14 Mar
2012
14 Mar
'12
11:10 p.m.
Balazs Scheidler <bazsi@balabit.hu> writes:
Ops, I haven't noticed this patch, however I think my solution (just posted in the other thread) is somewhat better.
You may have some data allocated by the parser, which wouldn't be freed in this case.
Please have a look at my patch and comment if that's ok.
Mmm, it's certainly nicer, but it calls the vp transform set finisher even on the error path, which might not be the best idea (that can also allocate stuff). On the other hand, that stuff might be freed elsewhere, so it's probably not a big deal. I'll take it to a test drive during the long weekend, but nevertheless, your approach does look a lot better. -- |8]