Re: [PATCH] Re: mutt IMAP seg-fault in imap/browse.c
On Wednesday, 04 January 2006 at 19:26, Phil Pennock wrote:
> On 2006-01-04 at 09:30 -0800, Brendan Cully wrote:
> > On Wednesday, 04 January 2006 at 15:13, Phil Pennock wrote:
> > > With patch below, I can no longer replicated this segfault.
>
> English really is my first language, although you wouldn't know it from
> that.
I just thought you were speaking hacker :)
> > Thanks! I've just committed what may be a slightly more robust version
> > of this patch (zeros out list.name before every imap_cmd_step, in case
> > non-list responses are interleaved with list responses).
>
> But not initialising the stack-allocated item before use? Ah well ...
this is ok, since the list response parser zeros out the struct
first. NULLing list.name is enough to ensure that we don't examine the
structure unless the list response parser has already gotten to it.
> I confirm that I can no longer replicate the segfault behaviour with the
> alternative fix in current CVS code.