Re: 1.27-tobe + 1.99_02-tobe

2002-05-23 Thread Doug MacEachern
On Thu, 23 May 2002, Randy Kobes wrote: > That was fast :) Unfortunately, and maybe I'm not > waiting long enough, but now I get _modperl_perl_sv_setref_uv > being undefined ... woops i did it again. needed to update the ModPerl::FunctionTable so it is included in the generated modperl.def

Re: 1.27-tobe + 1.99_02-tobe

2002-05-23 Thread Randy Kobes
On Thu, 23 May 2002, Doug MacEachern wrote: > On Thu, 23 May 2002, Randy Kobes wrote: > > > > with apache-2.0.37-dev (of today), I get an error in APR::OS when > > creating APR/OS/OS.lib - _sv_setref_uv is unresolved. > > whoops, referenced a function that doesn't exist in 5.6.x > this should be

Re: 1.27-tobe + 1.99_02-tobe

2002-05-23 Thread Doug MacEachern
On Thu, 23 May 2002, Randy Kobes wrote: > > http://perl.apache.org/~dougm/mod_perl-1.26_01-dev.tar.gz > > with apache-1.3.24, all tests pass ... excellent. > > http://perl.apache.org/~dougm/mod_perl-1.99_02-dev.tar.gz > > with apache-2.0.37-dev (of today), I get an error in APR::OS when > c

Re: 1.27-tobe + 1.99_02-tobe

2002-05-23 Thread Randy Kobes
On Wed, 22 May 2002, Doug MacEachern wrote: > both are pretty close, i've tested both on linux, solaris and hpux and > 1.99_02 on win32 only. with various perl flavors. if anybody has a > chance to run some make tests with cvs or the snapshots below, that'd be > appreciated as always... On Win

Re: [PATCH] HTML::Mason::ApacheHandler vs. mod_perl2

2002-05-23 Thread Autrijus Tang
On Thu, May 23, 2002 at 12:10:06PM -0500, Dave Rolsky wrote: > On Fri, 24 May 2002, Autrijus Tang wrote: > > Greetings. Below is what I've done to port Mason(CVS version)'s > > ApacheHandler to Apache2; it deals mostly with the args-parsing > > incompatibilities and changed method names. > Actuall

Re: [PATCH] CGI.pm vs. mod_perl2's POST.

2002-05-23 Thread Doug MacEachern
On Fri, 24 May 2002, Autrijus Tang wrote: > It's bleadperl-16610. I'll try to hunt down the bug -- any quick pointers? the bug i was thinking of has been fixed in bleedperl (changes 11536 + 11553), and there's a workaround in modperl2 for 5.6.x that bug caused STDIN to loose its tied-ness, so

Re: [PATCH] CGI.pm vs. mod_perl2's POST.

2002-05-23 Thread Autrijus Tang
On Thu, May 23, 2002 at 09:25:30AM -0700, Doug MacEachern wrote: > On Fri, 24 May 2002, Autrijus Tang wrote: > > As the new Apache->request provide a 'content' method already, I patched > > it so it tests for the method, and prefers it if it is available. > the 'content' method is deprecated in 2.

Re: [PATCH] HTML::Mason::ApacheHandler vs. mod_perl2

2002-05-23 Thread Dave Rolsky
On Fri, 24 May 2002, Autrijus Tang wrote: > Greetings. Below is what I've done to port Mason(CVS version)'s > ApacheHandler to Apache2; it deals mostly with the args-parsing > incompatibilities and changed method names. Actually, I've already done this in the CVS version. Which version were you

Re: [PATCH] CGI.pm vs. mod_perl2's POST.

2002-05-23 Thread Doug MacEachern
On Fri, 24 May 2002, Autrijus Tang wrote: > Hello there. I've been using CGI.pm with the CVS Apache2 module, and > found out that read_from_client doesn't quite work -- it sometimes > blocks the socket, but most of the time just returns an empty list. > > As the new Apache->request provide a 'co

[PATCH] HTML::Mason::ApacheHandler vs. mod_perl2

2002-05-23 Thread Autrijus Tang
Greetings. Below is what I've done to port Mason(CVS version)'s ApacheHandler to Apache2; it deals mostly with the args-parsing incompatibilities and changed method names. I think some of these chunks will need some work to maintain mod_perl 1.x compatibility, but as I don't have version 1 right

[PATCH] CGI.pm vs. mod_perl2's POST.

2002-05-23 Thread Autrijus Tang
Hello there. I've been using CGI.pm with the CVS Apache2 module, and found out that read_from_client doesn't quite work -- it sometimes blocks the socket, but most of the time just returns an empty list. As the new Apache->request provide a 'content' method already, I patched it so it tests for t

ModPerl::Registry root

2002-05-23 Thread Doug MacEachern
haven't looked to much, but it seems ModPerl::Registry does not compile everything under the same root package as 1.x does with Apache::ROOT:: but instead does something more obscure that also requires no strict 'refs' all over the place. 1.x was much cleaner the way it maintains the cache.

mod_perl benchmarking (was Re: documenting SetHandler perl-script|modperl)

2002-05-23 Thread David Jacobs
I've had trouble benchmarking some of my mod_perl scripts on my desktop. The http_load tool we usually use (all under RedHat) works fine with static pages or CGI, but when I use it on mod_perl scripts I get a "byte count wrong" error. This could be because of something unexpected in the header