Stas Bekman wrote:
> [EMAIL PROTECTED] wrote:
>
>> dougm 01/10/08 16:51:21
>>
>> Modified:lib/Apache compat.pm
>> Log:
>> add Apache::exit compat wrapper
>
>
>
> Uh, Doug, why this and a few other recent commits go into
> Apache/compat.pm? Does that mean that Apache::exit is
On Tuesday 09 October 2001 12:30, Stas Bekman wrote:
> Is Apache->server or its variation planned, because I cannot find it in
> todo/. I realize that now we can have more than one server, so it
> probably should be different but at least how do I get to the main
> server's object from modperl cod
[EMAIL PROTECTED] wrote:
> dougm 01/10/08 10:57:45
>
> Added: xs/ModPerl/Util ModPerl__Util.h
> Log:
> new module for util functions
>
> Revision ChangesPath
> 1.1 modperl-2.0/xs/ModPerl/Util/ModPerl__Util.h
>
> Index: ModPerl__Util.h
> =
[EMAIL PROTECTED] wrote:
> dougm 01/10/08 17:13:00
>
> Modified:src/modules/perl mod_perl.c modperl_perl.c modperl_perl.h
> Log:
> *CORE::GLOBAL::exit = \&ModPerl::Util::exit
Is there a way to freeze *CORE::GLOBAL::exit, once you set it? Otherwise
some code can set it to a di
[EMAIL PROTECTED] wrote:
> dougm 01/10/08 16:51:21
>
> Modified:lib/Apache compat.pm
> Log:
> add Apache::exit compat wrapper
Uh, Doug, why this and a few other recent commits go into
Apache/compat.pm? Does that mean that Apache::exit is deprecated in 2.0?
The same for runn
Doug MacEachern wrote:
> On Tue, 9 Oct 2001, Stas Bekman wrote:
>
>
>>Which means that if Apache-Test is used, the response handlers will be
>>configured only if they are located under one of these four dirs. I mean
>>if I want to use this feature under Apache-Registry subproject's own t/
>
On Tue, 9 Oct 2001, Stas Bekman wrote:
> Which means that if Apache-Test is used, the response handlers will be
> configured only if they are located under one of these four dirs. I mean
> if I want to use this feature under Apache-Registry subproject's own t/
> dir, I must make sure to put t
On Tue, 9 Oct 2001, Stas Bekman wrote:
> With what we have now, I cannot do:
> Apache->server->dir_config('ModPerl::RegistryCooker::DEBUG'),
> or do I?
you do now :)
> Is Apache->server or its variation planned, because I cannot find it in
> todo/.
anything that existed in 1.x and is not imp
While Gerald is working on generalizing the WrapXS stuff, I think that
there are many areas in Apache::Test that have hardcoded mod_perl stuff
that could be generalized.
For example the t/response/Test* are hardcoded in
Apache::TestPerlConfig::configure_pm_tests(). I suggest either configure
Doug MacEachern wrote:
> On Tue, 9 Oct 2001, Stas Bekman wrote:
>
>
>> for my $subdir (qw(Response Protocol Hooks Filter)) {
>>
>
> ok, well what do you want to add? these are there to configure all
> Perl*Handlers.
Which means that if Apache-Test is used, the response handlers will be
On Tue, 9 Oct 2001, Stas Bekman wrote:
> all I wanted to say that somebody will run a script that mungles
> *CORE::GLOBAL::exit. Since *CORE::GLOBAL:: is relatively new, I've
> called it modern :)
well if somebody really wants to mess with it, thats up to them.
> but I understand, it doesn'
Doug MacEachern wrote:
> On Tue, 9 Oct 2001, Stas Bekman wrote:
>
>
>>For example the t/response/Test* are hardcoded in
>>Apache::TestPerlConfig::configure_pm_tests(). I suggest either configure
>>the dirs to look under in Makefile.PL, or give the directory t/response
>>a magic meaning, so A
Doug MacEachern wrote:
> On Tue, 9 Oct 2001, Stas Bekman wrote:
>
>
>>This hangs in Apache::Registry tests, MARK never gets incremented.
>>
>
> right, because it (*MARK++) only happened if PL_tainting (see
> SvTAINTED_off macro). try with the recent commit, should have been that
> way to beg
>>I'm thinking of some modern registry script as an example.
