[reportlab-users] new pyRXP

Robin Becker reportlab-users@reportlab.com
Wed, 9 Apr 2003 08:25:17 +0100


In article <6CDB856C-6A39-11D7-9FD0-000393B63DDC@shangri-
la.dropbear.id.au>, Stuart Bishop <zen@shangri-la.dropbear.id.au> writes
.....
>There is still a double free around, and I can trigger it with my
>test suite. It seems to be in the parser cleanup routines. I'll have
>a closer look at this when I get a chance, but anyone more familiar
>with malloc debugging and the time is welcome to do it for me :-)
>I've attached the test cases that can trigger the issue in case anyone
>is feeling enthusiastic. This particular problem should only bite people
>who are parsing XML from untrusted sources since it takes particularly
>broken XML to trigger the malloc errors as far as I can tell.
.... if I get some more time I'll have a go at this, especially since we
now have a later rxp source. Thanks for the testing, are you now using
tests based on the later comment/processing instruction handling? I saw
also that you were attempting to get at the prolog in some way, is that
still required? It seems quite deeply buried in rxp.
-- 
Robin Becker