[Pidgin] #2094: Pidgin Causes Error, Causing entire machine to lock up (freeze) except for ctrl-alt-del

[Pidgin]

#2094: Pidgin Causes Error, Causing entire machine to lock up (freeze) except for
ctrl-alt-del
------------------------------+---------------------------------------------
  Reporter:  eliw             |       Owner:  datallah
      Type:  defect           |      Status:  new     
  Priority:  minor            |   Milestone:          
 Component:  winpidgin (gtk)  |     Version:  2.0.2   
Resolution:                   |    Keywords:  Synergy 
   Pending:  1                |  
------------------------------+---------------------------------------------
Changes (by datallah):

  * pending:  0 => 1

Comment:

 Replying to [comment:12 eliw]:
 > I understand that it doesn't appear to be directly related to Pidgin.
 >
 > However, my point (as a programmer), is just thus:  Gaim never had this
 problem.  I'm unsure therefore what changed between Gaim and Pidgin, that
 caused this problem to start being exercised.  But something did.

  Please feel free to experiment with any of the thousands of lines of code
 that changed; I don't think it is the best use of my limited time to look
 for a workaround for an issue (that appears to be a bug elsewhere) which
 only a couple people are likely to run into.  Keep in mind that there are
 plenty of bugs that I can actually do something about that I haven't had
 time to fully investigate.

 > Now, perhaps, if this is a new incompatibility between the latest GTK+
 (if versions were upgraded between Gaim and Pidgin) and Synergy, then yes,
 you are correct, nothing that you can do.

  I have no idea which version of gaim you're referring to, but if it was
 1.5.0 the GTK+ version has changed from 2.6.10 to 2.10.13.

 > However, I haven't had crash/lockup issues with any other GTK+
 applications that I run.  So I'm doubting that, though can be convinced
 otherwise.

  There are any number of different ways to use GTK+ and given that this is
 apparently an intermittent issue, I think it isn't at all correct to jump
 to the conclusion that it isn't a GTK+ issue; while it certainly is
 possible that Pidgin is doing something wrong, the crash trace simply
 doesn't indicate that.

 > On the other hand, if the way that GTK+ was used, has been changed, then
 perhaps, just perhaps, a workaround can be found, by reverting some of the
 code, and/or just not using certain bits, to stop this from happening.

  As a rule, this isn't the type of thing that would be worked around in
 Pidgin, but if you did track down something specific in GTK+, you could
 certainly report a bug there and hopefully get it resolved upstream.

 > I will also state that it seems rather odd to me, if this is a direct
 GTK+ problem only ... Since there is the strong coorelation between
 MSNchat ... MSNchat wouldn't be using GTK+ any differently than any other
 chat, would it?
 >
 > And if for some reason it does, then it would seem that the workaround
 might be quickly found, since the different use of GTK+ could be isolated.

  I think the MSN chat thing is a red herring; my guess is that it isn't
 related.

 > Thoughts?

  The backtrace doesn't indicate anything at all related to pidgin; it is
 happening entirely outside of Pidgin, crashing in the main loop due to a
 strcmp(NULL) somewhere  in glib based on a text selection event.

  If you run pidgin in gdb (available from [http://www.mingw.org]) and
 obtain a full backtrace (using the non-win32 gdb instructions
 [wiki:GetABacktrace here]), it might provide some clues.

  I'd be happy to give you any help that I can if you want to look into
 this.

-- 
Ticket URL: <http://developer.pidgin.im/ticket/2094#comment:13>
Pidgin <http://pidgin.im>
Pidgin