gEDA-dev: The infamous G_SLICE crash fixed (I hope) in CVS.

[Ales Hvezda]

Hi,

I spent some time (about 6 hours or so) this weekend tracking down the
G_SLICE crash that has been plaguing gEDA/gaf for a while now.  I think
I fixed it, but who knows.  All I can say is 1) it works for me, and 2)
valgrind isn't complaining anymore.

One of the bugs (memory corruption) has been in gEDA/gaf for
many many years and only now showed itself.  The other major
problem was the mixing of classic malloc/strdup/free and glib's
g_malloc/g_strdup/g_strdup_printf/g_free.  This is not allowed anymore.
Please use the g_* equivalent functions from now on.

The changes are in CVS right now (geda/gaf), so if anybody wants to test
them out with the new glib 2.10.x, please do so and let me know.  Be sure
to unset the G_SLICE environment variable if you set it before.  Thanks.

-Ales