perl-modperl mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Wyllie <>
Subject Re: [RELEASE CANDIDATE] please test mod_perl-1.99_13-dev.tar.gz
Date Sat, 06 Mar 2004 03:18:02 GMT

All tests passed on FreeBSD-5.2


Stas Bekman wrote:

> A release candidate for mod_perl-1.99_13 is available:
> Please test and report any failures to this list.
> If everything goes well, 1.99_13 will be released on Tuesday, so those 
> who do
> mod_perl at work can test on Monday and those at home over the weekend.
> Changes since 1.99_12:
> respect $ENV{APACHE_TEST_STARTUP_TIMEOUT} settings if any [Stas]
> Added tests for issuing subrequests from filters [Geoffrey Young]
> Updated to the new Apache License Version 2.0 [Gozer]
> Drop the support for making GATEWAY_INTERFACE special. It's not needed
> as $ENV{MOD_PERL}, available in both mod_perl generations, should be
> used to test whether the code is running under mod_perl. [Stas]
> Handle correctly the situation when response HTTP headers are printed
> from the handler and the response body starts with \000, which is the
> case with some images like .ico. [Stas]
> Apache::PerlSections->dump() and store(filename) [Gozer]
> expose $c->keepalive related constants and $c->keepalives counter
> [Stas]
> Perl handlers are now guaranteed to run before core C handlers for
> all request phases.  [Geoffrey Young]
> Fix the STDIN/OUT overriding process to handle gracefully cases, when
> either or both are closed/bogus (the problem was only with useperlio
> enabled perl) [Stas]
> copy apr_table_compress logic from later httpd versions in case mod_perl
> is built against 2.0.46, as mod_perl now requires it internally.  users
> should be aware that 2.0.47 may become the oldest supported httpd version
> in the near future.  [Geoffrey Young]
> Fix the corruption of the httpd process argv[0], caused by $0
> manipulating [Stas]
> ModPerl::MethodLookup::lookup_method now handles sub-classed objects
> [Stas]
> standard %ENV population with CGI variables and contents of the
> subprocess_env table (such as SetEnv and PassEnv) has been delayed
> until the last possible moment before content-generation runs.
> PerlSetEnv and PerlPassEnv are each an exception to this and are
> placed in both %ENV and the subprocess_env table immediately,
> regardless of the current [+-]SetupEnv setting.
> [Geoffrey Young]
> fix PerlAddVar configuration merging [Geoffrey Young]
> Anonymous subs are now supported in push_handlers, set_handlers,
> add_input_filter, etc. A fast cached cv is used with non-ithreaded
> perl. A slower deparse/eval approach (via B::Deparse) is used with
> ithreads enabled perls. Further optimizations are planned for the
> latter case. [Stas]
> ht_time w/o the pool is now available only via override/restore compat
> API. format_time, has been renamed back to ht_time, and the default
> values for fmt, time and gmt are now supported. [Stas]
> it's now possible to push new handlers into the same phase that is
> running at the moment [Stas].
> when $r->handler($new_handler) is called from a response phase, it now
> checks that the response handler type is not switched (e.g. from
> 'modperl' to 'perl-script') from the currently used one [Stas]
> Since Apache::SubProcess is now part of the mp2 API, add
> $r->cleanup_for_exec as a noop in Apache::compat. That function is no
> longer needed in Apache2. [Stas]
> When 'perl Makefile.PL PREFIX=/foo/bar' is used and mod_perl 1 is
> found, but at different prefix no longer require
> MP_INST_APACHE2=1. [Stas]
> modperl_mgv_resolve now croaks when a module scheduled for autoloading
> fails to load.  AutoLoaded modules shouldn't silently fail. [Stas]
> Perl(Input|Output)FilterHandler handlers are now always AutoLoaded, as
> if '+' prefix was used. This must be performed to get the access to
> filter attributes long before the filter itself is executed. [Stas]
> APR/Pool.xs has been reimplemented.  The problem with the previous
