Torsten Foertsch wrote:
On Fri 22 Feb 2008, Philippe M. Chiasson wrote:
-parms.path = path;
+parms.path = apr_pstrdup(p, path);
Yes, Philippe, you are right.
In current modperl there are 3 ways to get to modperl_config_insert():
- $r->add_config via modperl_config_insert_request
- $s
On Thu 21 Feb 2008, Fred Moyer wrote:
> patching file src/modules/perl/modperl_config.c
> Hunk #1 succeeded at 607 with fuzz 1 (offset -17 lines).
I am working with the threading branch. So there may be an offset. Also, some
time ago I had to split up a jumbo patch. Since the bits built up on top
On Fri 22 Feb 2008, Philippe M. Chiasson wrote:
> - parms.path = path;
> + parms.path = apr_pstrdup(p, path);
Yes, Philippe, you are right.
In current modperl there are 3 ways to get to modperl_config_insert():
- $r->add_config via modperl_config_insert_request
- $s->add_config via modperl
Torsten Foertsch wrote:
On Mon 18 Feb 2008, Torsten Foertsch wrote:
Here is a patch and a test case.
Forget the patch. It fixes the wrong place. The only other directive (that I
know of) that uses the location path ProxyPassReverse also assumes that the
pointer is valid for the whole request
> On Mon 18 Feb 2008, Torsten Foertsch wrote:
>> Here is a patch and a test case.
>
> Forget the patch. It fixes the wrong place. The only other directive (that
> I
> know of) that uses the location path ProxyPassReverse also assumes that
> the
> pointer is valid for the whole request.
>
> So the c
On Mon 18 Feb 2008, Torsten Foertsch wrote:
> Here is a patch and a test case.
Forget the patch. It fixes the wrong place. The only other directive (that I
know of) that uses the location path ProxyPassReverse also assumes that the
pointer is valid for the whole request.
So the correct place to
Hi,
modperl_config_dir_create begins with these lines:
void *modperl_config_dir_create(apr_pool_t *p, char *dir)
{
modperl_config_dir_t *dcfg = modperl_config_dir_new(p);
dcfg->location = dir;
While dcfg is created anew the dir pointer is simply stored. This means the
lifetime of dcfg