> > Sigh. Wonder if this is related to what FirstBill saw earlier today with > > the proxy? Seeing as nothing has changed in the buckets code any time > > recently... OHHHHHHHH WAIT, yes it has. Brian apparently committed a > > change to allocate the apr_bucket_brigade itself out of the brigade > > allocator. > > Revert Brian's patch to apr_brigade.c (attached) and the segfaults go away. > Interesting... I suspect this patch tickled a bug elsewhere in the code. > > Index: apr_brigade.c
Humm, the apr_brigade_destroy calls in core_output_filter look suspect. Freeing storage pointed to by almost every other higher level routine in the call stack can't be a good thing. We get away with it when the brigade is allocated out of a pool but that still does not make it right. Bill
