6/14, hapee <email address hidden>:
> another one:
>
> ***MEMORY-WARNING***: [19535]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...
> *** glibc detected *** filezilla: double free or corruption (out): 0x085ae040 ***
>
> file a bug at filezilla?
>
> --
> [gutsy] GSlice: g_thread_init() Error messages
> https://bugs.launchpad.net/bugs/116870
> You received this bug notification because you are a direct subscriber
> of the bug.
>

no need to add random comment, if you are not comfortable getting a warning display when a program does something wrong maybe you should stick to a stable version of the distribution

Sebastian, I am very proud to be a member of a Ubuntu community and
one of thousands of testers and I am fully aware of alpha state of
Gutsy. I want to help to beGutsy as excellent as possible to be proud
of it at recommending to friends, so I report two similiar bugs to
compiz. The result of that bugs (as far as I am able to underestand)
is that compiz (and according to this bug comments many other
applications) isn't working. :( So for that reason I see reporting
this on every thoughable applications is pointless and waste of time.

Agree with Petr, this is not about warning but about not functioning application, they just don't start or crash after a few seconds, look at the examples above and if someone reports it it should be taken serious instead of an arrogant response. Can we take this higher up within the community?

You are commenting on the wrong entry then, this bug is about the glib warning displayed on the command line and there is no indication those crashes are due to it, there is lot of desktop applications displaying the warning and not crashing

I had the same error with either vlc and amule. I'm running kubuntu gutsy

petr, hapee: this bug is clearly not a crasher. if your app crashes, you _should_ file another bug, because the GSlice stuff will not be the issue. as sebastian said: there are a lot of packages that work correctly even with the bug. (see the entry in the gnome bugzilla for details)

to really help with the gutsy development, filing bugs correctly is the least you can do.

Joachim + Sebastian: OK Thank you for your time and explanations. I
will do my best.

Gramps (the genealogy program) and Audacity do not load at all. This may be due to this bug or something else, if I am reading the comments properly, and I should amend or file bugs for both of these.
Audacity has previously reported stability problems:
bug 106930
So I added my traceback report to that bug.
I filed a new bug for gramps.
bug 120569
Susan

I am getting this same bug when I try to run a Java OpenGL (JOGL/SWT) application (avoCADo CAD) that worked cleanly before I installed [gutsy]. Have there been any updates or decisions since the last post about 12 days ago? I realize that [gutsy] is still in the debug process and just want to do my part in reporting that the bug still exists for me.

thanks.

After all it seems that the "GSlice: g_thread_init()" bug was not due to libglib, but to wxwidget 2.6 and 2.8. So it affected most programs dependent from those libraries. With new versions of those libraries released between yesterday and today it seems that the issue is gonna be resolved. I still have problems with amule, but other programs (vlc, audacity, etc.) function well.

@ Susan:
the gramps bug seems to be completely different from this one, maybe due to aspell dictionary. Audacity, on the other way, should funtion now.

Many thanks to all the developers for the solution. Bye

Same warnings are in bug 120344

The bug is a mono bug, it should NOT depend on glib!!
It is already fixed in the upstream mono SVN and in debian unstable repository as in
http://bugzilla.ximian.com/show_bug.cgi?id=8186
http://packages.debian.org/changelogs/pool/main/m/mono/mono_1.2.4-4/changelog
I'm attaching a patch which applies against 1.2.4.

Wrong link for ximian bugzilla, sorry
http://bugzilla.ximian.com/show_bug.cgi?id=81862

?! This patch is already in the gutsy packages since weeks... also mono must depend on glib as it uses many glib functions to not reinvent the wheel.

Sorry, i just saw you committed the patch on 12/06, the bug report is dated 30/05, so i guess it is solved now, at least for mono...
I did not say mono must not depend on glib, i was saying this is not a glib bug as stated, but a mono bug.
I think this should be closed...

Well, it's arguable whether this is a glib or $random_application bug... see the upstream bugreport linked to this one here.

And this can't be closed because it still affects n+1 different applications ;)

I'm getting these too. I don't seem to see any pattern to them. It seems completely random because I'm getting them by just installing the updates.

I realize there is a bit of dispute over whether or not this is a bug in GLib or the applications that are using GLib, but regardless, it seems to have an effect on a handful of programs. Could this be a case of an older bug in GLib that was "fixed" but possibly now breaks existing apps that depended on that bug?

For the app I am running (JOGL/SWT in Java) the warning message is not just a nuisance, it is immediately followed by a stack trace and the end of my program's happy execution. :(

there is a workaround described in the bug. The glib2.0 upstream task state they want to make it handle the error better, that will likely happen before for gutsy, you are using an unstable distribution and such bugs happen there

I understand that Gutsy is not "stable" yet. :)
I am just trying to help by validating this bug's importance for
resolution.

Handling the error better is a good starting point, but ideally someone who
knows the inner workings of GLib or one of the libraries that calls to it
can isolate the bug and either update GLib or clearly state what
libraries/apps are doing wrong now (i.e., what needs to change to be
compatible with newer version of GLib).

On 7/5/07, Sebastien Bacher <email address hidden> wrote:
>
> there is a workaround described in the bug. The glib2.0 upstream task
> state they want to make it handle the error better, that will likely
> happen before for gutsy, you are using an unstable distribution and such
> bugs happen there
>
> --
> [gutsy] GSlice: g_thread_init() Warning messages
> https://bugs.launchpad.net/bugs/116870
> You received this bug notification because you are a direct subscriber
> of the bug.
>

> or clearly state what libraries/apps are doing wrong now

the warning is clear no, "g_thread_init() must be called before all other GLib functions"?

In the upstream bug linked to this one is stated that these errors occurr if g_thread_init() is not the first glib function called and you use gslice functions.

