[g-a-devel] at-spi segfaults

[Matthias Clasen <mclasen redhat com>]

At the request of Bill H. I have committed a "fix" for bug 337239
(immediate segfaults on x86_64), which made the segfaults go away
on x86_64. Unfortunately, as I have only now noticed, it makes the
same code segfault on i386...

At this point, I'm pretty much out of ideas and out of time to
dig further into this problem. My current hypothesis is that maybe
the struct copying and stack allocation of CORBA_any's thats going
on in the code is forgetting to set up/copy the mem-prefix that
CORBA_free needs to decide how to free things. Someone who actually
knows CORBA should look into this...

Matthias