Re: [gnome-desktop-testing] patch - gedit/gedit_chains.py

[Ara Pulido <ara ubuntu com>]

Hey!
> I would prefer moving FAIL/SUCCESS and other "status codes" to something
> like desktoptesting.check.StatusCodes and have it imported specifically
> instead of a "import *"
The import * was not added to trunk.
> >  
> >  try:
> >  
> > @@ -29,15 +29,15 @@
> >      elapsed = stop_time - start_time
> >      
> >      testcheck = FileComparison(oracle, test_file)
> > -    check = testcheck.perform_test()
> > +    status = testcheck.perform_test()
> >  
> >      ldtp.log (str(elapsed), 'time')
> >      
> > -    if check == desktoptesting.check.FAIL:
> > +    if status == FAIL:
> >          ldtp.log ('Files differ.', 'cause')
> >          ldtp.log ('Files differ.', 'error')
> >  
> > -except ldtp.LdtpExecutionError, msg:
> > +except ldtp.LdtpExecutionError:
> >      raise
> Wouldn't it be better if we abstract ldtp.LdtpExecutionError into
> desktoptesting.Exceptions.LdtpExecutionError?
We wouldn't want to add to much complexity to the framework. The 
important thing are tests, and not maintaining a complex framework. 
Nevertheless, your design comments are more than welcome in the wiki [1].

Cheers,
Ara

[1] http://live.gnome.org/DesktopTesting/Documentation/Design