[Catalyst] $c->res->redirect and $c->detach

[John Napiorkowski]

--- Ash Berlin <ash at cpan.org> wrote:

> John Napiorkowski wrote:
> > --- Matt S Trout <dbix-class at trout.me.uk> wrote:
> >
> >   
> >> John Napiorkowski wrote:
> >>     
> >>> Hi,
> >>>
> >>> I have a controller with some chained actions. 
> >>>       
> >> One of
> >>     
> >>> the actions (not an endpoint) will redirect
> given
> >>>       
> >> a
> >>     
> >>> certain value for the arguments.  However I find
> >>>       
> >> that
> >>     
> >>> this doesn't stop the chain from completing.  I
> >>>       
> >> still
> >>     
> >>> get a line about the redirect in the log, but
> the
> >>> endpoint in that chain still activates.
> >>>
> >>> I find that this doesn't happen when I use
> >>>       
> >> $c->detach,
> >>     
> >>> although in this circumstance detach is not the
> >>>       
> >> best
> >>     
> >>> use for me.
> >>>
> >>> Has anyone run into this before and found a way
> >>>       
> >> around
> >>     
> >>> it?
> >>>       
> >> Calling $c->res->redirect never stops anything
> from
> >> completing, all it does is 
> >> set the appropriate data on the response.
> $c->detach
> >> is the correct thing to 
> >> do here.
> >>
> >> Perhaps you need to rephrase your question in
> terms
> >> of why you feel detach 
> >> isn't ideal and how that interacts with the way
> your
> >> architecture is put together?
> >>     
> >
> > It is probably more a case of my not understanding
> the
> > best way to do this.  I'm still a bit stuck in my
> old
> > way of putting these types of applications
> together. 
> > Let me explain a little more what I am doing and
> > perhaps this is solvable with a more
> catalyst-think
> > solution.
> >
> > I'm using chained actions as part of a wiki
> system. 
> > Basically I have actions that descend down a url
> which
> > first finds a wiki (I have multiply wikis in the
> > system) and then a wiki page and then an action
> > associated, such as create, view or edit.  Now
> when I
> > run the action associated with finding the page in
> the
> > database based on the page title what I'd like to
> do
> > is redirect to the create action when we don't
> find
> > that a  page exists.  I think this is the best
> action
> > for a wiki and what is expected (as long as you
> have
> > create rights).
> >
> > Now I'd like the URL to actually change to
> represent
> > that you are on an edit page, which is why the
> detach
> > doesn't do it for me.  If this is really the best
> > thing to do I guess I can bend my head around it,
> but
> > I find in not so intuitive to have a give url show
> > both content or a form.
> >
> > I find it easier to write the tests for this if I
> can
> > detect the redirect action.
> >
> > Doing a $c->detach('/end') after the redirect was
> > suggested to me, although I find this can give an
> > error since my default end action looks for a
> template
> > which is this case doesn't exist.
> >
> > Maybe I should do both a redirect and a detach?
> >
> > I can show you some code for this if that would
> help
> > you understand my thinking, although it's in two
> > different modules (I'm using the :Chained('.')
> feature
> > to help organize the code and it works
> fantastically).
> >
> > Let me know what you think.  Thank you for your
> > feedback and suggestions.  --john
> >
> >   
> I think the problem here is that you're expecting
> redirect to do more 
> for you than it does:
> 
>        $res->redirect( $url, $status )
> 
>        Causes the response to redirect to the
> specified URL.
> 
>            $c->response->redirect(
> 'http://slashdot.org' );
>            $c->response->redirect(
> 'http://slashdot.org', 307 );
> 
> Thats all it does. It doesn't stop the chain. Try
> this instead
> 
> $c->res->redirect($url); return 0;
> 
> Although I'm not sure if the `return 0' will dwim in
> chained dispatch. 
> Try it and see.

I didn't find that return 0 stopped aborted that chain
at tht point.  Right now the way I am gettin around
this it is that I have each endpoint test to make sure
the database resultset exists.  This is quite a hack
for me since the whole point of creating a point in
the chain to load the page from the database from was
to avoid having it in three locations.  I'll leave it
this way for now but I'll have to come back and
rethink this more carefully since I'm guessing this
type of pattern will reoccur for me.

Thanks again,

--john

> 
> Ash
> 
> 
> _______________________________________________
> List: Catalyst at lists.rawmode.org
> Listinfo:
> http://lists.rawmode.org/mailman/listinfo/catalyst
> Searchable archive:
>
http://www.mail-archive.com/catalyst@lists.rawmode.org/
> Dev site: http://dev.catalyst.perl.org/
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com