100 percent.

Doug


On May 8, 2008, at 1:36 PM, Paul Winkler wrote:

> Okay. Do you concur with my assessment about what the test should
> expect?
>
> On Thu, May 08, 2008 at 01:13:29PM -0400, Douglas Mayle wrote:
>> To understand what's going on hear, you really have to go back and  
>> look at
>> the README.txt it was originally copied from.  The key lines are:
>>
>> 387 	Login [to be done]
>> 388 	
>> 389 	    >>> view.request.form['__ac_name'] = 'foobar'
>> 390 	    >>> view.request.form['__ac_password'] = 'testy'
>> 391 	    >>> output = view()
>>
>> When I looked at this stuff, no one could tell me what the missing  
>> steps
>> were to make login actually work through the views (as opposed to  
>> the magic
>> forcing of login through zope).  If you know how to get a login  
>> working
>> properly, than we should be able to fix these tests (among others)...
>>
>> Doug
>>
>>
>> On May 8, 2008, at 12:23 PM, Paul Winkler wrote:
>>
>>> Okay, this gets weirder.  The test fix I comitted in r16643 now  
>>> fails
>>> - the top nav has reverted to its previous (wrong-looking to me)
>>> behavior.  Does anybody have any idea what's going on?
>>>
>>> - PW
>>>
>>> On Wed, May 07, 2008 at 08:02:13PM -0400, Paul Winkler wrote:
>>>> Or rather, what's mysterious to me is why this test ever passed
>>>> before.
>>>>
>>>> In http://trac.openplans.org/openplans/changeset/16643 I changed a
>>>> test in oc-vacuum to assert somethign that looks more sensible to
>>>> me. If we're testing the topnav for a logged-in user, I don't  
>>>> know why
>>>> we'd expect to see "login" and "join" links. We should instead  
>>>> see a
>>>> "logout" link.  So that's what I changed it to do.
>>>>
>>>> Is there a reason it was the way it was before?  If so, I'm  
>>>> wrong, and
>>>> we have a new bug, and r16643 should be reverted.
>>>>
>>>>
>>>> -- 
>>>>
>>>> Paul Winkler
>>>> http://www.openplans.org/people/slinkp/profile
>>>> yahoo: slinkp23
>>>> AIM:   slinkp1970
>>>>
>>>>
>>>> --
>>>> Archive:
>>>> http://www.openplans.org/projects/opencore/lists/opencore-dev/archive/2008/05/1210206525293
>>>> To unsubscribe send an email with subject unsubscribe to
>>>> opencore-dev@....  Please contact
>>>> opencore-dev-manager@... for questions.
>>>>
>>>>
>>>>
>>>>
>>>
>>> -- 
>>>
>>> Paul Winkler
>>> http://www.openplans.org/people/slinkp/profile
>>> yahoo: slinkp23
>>> AIM:   slinkp1970
>>>
>>>
>>> --
>>> Archive:
>>> http://www.openplans.org/projects/opencore/lists/opencore-dev/archive/2008/05/1210263829341
>>> To unsubscribe send an email with subject unsubscribe to
>>> opencore-dev@....  Please contact
>>> opencore-dev-manager@... for questions.
>>>
>>>
>>>
>>>
>>
>>
>>
>>
>
> -- 
>
> Paul Winkler
> http://www.openplans.org/people/slinkp/profile
> yahoo: slinkp23
> AIM:   slinkp1970
>
>
> --
> Archive: http://www.openplans.org/projects/opencore/lists/opencore-dev/archive/2008/05/1210268189952
> To unsubscribe send an email with subject unsubscribe to opencore-dev@... 
> .  Please contact opencore-dev-manager@... for  
> questions.
>
>
> !DSPAM:4037,48233ada222541439371379!
>

I mean, simulating a login the way a user would login in the tests.   
When you use the convenience methods to login, your bypassing our  
login interface, and rmarianski and I couldn't get our login interface  
to work in the tests.  I think Jeff can be svn blamed for those lines  
of code below, but he had no memory of having ever touched that.  Do  
you know how to use our login interface (the views with a request  
object) to log a user in?  If not, we can't 1) test the login  
interface in doctests, and 2) test the logout interface (either the  
defualt, or the js based logout) in doctests.  The logout js based  
interface is currently tested exclusively through flunc...

Dug

On May 8, 2008, at 3:19 PM, Rob Miller wrote:

> Douglas Mayle wrote:
>> To understand what's going on hear, you really have to go back and  
>> look at the README.txt it was originally copied from.  The key  
>> lines are:
>> 387 Login *[to be done]*
>> 388
>> 389    >>> view.request.form['__ac_name'] = 'foobar'
>> 390    >>> view.request.form['__ac_password'] = 'testy'
>> 391    >>> output = view()
>> When I looked at this stuff, no one could tell me what the missing  
>> steps were to make login actually work through the views (as  
>> opposed to the magic forcing of login through zope).
>
> i don't understand this bit.  what do you mean by making login work  
> "through the views"?  we have views that are designed to log you in  
> (i.e. the ones that the login form submits to), and the test case  
> classes have convenience methods if you want to log in as an  
> arbitrary user while you're in a test case, but the idea of logging  
> in "through" an arbitrary view makes no sense to me.
>
> if you clarify the need, i can probably help.
>
> -r
>
>
> --
> Archive: http://www.openplans.org/projects/opencore/lists/opencore-dev/archive/2008/05/1210274271115
> To unsubscribe send an email with subject unsubscribe to opencore-dev@... 
> .  Please contact opencore-dev-manager@... for  
> questions.
>
>
> !DSPAM:4037,4823525c280101431913854!
>