httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Yann Ylavic <ylavic....@gmail.com>
Subject Re: svn commit: r1707087 - /httpd/httpd/trunk/modules/debugging/mod_bucketeer.c
Date Tue, 25 Apr 2017 23:02:29 GMT
Hi Jacob,

On Tue, Apr 25, 2017 at 11:58 PM, Jacob Champion <champion.p@gmail.com> wrote:
>
> Unfortunately the patch just moves the crash to another location. We can't
> call APR_BRIGADE_EMPTY() on a brigade that's pointing at junk. I think a
> bucket that cleans up the brigade it's a part of is just not a good idea. :)
> So moving to c->pool is an option for a quick fix, I suppose...
>
> ...but I'm more inclined to look at the whole EOR bucket situation --
> specifically eor_bucket_destroy() and its helpers. Why is the EOR bucket
> responsible for freeing the request's pool in the first place? It doesn't
> own the request. Surely we should be freeing the pool in the same code
> context in which we've allocated the pool?
>
> One of the comments in eor_bucket.c is "eor_bucket_destroy might be called
> at a point of time when the request pool had been already destroyed", which
> makes me incredibly suspicious of the whole thing. Finalizers are not a
> great place to put business logic.

Let me remind this a bit, it's been a long time :)
Will have a look at it tomorrow hopefully..


Regards,
Yann.

Mime
View raw message