> implementation is that a dead perl pool object could hijack a newly
> created pool, which didn't belong to that object, but which happened
> to be allocated at the same memory location. The problem is that
> apr_pool_user_data_set/get has no mechanism to check whether the pool
> has changed since it was last assigned to (it does but only in the
> debug mode). It really needs some signature mechanism which can be
> verified that the pool is still the same pool. Since apr_pool doesn't
> have this feature, the reference counting has been reimplemented using
> a plain sv reference. Several new (mainly hijacking) tests which badly
> fail with the previous impelementation have been added. [Stas]
> fix calling $r->subprocess_env() in a void context so that it only
> populates %ENV if also called with no arguments.  also, make sure it
> can be called more than once and still populate %ENV.
> [Geoffrey Young]
> add APR::Brigade::pool() to allow access to the pool associated with
> the brigade  [Geoffrey Young]
> make 't/TEST -startup_timeout secs' working (previously user's value
> was ignored) [Stas]
> ModPerl::Registry and friends now support non-parsed headers scripts,
> whose filename =~ /^nph-/, identically to mod_cgi. + test [Stas]
> implement APR::Brigade::length() and APR::Brigade::flatten() (the
> latter implements a wrapper for apr_brigade_flatten, but also includes
> an emulation of apr_brigade_pflatten) as [Geoffrey Young]
> ($r|$s)->add_config() now die if failed (previously returned the
> error) [Stas]
> fix context problems in <perl> sections and
> PerlModule/PerlLoadModule/PerlRequre under threaded mpms w/
> PerlOptions +Parent/+Clone in Vhosts + TestVhost::config test. [Stas]
> Implemented Apache::get_server_version and Apache::get_server_built
> as constant subroutines [Geoffrey Young]
> Moved some functions out of the Apache:: namespace:
>   Apache::unescape_url()       is now Apache::URI::unescape_url()
>   Apache::log_pid()            is now Apache::Log::log_pid()
>   Apache::LOG_MARK()           is now Apache::Log::LOG_MARK()
> [Geoffrey Young]
> if MP_AP_PREFIX is used apxs and apr-config from the apache build tree
> won't work, so it can't co-exist with MP_APXS and MP_APR_CONFIG build
> options - ensure that this doesn't happen. [Stas]
> server_root_relative() now requires either a valid pool or an $r, $s,
> or $c object as a first argument.  also, the returned result is a
> copy, protecting against cases where the pool would go out of scope
> before the result.  [Geoffrey Young]
> Check the success of sysopen in tmpfile() in compat [Geoffrey Young]
> make sure DynaLoader is loaded before XSLoader, not only with perl
> 5.6.1, but always because of the issues with <Perl> sections are
> loaded from +Parent vhost [Stas]
> added ($r|$s)->is_perl_option_enabled($option_name), to test for
> PerlOptions + tests [Stas]
> On Solaris add a workaround for xs/APR/APR/Makefile.PL to build
>, correctly linked against apr and apr-util libs, by addding the
> missing -R paths corresponding to -L flags. EU::MM was adding them via
> LD_RUN_PATH instead of using -R, but since perl's lddflags may have -R
> it overrides LD_RUN_PATH. So explicitly add anything that may go into
> LD_RUN_PATH via -R. Also make sure that -R coming from Apache will
> appear first. [Brad Lanam <>]
> 'make dist' now generates and picks Apache-Test/META.yml which was
> always reported missing, as it was included in Apache-Test/MANIFEST
> [Stas]
> fix the $r->read function to return undef on failure similar to the
> core perl function and make $! available for those who test for read()
> failures. [Stas]
> Make sure that pnotes are destroyed after PerlCleanup handlers are
> finished and not before + test. [Stas]
> __________________________________________________________________
> Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
>     mod_perl Guide --->

Report problems:
Mail list info:
List etiquette:

View raw message