[Catalyst] Testing controller which require login.

[Peter Karman]

Louis Erickson wrote on 05/14/2009 09:39 AM:
>  
> I'm using AuthTkt, and it works very well.

glad to hear it.

If you're using AuthTkt, then the whole cookie this is a non-issue. Just
pass the ticket as a URL param like the tests do. The plugin supports
both ways (cookie and param).

> 
> Why didn't it occur to me to look at the tests for it?

I have had to learn that habit myself.

> 
> On Thu, 14 May 2009, Peter Karman wrote:
> 
>> You can see the AuthTkt tests here[0]. The my_request() function just
>> sets $ENV{COOKIE} and lets the standard AsCGI magic work. I'm sure
>> that's a Bad Way to Do It, but it works.
> 
> Why is it a bad way?  If it's a standard and documented part of AsCGI, 
> then it should continue to work properly, shouldn't it?
> 

It should probably be:

 local $ENV{COOKIE} = $cookie;

instead. It's a little nitpicky, since in this particular case setting
the global env var doesn't hurt, but in general I try to avoid setting
class/global/env vars unless I really want global behaviour.

> The question I asked in the other message about cookies and Catalyst::Test 
> is: Is that a good idea?  Should it get added and/or documented, and let 
> people use it, or should people be pointed at Mechanize or other tools?
> 

Seems like a good idea to me to make some sort of official way to
get/set cookies in Catalyst::Test. I won't have time soon to try it
myself; care to try a patch yourself?


-- 
Peter Karman  .  peter at peknet.com  .  http://peknet.com/