> Fred Moyer wrote:
>> New diff file [3] based on comments by Geoffrey Young [1] and Philippe
>> M.
>> Chiasson [2]
>
> great work Fred!
>
Thanks Geoff, and thanks to the mod_perl dev team for this great software.
Much credit is due there, especially for mp2 which is great fun to
work with.
Currently, soak_end_container resides under Apache::Command, while expecting an
Apache::CmdParam as first argument. It was apparently misplaced to start with.
I suggest either moving it to Apache::CmdParams or dropping it alltogether.
--
-
For now, things in Apache::Directive, Apache::Command and Apache::Module having to
do with apache's configuration tree and modules have been made mostly read-only.
After thinking about this a bit, I believe they might be changed to writable before
child-init, read-only otherwise. I am not sure why,
Philippe M. Chiasson wrote:
This is an intended wrapper around ap_walk_config, but we already have a simpler
and friendlier Apache->server->add_config(). On top of it, to call walk_config,
you need a valid ConfVector object, something we don't expose at all right now.
Basically, this function is u
Philippe M. Chiasson wrote:
Right now, we are exposing the data member of ap_directive_t and I can't think
of a good reason why. It's a void pointer and it's private to the module implementing
the directive.
I'll remove it unless somebody screams.
Removed.
--
--
This is an intended wrapper around ap_walk_config, but we already have a simpler
and friendlier Apache->server->add_config(). On top of it, to call walk_config,
you need a valid ConfVector object, something we don't expose at all right now.
Basically, this function is unusable and has a good altern
Right now, we are exposing the data member of ap_directive_t and I can't think
of a good reason why. It's a void pointer and it's private to the module implementing
the directive.
I'll remove it unless somebody screams.
--
As stas pointed out in todo/release, since glue_pods is kinda broken right now,
why not simply use ExtUtils::MakeMaker to manify and install all pods in docs/api ?
--
Philippe M. Chiasson m/gozer\@(apache|cpan|ectoplas
Fred Moyer wrote:
> New diff file [3] based on comments by Geoffrey Young [1] and Philippe M.
> Chiasson [2]
great work Fred!
--Geoff
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTE
New diff file [3] based on comments by Geoffrey Young [1] and Philippe M.
Chiasson [2]
[1]
http://marc.theaimsgroup.com/?l=apache-modperl-dev&m=109215191529527&w=2
[2]
http://marc.theaimsgroup.com/?l=apache-modperl-dev&m=109215411230572&w=2
[3]
--- apache_structures.map 8 Aug 2004 18:04:55
it's a duplication of Apache::Module->name
$mod->name and $mod->find_module_name are identical
I'd like to remove the longer and stranger sounding find_module_name
method for clarity.
--
Philippe M. Chiasson m/gozer\@(
We are currently exposing dynamic_load_handle (i.e. the dlopen handle) of a module
to the Perl API.
I don't think there is any usefullness to this and I'd like to see it removed...
--
Philippe M. Chiasson m/gozer\@(apa
Stas Bekman wrote:
> +1 to remove all the code that isn't helping to test
> $r->note_digest_auth_failure.
is this ok?
--Geoff
Index: t/hooks/authen_digest.t
===
RCS file: /home/cvs/modperl-2.0/t/hooks/authen_digest.t,v
retrieving r
Fred Moyer wrote:
Per Stas Bekman's request I have reviewed the accessor settings in
modperl-2.0/xs/maps/apache_structures.map against the doc pages for
entries that should be read/write and generated a diff [1] against the
current cvs.
Excellent, thanks a lot!
I have found methods in the man page
> - no_local_copy
> +< no_local_copy
I don't think I'd do that, as no_local_copy has the potential to be useful
due to the effects it has on some underlying httpd functions.
> - local_addr
> - remote_addr
> - remote_ip
> - remote_host
> +< local_addr
> +< remote_addr
> +< remote_ip
Per Stas Bekman's request I have reviewed the accessor settings in
modperl-2.0/xs/maps/apache_structures.map against the doc pages for
entries that should be read/write and generated a diff [1] against the
current cvs.
I have found methods in the man pages which are not in structures.map, and
some
Stas Bekman wrote:
> Some of the autoconfig magic starts to create more overhead than it used
> to safe. e.g. all this magic of t/hooks/TestHooks is no longer useful,
> as we add more tests and the test names aren't mapping 1:1 to the phase
> name.
> I think we should start stripping those spec
17 matches
Mail list logo
