(reposted since I think it got lost in the thread...) After giving this some thought last night, I realized that it's probably better if we don't create a structure full of possible holes (we trust h.sid a little too much at this point). So here's a potentially better fix. It renders my previous patches unneccessary, since we never hit a ctx->hdr that is uninitialized. -Ken
Attachment:
patch-1.5.13.kb.imap_segfault_on_corruption.3
Description: Binary data