[jifty-devel] I'm designing a DSL for jifty's web testing...

[Jesse Vincent]

> A couple of comments though (nitpicking really). I read the file
> declarative-test-design.

Cool. have a look at Test::WWW::Declare, where I'm starting to actually
write code?

> 
> expect_url initially made me think that this url was to be found on
> the page, rather than being the url of the page itself. Perhaps
> something like 'expect_page_url' or 'page_url_is' would be clearer.
> For checking that urls are on the page will you use something like
> find_url '/foo/bar' ?

I don't actually like the naming of that stuff much yet ;) I agree with
'expect_url' not being right. and it doesn't quite fit with the scheme
I'm currently using.
 
> 
> Each 'test' block is a separate test according to the comments, and if
> anything inside a test block fails it is aborted. Does this mean that
> the state is reset on exiting a test block (ie mech returns to the
> page it was on when entering the test block)?

I'm not sure yet, but I could see each "flow" localizing its mech
object. I'm not sure that would fully dwym.


> If the state is not reset then the order of blocks becomes important,
> and a block failure could lead to other blocks failing - which is not
> ideal.

I agree 
> 
>    follow_link 'my folks' => 'My tasks' => 'Braindump';
> 
> Inspired - I am sold on this syntax just by seeing that.
> 
> 
> Will there be a way to name the test blocks so that failure reporting
> is clearer? perhaps something like:
> 
> test 'login and invite a friend' {
>  ... tests.
> }
> 
> 
> One of the real strengths of this sort of testing is that it reads as
> a natural language and so non-techs (i.e. customers) can make sense of
> it. With this in mind I think that it might well be worth showing it
> to some people who know nothing of WWW::Mech etc to see if it reads
> clearly for them as well.
> 
> Excellent stuff.
> 
> Cheers,
>  Edmund.
> 
> 
> PS I gather that BTDT is your internal name for hiveminder. What does
> it stand for? "bout time [I] did that" perhaps?
> _______________________________________________
> jifty-devel mailing list
> jifty-devel at lists.jifty.org
> http://lists.jifty.org/cgi-bin/mailman/listinfo/jifty-devel
> 

--