[Catalyst-dev] hacking on Session + Session::State::Cookie

[Jonathan Rockway]

I've been working on these two modules recently.  That failing expires
test in the cookie dist (that has been giving me nightmares for the last
month or so :) is finally passing!

However, there is still some weird behavior that I'm noticing. =

C::P::Continuation only passes its tests if CATALYST_DEBUG is enabled,
and FormCanary (in git) can only access $c->session->{foo} if I call
$c->session in void context before accessing $c->request and /then/
$c->session->{whatever}.  That smells buggy.

I tested some random dists in svn against the new Session code in trunk,
and all of that is passing.  random =3D Store::[FastMmap|DBI],
Catalyst::Plugin::Authentication, and Angerwhale.

Anyway, below is a summary from IRC.  I need some sleep before I can
tackle the remaining issues, so just know that there is some potentially
broken/fixed/incomplete/awesomely delicious code in trunk.  No action is
required on anyone's part, I just want to let everyone know what's going
on so you don't have to wonder :)

IRC log:

18:57 <@jrockway> ok, i've figured out ::session
18:58 <@jrockway> the problem was that we were sending headers before
updating
                  the expiry, so it just got ignored
18:58 <@jrockway> that affected the $c->res->body case
18:58 <@jrockway> streaming ($c->res->write) was still broken
18:58 <@jrockway> because that never calls finalize, only finalize_headers
18:58 <@jrockway> so i move the expiry update to finalize_headers
18:58 <@jrockway> and then NEXT::finalize_headers
18:58 <@jrockway> and now everything works again

(editor's note: after i said that, I moved everything but the expiration
update back to finalize.  so anything cookie related happens before
sending headers, but everything else happens after the headers and body
have been sent.  as a result, that 06finalize test now passes again)

21:19 <@jrockway> catalyst::pluigin::continuation won't pass tests unless
                  CATALYST_DEBUG=3D1
21:19 <@jrockway> but, it appears that was the case before i touched
session...
21:22 <@jrockway> ahh, i know
21:22 <@jrockway> recently, calling "$c->session" is required before
                  $c->session->{whatever} will work
21:22 <@jrockway> i guess that depends on debug/no debug...
21:32 <@jrockway> unrelated
21:32  * jrockway pulls out hair

gnite.  sorry this is incoherent :)

Regards,
Jonathan Rockway

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 370 bytes
Desc: OpenPGP digital signature
Url : http://lists.scsys.co.uk/pipermail/catalyst-dev/attachments/20070921/=
e2a5625e/signature.pgp