>>
>
> such as?
all I wanted to say that somebody will run a script that mungles
*CORE::GLOBAL::exit. Since *CORE::GLOBAL:: is relatively new, I've
called it modern :)
but I understand, it doesn't matter whether it's a script or a
On Tue, 9 Oct 2001, Stas Bekman wrote:
> For example the t/response/Test* are hardcoded in
> Apache::TestPerlConfig::configure_pm_tests(). I suggest either configure
> the dirs to look under in Makefile.PL, or give the directory t/response
> a magic meaning, so Apache::Test will always look at
Doug MacEachern wrote:
> great! i haven't reviewed yet (and will need some time to). i should
> have mentioned this thought earlier, but it would make the dev/transition
> much easier if the new kit has new names (i.e. other than
> Apache::{Registry,PerlRun}). the 1.x stuff currently works wit
On Tue, 9 Oct 2001, Stas Bekman wrote:
> Is there a way to freeze *CORE::GLOBAL::exit, once you set it?
no.
> Otherwise some code can set it to a different value and wreck havoc in
> the whole system.
then they get what they deserve.
> I'm thinking of some modern registry script as an examp
On Tue, 9 Oct 2001, Stas Bekman wrote:
> > Uh, Doug, why this and a few other recent commits go into
> > Apache/compat.pm? Does that mean that Apache::exit is deprecated in 2.0?
the modperl version of "exit()" has nothing todo with Apache, never had
any business in that namespace, as we all kn
Geoffrey Young wrote:
>>
>> if ($mtime > $Stat{$file}) {
>> delete $INC{$key};
>> # warn "Reloading $key\n";
>> if (my $symref = $UndefFields{$key}) {
>> #warn "undeffing fields\n";
>> no strict 'refs';
>>
On Tue, 9 Oct 2001, Stas Bekman wrote:
> Another suggestion would be to point to itseldf, i.e. Apache/compat.pm
> instead of '1'. I think that this is a safe assumption, and if it gets
> reloaded, the module itself will reassign correct values in %INC.
good idea, done.
On Tue, 9 Oct 2001, Stas Bekman wrote:
> So you suggest to just rename the leading package s/Apache/ModPerl/ but
> to keep the Registry, PerlRun, RegistryNG and RegistryBB names? They
> preserve exactly the same functionality.
that's what i had in mind.
> We have discussed this before a bit
>
> if ($mtime > $Stat{$file}) {
> delete $INC{$key};
> # warn "Reloading $key\n";
> if (my $symref = $UndefFields{$key}) {
> #warn "undeffing fields\n";
> no strict 'refs';
> undef %{$symre
Doug MacEachern wrote:
> On Sat, 6 Oct 2001, Stas Bekman wrote:
>
>
>>As I go with Apache::Registry I'm playing with stuff from 1.x, and see
>>some problems. Let me know whether these should be fixed==implemented
>>or fixed==compat, ignored, documented.
>>
>>Trying Apache::Reload:
>>
>>- PerlIn
On Tue, Oct 09, 2001 at 01:06:43PM +0800, Philippe M. Chiasson wrote:
> Here is $r->pnotes() implementes, mostly stolen from 1.x but it works and all
> tests pass fine for me.
But darn, it leaks memory, as the pnotes HV never gets freed. Please wait
for an updated patch later...
Oupss...
> /ho
With what we have now, I cannot do:
Apache->server->dir_config('ModPerl::RegistryCooker::DEBUG'),
or do I?
Is Apache->server or its variation planned, because I cannot find it in
todo/. I realize that now we can have more than one server, so it
probably should be different but at least how do I
On Tue, 9 Oct 2001, Stas Bekman wrote:
> This hangs in Apache::Registry tests, MARK never gets incremented.
right, because it (*MARK++) only happened if PL_tainting (see
SvTAINTED_off macro). try with the recent commit, should have been that
way to begin with.
--
Here is $r->pnotes() implementes, mostly stolen from 1.x but it works and all
tests pass fine for me.
/home/gozer/sources/mod_perl2/deps/perl/bin/perl build/cvsdiff
Index: t/response/TestAPI/request_rec.pm
===
RCS file: /home/anoncv
27 matches
Mail list logo
