-1
This code cannot run under mod_perl 2 because it has a 'use
Apache::Constants' right at the top of SizeLimit.pm.
Adam
On 10-09-28 11:52 PM, Fred Moyer wrote:
Please take a minute to download and test this release candidate for
Apache::SizeLimit. This is especially import for those of you
after actually reading the documentation i found out this isn't really
dual lifed, and i should be running Apache2::SizeLimit under mp2.
However, this still fails with:
\t(in cleanup) Can't locate object method "set_max_process_size" via
package "Apache2::SizeLimit" at /usr/local/prefork/conf/
On Wednesday, September 29, 2010 17:01:15 Adam Prime wrote:
> PerlModule Apache2::SizeLimit
>
>
> Apache2::SizeLimit->set_max_process_size(150_000);
> Apache2::SizeLimit->set_min_shared_size(10_000);
> Apache2::SizeLimit->set_max_unshared_size(120_000);
>
I suspect that should be
On 10-09-29 11:07 AM, Torsten Förtsch wrote:
I suspect that should be either
use Apache2::SizeLimit;
Apache2::SizeLimit->set_max_process_size(150_000);
Apache2::SizeLimit->set_min_shared_size(10_000);
Apache2::SizeLimit->set_max_unshared_size(120_000);
or
Pe
This turned out to be an @INC ordering problem. Apache will load now
with the default config, though it may be a good idea to include the
PerlLoadModule Apache2::SizeLimit and/or use Apache2::SizeLimit in the
synopsis.
I'm going to fiddle with this more later
Adam
--
Make sure that you nuke any previously installed versions of SizeLimit
when testing. The version bundled with mod_perl2 is installed in
i686-lib (or you own arch), and this pure perl version is installed in
site_perl.
I'm wondering how much trouble may be caused if someone installs A::SL
bundled
+1, 2.6 Linux, 5.8.8, 2.2.8
+1, Darwin, 5.12.2, 2.2.15
On Tue, Sep 28, 2010 at 8:52 PM, Fred Moyer wrote:
> Please take a minute to download and test this release candidate for
> Apache::SizeLimit. This is especially import for those of you users
> who actually *use* Apache::SizeLimit :) There
Here's my attempt at removing the interactive and custom config, and
instead gathering all configuration data from the current environment.
It should save us a *huge* amount of support headache, and fixup a
lot of false positive failing tests.
I realize this is an enormous paradigm shift, but I t
Note that this will require additional releases of SizeLimit and
Reload to deal with the API change. But it makes the bootstrapping
process a bit simpler also.
On Wed, Sep 29, 2010 at 6:39 PM, Fred Moyer wrote:
> Here's my attempt at removing the interactive and custom config, and
> instead gath
