[Pidgin] #442: Program received signal SIGSEGV on startup, pidgin-2.0.0beta7 on FreeBSD (compiled manually, not through Ports)

[Pidgin]

#442: Program received signal SIGSEGV on startup, pidgin-2.0.0beta7 on FreeBSD
(compiled manually, not through Ports)
-------------------------+--------------------------------------------------
  Reporter:  fongsaiyuk  |       Owner:     
      Type:  defect      |      Status:  new
  Priority:  minor       |   Milestone:     
 Component:  libpurple   |     Version:  2.0
Resolution:              |    Keywords:     
   Pending:  0           |  
-------------------------+--------------------------------------------------
Comment (by fongsaiyuk):

 Thanks for the correction on the install order. :) Not to tell a man his
 business, but would it be desirable to set the corresponding libpurple
 version as a dependency for pidgin in the Makefile, a la "pidgin-2.0.0
 depends on libpurple-2.0.0. . ." (as opposed to libpurple without a
 version number) and then during a portupgrade of pidgin, libpurple is
 upgraded before pidgin? I want to say earlier betas did this with gaim and
 libgaim. Just a thought, as there are probably other silly users like
 myself who would see that a new version of pidgin is out, then immediately
 attempt to "portupgrade pidgin" without giving much thought to libpurple.

 Alas, after upgrading to libpurple-2.0.0 and pidgin-2.0.0 with the above
 patches, I get a segfault when pidgin launches. As was the case when I
 compiled beta7 with gnutls disabled and nss enabled as per my 05/01/2007
 03:33:42 PM post, gdm now crashes before it can generate a backtrace:

 {{{

 (gdb) run
 Starting program: /usr/local/bin/pidgin
 warning: Unable to get location for thread creation breakpoint: generic
 error
 [New LWP 100092]
 [New Thread 0x80e7000 (LWP 100117)]
 [New Thread 0x80e7200 (LWP 100115)]
 gdb in realloc(): error: pointer to wrong page
 zsh: abort      gdb /usr/local/bin/pidgin
 }}}

 I tested those patches on a second machine running FreeBSD 6.2-RELEASE-p0
 on amd64, and pidgin-2.0.0 runs with no problem using nss. However, on
 ''this'' machine, also FreeBSD 6.2-RELEASE-p0, albeit on i386, I get that
 segfault at startup. nss is the current version in Ports.

 Out of curiosity, is there a reason why Mozilla NSS is preferred over
 GnuTLS? (Besides the fact that most desktop users already have nss
 installed).

-- 
Ticket URL: <http://developer.pidgin.im/ticket/442#comment:9>
Pidgin <http://pidgin.im>
Pidgin