Re: [gnome-desktop-testing] patch - gedit/gedit_chains.py
- From: Ara Pulido <ara ubuntu com>
- To: desktop-testing-list gnome org
- Subject: Re: [gnome-desktop-testing] patch - gedit/gedit_chains.py
- Date: Tue, 17 Feb 2009 09:01:20 +0100
Sandro Millien wrote:
I totally agree, that is the whole purpose of the gnome-desktop-testing
and the ubuntu-desktop-testing projects. If you look to the comparison
of two Gedits scripts, one using pure LDTP code
and the other using the library
you will see the difference.
Either we have a complex framework but simple test cases.
Either we have a simple framework but complex test cases.
We cannot have both simple framework and simple test cases. This is
I think it is better to put the complexity in the framework, and enable
people to write test cases easily.
In the case of the gnome-desktop-testing library we wanted to go a step
further to make the tests framework independent. If the chosen framework
changes, the scripts will remain the same.
This is why, I wrote ltccompiler.py a python script that takes in input
a .ltc file then translates it in a .ltc.py file .
A .ltc file can only see the ldtp namespace. i.e. : the "from ldtp
import *" is implicit. No other namespace is aloud.
Commands issued in a .ltc file are implicitly enclosed in try...except
I really like your ideas of reducing the power of test cases. Have you
had a look to the gnome-desktop-testing library code? Could you explain
(here and or in the wiki) how would you make both ideas (the ltc files
and the library) work together?
I am still working on it, there are many other features I want to put in
in it, but the goal is to absorb the complexity of writing test cases in
the framework and let people write test cases, easily an securely.
I putted some initial documentation on it here :
Please, do not hesitate branching the gnome-desktop-testing project or
the ubuntu-desktop-testing project to add your changes and proposal.
] [Thread Prev