Well, hmmm...
I tried 2 tests with php disabled. A 1mb commit succeeded, and a 2.5mb
commit failed. So either it improved things but didn't fully correct
the issue, or my conclusions about always failing over 100k was wrong.
Rob
Garrett Rooney wrote:
> On 9/15/06, Rob Brandt <bronto@csd-bes.net> wrote:
>> Yes, as a matter of fact I am.
>
> Could you possibly turn it off and see if the problem goes away?
>
> We've had a number of reports of weird errors like this that all
> basically get tracked back to PHP changing some sort of global state
> (like the umask, or signal handlers, or other things like that) and
> then not setting it back, or if you're using a multithreaded MPM even
> setting it then setting it back, but just at the wrong time... It's
> quite possible that PHP is causing your issue.
>
> -garrett
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Fri Sep 15 19:23:28 2006