[Seaside-dev] Seaside-Core-lr.62

[Philippe Marschall]

2008/2/28, Lukas Renggli <renggli at gmail.com>:
> Thanks Philippe for the in-depth testing.
>
>
>  > WACacheTest clicking the links of increases a continuation count only
>  >  the second time, often clicking a link will reduce "foreign"
>  >  continuations. Might be a GC effect.
>  >  an explaining text couldn't hurt
>
>
> Yes, and the list item EscapeContinuation should could be removed, as
>  it is not needed anymore.
>
>
>  >  WACallbackTest 'Side Effect' doesn't seem to work because I see no
>  >  different to 'Idempotent'
>  >  an explaining text couldn't hurt
>
>
> It looks to me like it works. But frankly I don't understand that test
>  as well. What are the links and the counter needed for? I think the
>  only thing that is tested here is the order of the textInput
>  callbacks, and that the cancel has a higher priority. Who wrote that
>  test?
>
>
>  >  WAColusreTest, I'm not sure if 'go with ensure' does the right thing (tm)
>  >  the back button behavior seems broken
>  >  'ensure' -> 1a -> 'ensure' -> back-button to the first ensure -> 1a ->
>  >  1a -> 1a -> ...
>  >  an explaining text couldn't hurt
>
>
> It looks ok to me, but we have to check how the behavior was in
>  earlier versions of Seaside.
>
>
>  >  WACookieTest the remove button removes only the value
>
>
> Looks like it works to me. You are the expert on cookies, I usually
>  just eat them ;-)
>
>
>  >  WAErrorTest couldn't test, test will follow
>
>
> I tested it, it should work.

There is a back button issue with the resumable error test.
'Raise zero divide' -> 'Debug' -> Proceed -> back-button to the
walkback -> 'Debug' -> Proceed -> 'To be displayed' will be displayed
twice.

>  >  no idea what WAHomeTest home test should do
>
>
> I will add a comment. It tests that sending #home to a component
>  clears all its delegates.
>
>
>  >  WATableReportTest broken in my image
>
>
> Why? It works in my image.

Now works here too, was an image issue.

Thanks for the new Home and Callback tests.

Cheers
Philippe

>  It is probably a bit early now, but would it make sense to move all
>  the continuation based stuff into a separate package? Something like
>  Seaside-Flow-Core/Seaside-Flow-Tests? Like this, Smalltalk vendors
>  without a first-class execution-stack don't need to load that part of
>  Seaside. Of course, not loading Seaside-Flow, will result in a big
>  loss of features.
>
>  Lukas
>
>
>  --
>  Lukas Renggli
>  http://www.lukas-renggli.ch
>  _______________________________________________
>  seaside-dev mailing list
>  seaside-dev at lists.squeakfoundation.org
>  http://lists.squeakfoundation.org/mailman/listinfo/seaside-dev
>