Indeed. thanks for the clarification.
I will begin digging through some 3rd party library code to see if I can
remedy my problems.

I guess I am most surprised that the same libraries worked (as far as I
could tell) with older versions of GLib. I am glad that the errors are now
being caught and reported.

On 7/5/07, Cosimo Cecchi <email address hidden> wrote:
>
> In the upstream bug linked to this one is stated that these errors
> occurr if g_thread_init() is not the first glib function called and you
> use gslice functions.
>
> --
> [gutsy] GSlice: g_thread_init() Warning messages
> https://bugs.launchpad.net/bugs/116870
> You received this bug notification because you are a direct subscriber
> of the bug.
>

I would just like to add that I recently compiled the latest stable release of glib, 2.13.2 at the time, under feisty and installed it. Suddenly I started getting this error everywhere. Therefore, it IS glib related. Whether this has always been the case, but the error was not previously printed (as suggested above), or whether this is a new (and significant) but in glib, I don't know...

Pascal.

Yea, this command happens with lots of commands on Gutsy.

Just to add to the list :gftp
gFTP Warning: Skipping line 229 in config file: entropy_source
gFTP Warning: Skipping line 232 in config file: entropy_len
gFTP Warning: Skipping line 235 in config file: verify_ssl_peer

***MEMORY-WARNING***: [11262]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...

gksudo gedit test

***MEMORY-WARNING***: gksudo[11029]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...

abiword

***MEMORY-WARNING***: abiword[11033]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...

beryl-manager

***MEMORY-WARNING***: beryl-manager[11040]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...

gxine

***MEMORY-WARNING***: [11145]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...
gtkvideo: failed to get a proxy for gnome-screensaver
lirc: cannot initialise - disabling remote control
lirc: maybe lircd isn't running or you can't connect to the socket?

gftp
gFTP Warning: Skipping line 229 in config file: entropy_source
gFTP Warning: Skipping line 232 in config file: entropy_len
gFTP Warning: Skipping line 235 in config file: verify_ssl_peer

***MEMORY-WARNING***: [11262]: GSlice: g_thread_init() must be called before all other GLib functions; memory corruption due to late invocation of g_thread_init() has been detected; this program is likely to crash, leak or unexpectedly abort soon...

As a side note, I have seen the comments in this thread , like Sebastien Bacher said on 2007-06-14: Re: [gutsy] GSlice: g_thread_init() Error messages (permalink)

"no need to add random comment, if you are not comfortable getting a warning display when a program does something wrong maybe you should stick to a stable version of the distribution"

Sebastien : I do not mind error messages and understand this is in the nature of running Gutsy. Developers should be thankful that the Ubuntu community takes the time to test and report back. This kind of feedback should be encouraged ... If you think these comments are "random", where would you have them posted ? And do you not think it would be helpful if the developers would give feedback on this thread as well ? I think that would give structure to bug reports.

You can add comments if they have new informations, complains about software not working correctly or printing warning on gutsy are of no use. If an application is displaying an error you can open a bug against it or a task on this one

Erm, just to add my $0,02, this warning has been displayed in all the glib development releases for a while and is hidden in stable release, hence the fact that one doesn't ever see it usually.
And for the record, it affects all pyGTK apps, and it's definitely not a problem most often, but a "stupidity" from glib developers (that's what I have been told by one of the pyGTK core developers).

Hi, having browsed through the Gutsy milestones the first entry entry that I found is for Ubuntu 7.10 Beta scheduled for release on the 27th of September. Just over two months away.

I'm curious as to why this is the case as,

a) an upstream fix has been released,
b) the bug has been rated as high in importance &
c) a number of applications have been affected.

Personally the three most commonly affected applications I use are Firefox, (hangs and crashes on occasion), Thunderbird, (virtually unusable) and Gnome Terminal, (no apparent affect).

Cheers.

Well it will certainly be fixed by the developers long before Gutsy is released as stable, currently it's an Alpha release and should not be used for anything other than testing and bug-hunting.

I'm have no doubt that it will be fixed prior to the final release thankyou.

I would have thought that obvious as have read through the various milestones schedule up to the first beta.

Not to mention that I'm well aware of the unstable nature of alpha releases. Gosh, such 'warnings' pop up all the time in the forums, I'm a little surprised that someone should feel the need to do so in launchpad.

In fact, for your information I run parallel installs one stable and one for testing.

I was merely curious as to the reason that given this bug / is marked as serious / effects a range of applications and an upstream fix is currently available, that a downstream fix is not scheduled for another two months.

I would consider an informative response as part of furthering my understanding of the development process.

As a result I look forward to a constructive response.

glib2.0 (2.13.7-1ubuntu1) gutsy; urgency=low

  * Sync with Debian and new upstream version:
    - The memory corruption warning from the slice allocator that
      occurred when threads were initialized after the slice allocator
      has been removed, as the slice allocator now works fine
      in this scenario (LP: #116870)
    - New functions g_once_init_enter() and g_once_init_leave() make
      it easier to write threadsafe one-time initialization functions
    - Bugs fixed:
      454473 Simple XML Subset Parser terminates on invalid XML
      445813 g_module_open error, add file name
      453796 errno gets clobbered by g_filename_display_name
      341988 don't use "-c" with msgfmt in Makefile.in.in
      447048 Please produce slightly more output during long tests
      454785 GModule documentation lists same block of code twice.
      454786 GModule documentation lists same paragraph twice.
      383155 small docs quirks in gobject/closure API documentation
      65041 _get_type() functions aren't thread safe
  * debian/control.in:
    - the package is maintained by the Ubuntu Desktop Team
  * debian/patches/series:
    - use 01_gettext-desktopfiles.patch

 -- Sebastien Bacher <email address hidden> Mon, 23 Jul 2007 10:03:37 +0200