Re: [GsoC] your latest commit of tasks_udd.py

[Akshita Jha <zenith158@gmail.com>]

Hi,

On Mon, Jul 27, 2015 at 11:18 PM, Andreas Tille <andreas@an3as.eu> wrote:
>
> We should allow some sloppyness for sure.  In any case the packages (and
> the sequence of packages) should fit.  Also the metainformation should
> be identical (popcon, versions, Vcs-fields etc.)  Since we have nearly
> identical templates it might possibly the simplest way to do a diff by
> ignoring spaces.  Just try to find something "sensible" that ensures
> that the html rendering looks somehow identical regarding its content.
>

I have written a script: test_output.py which parses the old and the new tasks files and gives the difference between the two. The script test_ouput.py makes use of HTMLParser library[0] and difflib[1] standard libraries of Python 3.

To run the script : ./test_output.py <Blend> eg: ./test_output.py debian-med

Ouputdir : ./tests/<Blend>/<task_file.diff>

For now, I am just computing the difference between the english versions of task files. I feel it'll be faster this way and I am assuming the translations will have the same differences. This is not an issue and is easy to change.

The test_output.py for now is in webtools_py3 folder. Once you confirm that the diff output is as desired, I'll move it to tests folder. FWIW, the difference in the tasks files is mainly because of the order  (by order I mean that the order of packages within a section may differ but the packages in each section remain the same).

Needless to say, the script may require lots of changes. Can you please go through the output and let me know.

Regards,

Akshita

[0] https://docs.python.org/3/library/html.parser.html
[1] https://docs.python.org/2/library/difflib.html