[Catalyst] using catalyst as a developer

[Yuval Kogman]

On Mon, Mar 20, 2006 at 23:48:12 -0800, David Rio Deiros wrote:
> Hi there,
> 
> I would like to know how catalyst developers work with the sources
> in order to improve/patch/fix catalyst.

Yay!

> 2. checkout catalyst repository in one of the @INC perl directories.

Nope. The tree is distribution oriented, so it doesn't really match
@INC...

What I do is install the plugins I don't develop, and for the ones I
do (Catalyst itself, sessions, authentication, etc) i have a simple
alias in my .profile

	alias catlib='export PERL5LIB=`ls -d /Users/nothingmuch/Perl/catalyst/{Catalyst,*Auth*,*Session*,*Log*,*HTML-Widget*,*BindLex*,*Workflow*,*Stack*}/lib | tr \\\\n :`'

I could make this static but I sometimes want to use installed
modules over existing ones.

Ingy has a small script to generate an @INC valid dir out of a
dist-tree using symlinks... I'll ask him to post it.

> 3. create a test catalyst application that uses the feature you are working
>    on.

Yes, and make sure that this can be converted to unit tests or
TestApp code later.

> 4. patch/modify code
> 5. reload application

foo_server.pl -R might do this if i'm really desparate. I always
wanted it to respond to changes in all of keys %INC with some
special flag.

> 6. test new feature

Not enough - functionality testing is only half the story. Write
real unit tests.

> 7. generate patch and submit to the mailing list/ #catalyst for revision
> 8. if patch approved it and you have commit privileges, commit changes

If it's a small patch and you're sure of yourself, you can just
commit it.

Otherwise ask on #catalyst.

Remember that this is revision control and changes can be undone
easily. If it seems like a right thing to do and nobody violently
disagrees then worst case scenario someone will revert it later.

If it's a medium patch, ask catalyst-dev people.

If it's a big patch there's the 3 dev rule - at least 3
catalyst-dev members must agree that it's a good thing before it's
comitted.

Also, mst's general rule for who gets commit: if you're working on
it and you've shown that you're sane based on unidiffs, then you get
access.


A slight digression: we've discussed the issue of future directions
of Catalyst, and hopefully we will have a simple road map that will
help clarify the type of code we want, like what is high priority,
etc. The general idea is that we want to refactor some things that
don't need to be core, and get stability improvements. Bigger
changes will need to be more thoroughly discussed.

I hope sri gets well soon and publishes this roadmap thing ;-)

-- 
 ()  Yuval Kogman <nothingmuch at woobling.org> 0xEBD27418  perl hacker &
 /\  kung foo master: /me wields bonsai kittens: neeyah!!!!!!!!!!!!!!!!

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 191 bytes
Desc: not available
Url : http://lists.rawmode.org/pipermail/catalyst/attachments/20060321/d398793f/attachment.pgp