X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=TODO;h=3a5094aad8a24c04c7ba269fb4d6aba33d33b51f;hb=9bc5a5f177eb2ae1c3eb64909c75de15de843ba4;hp=b304f696a337bede1fac08516a96d8946eb1e6d4;hpb=b1e0cb6d4ded3d4223517d4adce6caef23112946;p=catagits%2FCatalyst-Runtime.git diff --git a/TODO b/TODO index b304f69..3a5094a 100644 --- a/TODO +++ b/TODO @@ -1,93 +1,56 @@ -Pending patches: - - meta test for MX::Emulate::CAF needed by Catalyst::Plugin::Cache::Curried - - - Class::Accessor::Chained::Fast test for MX::Emulate::CAF, to fix - HTML::Widget - - - Re-opening packages with MX::Emulate::CAF (for - Catalyst::Plugin::HashedCookies) - -Back-compat investigation / known issues: +Known Bugs: - - Catalyst::Plugin::Server is still buggered in various ways.. - - Fix TODO tests in t/aggregate/live_engine_request_body.t + - Bug ->go or ->visit causes actions which have Args or CaptureArgs called + twice when called via ->go or ->visit. - - Common engine test failures, look into and get tests into core. + Test app: http://github.com/bobtfish/catalyst-app-bug-go_chain/tree/master - - CatalystX-Imports, Class::MOP doesn't consider anon subs in the symbol - table as methods, tests + fix, or explanation and documentation? - (rafl & phaylon) +Compatibility warnings to add: - - Run another round of repository smokes against latest 5.80 trunk, manually - go through all the things which are broken (t0m). - - - Catalyst-Plugin-Cache dies due to mk_accessors('meta') - - - Catalyst-Action-REST appears to have real issues, investigate - - - Catalyst-Model-Search ditto - - - CatalystX-CRUD ditto - - - Catalyst-Engine-Apache - Deep recursion on subroutine - "Catalyst::Action::execute", investigate once all other tests pass - - - Catalyst-Plugin-Scheduler - - Catalyst-Engine-HTTPEngine - - Catalyst-View-HTML-Template - - Catalyst-View-XSLT - - Catalyst-Plugin-Setenv - - Catalyst-Plugin-RequireSSL - Bullshit 'cannot locate' errors - Bug in smoke test rig. + - $self->config should warn as config should only ever be called as a + class method. - - Issues with TWMC not being loaded when it used to be in 5.70 - (Bill Moseley) +Proposed functionality / feature additions: - - Fix memory leaks (I already tried Devel::Leak::Object, but no joy). + - Log setup needs to be less lame, so Catalyst::Plugin::Log::* can die + in a fire. Having $c->log_class would be a good start. kane volunteered + to do some of this. - - In 5.70, you can have a component which doesn't inherit from ::Component, - and has no new method, and it'll do the right thing for you. We just - ignore that, we should still support that, but warn for/deprecate it so - it can go for 5.9X... - This all obviously needs better tests :/ + Simple example: Catalyst::Plugin::Log::Colorful should just be a + subclass of Catalyst::Log, no ::Plugin:: needed. -Cleanups: + See also: Catalyst::Plugin::Log::Dispatch and + http://github.com/willert/catalyst-plugin-log4perl-simple/tree - - Catalyst-Plugin-Authorization-ACL, Can't locate object method "tree" - via package "Catalyst::Dispatcher", fix the plugin as tree was never - a public method. + Fixes to component interfaces to deal with anon classes: - - Catalyst-Model-Akismet, incorrectly inherits from - Catalyst::Component::InstancePerContext, should compose as a role. - Fix the module. - - - Update Test suite to not assume MyApp ISA Controller - - After that set up attr handlers that will output helpful error messages - when you do it as well as how to fix it. (done already?) +22:16 <@mst> t0m: explain what breaks +22:16 <@mst> I'm not 100% sure we have the same problem in mind +22:17 <@t0m> well, various shit in Catalyst utils to resolve app from MyApp::Controller::Foo breaks +22:17 <@mst> Caelum: http://scsys.co.uk:8001/31240 +22:17 <@mst> t0m: the trick would be to not need to do that +22:17 <@mst> t0m: $self->_application should be fine +22:18 <@t0m> $action->class can't be fed back into $c->component() +22:18 <@t0m> which breaks ::REST +22:18 <@mst> right +22:18 <@mst> we need to +22:19 <@mst> (a) add an ->_component_name or similar method to controllers +22:19 <@mst> (b) pass action objects either that or the controller +22:20 <@t0m> do you have a controller calling $action->dispatch? Course you must, the action lives in a controller.. +22:21 <@mst> and that doesn't matter anyway +22:21 <@mst> dispatch has $c +22:21 <@mst> it needs to do +22:21 <@mst> $c->component($self->_component_name) +22:21 <@mst> and that _component_name needs to come from the controller at register_actions time +22:22 <@t0m> confound: Oi, please to be releasing my Action::REST fixes, I'm about to do more.. +22:22 <@t0m> mst: right, gotcha +22:22 <@mst> note that needs to happen in core +22:22 <@mst> then actions can use that instead of $self->class +22:22 <@mst> and shit will work - - Comments marked /Moose TODO/i in Catalyst::Request re {_body} (t0m) + The horrible hack for plugin setup - replacing it: - - Eliminate all instances of $instance->{$key}, I think the only thing - left is lib/Catalyst/Engine/HTTP.pm: $self->{inputbuf}, which I haven't - touched as it is used as an lvalue in a lot of places (t0m) +15:30 <@mst> t0m: have a look at the Devel::REPL BEFORE_PLUGIN stuff +15:30 <@mst> I wonder if what we need is that combined with plugins-as-roles - - Find whatever in the test suite craps a file called GLOB(0xfoo) onto my - disk and fix it. (Believed to be in the optional tests?) (t0m) -Documentation: - - - extends in components must be inside a BEGIN block. - - - How to write Moosified Catalyst components. - - - Manual / Tutorial updates - - - Fix the Roadmap to be less full of lies. - -Profiling: - - - vs 5.70 and optimisation as needed. - -Tests: - - - Moosified test application?