You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Ben Reser <be...@reser.org> on 2013/01/03 21:38:08 UTC
Re: svn commit: r1427254 - /subversion/trunk/tools/server-side/svnauthz-validate.c
On Thu, Jan 3, 2013 at 12:17 PM, Daniel Shahaf <da...@elego.de> wrote:
> What's the benefit of this? Destroying the pool at process exit ---
> it's not to reclaim memory. Is it to ensure some pool cleanup callback
> fires?
It's mostly just a correctness thing. I noticed it when I noticed
that 1.7.x's svnauthz-validate crashes on error handling.
Right now it has no impact whatsoever since we're doing only read-only
ops on a repository. But if we weren't then yes the pool cleanups
might be an issue.
Re: svn commit: r1427254 - /subversion/trunk/tools/server-side/svnauthz-validate.c
Posted by Ben Reser <be...@reser.org>.
On Thu, Jan 3, 2013 at 12:40 PM, Branko Čibej <br...@wandisco.com> wrote:
> We have maintainer-mode magic that aborts if it detects an svn_error_t
> that was not properly cleaned up.
That's exactly what it was, we didn't call svn_error_clear() after
svn_handle_error2().
Re: svn commit: r1427254 - /subversion/trunk/tools/server-side/svnauthz-validate.c
Posted by Branko Čibej <br...@wandisco.com>.
On 03.01.2013 21:38, Ben Reser wrote:
> On Thu, Jan 3, 2013 at 12:17 PM, Daniel Shahaf <da...@elego.de> wrote:
>> What's the benefit of this? Destroying the pool at process exit ---
>> it's not to reclaim memory. Is it to ensure some pool cleanup callback
>> fires?
> It's mostly just a correctness thing. I noticed it when I noticed
> that 1.7.x's svnauthz-validate crashes on error handling.
We have maintainer-mode magic that aborts if it detects an svn_error_t
that was not properly cleaned up.
-- Brane
--
Branko Čibej
Director of Subversion | WANdisco | www.wandisco.com