python25 core dumps

David J Brooks freysman at
Thu Nov 15 07:24:40 PST 2007

On Thursday 15 November 2007 03:07:03 am Heiko Wundram (Beenic) wrote:

> This seems like a problem in libaspell; maybe you should simply try to
> reinstall the aspell port. See below for more info.

I rebuilt aspell, but gramps still core dumps. The backtrace shows the same as 
previously. :/

> > For comparison, this is what a crash from eric4 (built with PyQt4) looks
> > like:

<snip backtrace>

> This seems like a problem in qt4 (I don't think the problem is in PyQt),
> simply try reinstalling that, too (completely; qt4 is split into several
> ports and "pkg_info | grep qt4" is your friend here).
> Generally, from what I interpret into the second backtrace, you upgraded
> from some 6 release to 7.0-BETA2, which (amongst other things) means that
> the C++ libraries have changed (because of a newer compiler, gcc 3.3 vs.
> 4.2). The compiler has also had changes introduced to the C++ type info
> descriptor layout (which I should think causes the segmentation fault in
> typeinfo name in the second backtrace), so that if you have a program
> that's linked against different versions of libstdc++ (PyQt is linked
> against that, just as qt4 is), you'll see behaviour like this.
> To check whether my hypothesis is correct, simply do an ldd on both a PyQt
> library, and a qt4 shared library (locations of both of which you can
> extract from the backtrace). If the version of libstdc++ is different, you
> didn't follow the upgrading procedure which explicitly states to recompile
> _all_ ports for the new system.

These appear to be the same, but its possible my blurry eyes may be 
overlooking something obvious.

/usr/local/lib/ => /usr/local/lib/ (0x28185000) => /usr/local/lib/ (0x281aa000) => /usr/local/lib/ (0x281b2000) => /usr/local/lib/ (0x28a1a000) => /lib/ (0x281c9000) => /usr/local/lib/ (0x281db000) => /usr/local/lib/ (0x28b89000) => /usr/local/lib/ (0x28c27000) => /usr/local/lib/ (0x281e0000) => /usr/local/lib/ (0x281e8000) => /usr/local/lib/ (0x281f0000) => /usr/local/lib/ (0x281f7000) => /usr/local/lib/ (0x28d1c000) => /usr/local/lib/ (0x281fc000) => /usr/local/lib/ (0x28d25000) => /usr/local/lib/ (0x28d93000) => /usr/local/lib/ (0x28dbd000) => /usr/local/lib/ (0x28dcb000) => /usr/lib/ (0x28eb8000) => /lib/ (0x28fa2000) => /lib/ (0x28fb7000) => /lib/ (0x28fc2000) => /lib/ (0x28089000) => /usr/local/lib/ (0x28fd4000) => /usr/local/lib/ (0x28fdd000) => /usr/local/lib/ (0x29003000) => /usr/local/lib/ (0x29023000) => /usr/local/lib/ (0x29026000) => /usr/lib/ (0x2902b000)

/usr/local/lib/python2.5/site-packages/PyQt4/ => /usr/local/lib/ (0x28b09000) => /usr/local/lib/ (0x28185000) => /usr/local/lib/ (0x281aa000) => /usr/local/lib/ (0x281b2000) => /usr/local/lib/ (0x29223000) => /lib/ (0x281c9000) => /usr/local/lib/ (0x281db000) => /usr/local/lib/ (0x29392000) => /usr/local/lib/ (0x29430000) => /usr/local/lib/ (0x281e0000) => /usr/local/lib/ (0x281e8000) => /usr/local/lib/ (0x281f0000) => /usr/local/lib/ (0x281f7000) => /usr/local/lib/ (0x29525000) => /usr/local/lib/ (0x281fc000) => /usr/local/lib/ (0x2952e000) => /usr/local/lib/ (0x2959c000) => /usr/local/lib/ (0x295c6000) => /usr/local/lib/ (0x295d4000) => /lib/ (0x296c1000) => /usr/lib/ (0x296d6000) => /lib/ (0x297c0000) => /lib/ (0x297cb000) => /lib/ (0x28089000) => /usr/local/lib/ (0x297dd000) => /usr/local/lib/ (0x297e6000) => /usr/local/lib/ (0x2980c000) => /usr/local/lib/ (0x2982c000) => /usr/local/lib/ (0x2982f000) => /usr/lib/ (0x29834000)

> There are also some other changes, especially in the threading libraries,
> which can cause errors like these on non-C++ applications, although I
> wouldn't know whether aspell is affected by this, but possibly gtk2 is.
> Anyway, posting some more info on the upgrade process to 7.0 you did would
> help, I guess.

I initially tried to upgrade all ports using 'portupgrade -fa', which ran for 
days and left me with 11 ports failed and about 300 skipped as a result. I 
managed to upgrade the failed ones with 'make install' then 'portupgrade -fa' 
again, rather than trying to upgrade the skipped 300 one at a time. When this 
ran into the same series of failures as before I realized the error and 
simply removed all ports 'pkg_delete -f \*' and reinstalled them 
using 'pkg_add -r'. I am currently in the process of 'portupgrade -aR' 
following a recent portsnap which shows a small handful of packages needing 

This message is LOADING...

More information about the freebsd-questions mailing list