[Soc-2008-dev] Week 6 report

[Dhanannjay (DJ) Deo]

Work this week
===========

*1: Code to write blender databases to a flat file, human readable -*
I am able to nicely read through the blender library data and dump the
object info as necessary (this completes 1), also with the option of dumping
python dictionary to file.
*
2: code to compare two of these dump files against each other and
generate warnings*

I am not using best of the graph comparison algorithms, but I have
something which works for 2. If there are some differences it can
generate warnings, which can be selectively suppressed.
*
3: code to import a file, dump it to the human readable format, then
compare against file in the (already created regression test suite).*

Tried for a controller2 file from blender regression suite. Requires further
testing.
Mostly because not many other file formats do not support hierarchial data.
*
4. demonstrate the mechanism to include exceptions, for the process of
   native file-> imported data-> regression file comparison*

Have a simple exception which can be thrown from and caught at the test
suite.
Working on it right now, mostly will be done tonight, not willing to keep
the report until then.

*5. Also *
I was also supposed to create an input file for each file format (some
cases)
which describes which parameters to call the exporter with, for
exhaustive calling of exporters within a test suite. This was
necessary to bring consistency in the test suite, since not only
exporters use parameters, but they also use different number of
parameters.

I am therefore adding a "test_exporter" function for each exporter
(and later importer), which accepts a single parameter, a list of
further parameters.
The list which can be stored into a dump file, and created from it by
the test suite in runtime.

*More details and tangible outputs have been reported to Mentor earlier this
week.*

Work planned next week
==================

I can now see the components and functional prototypes of are shaping up.
The mid term report will contain testing and demonstrations of most of the
features of the test suite.

The mechanisms are now ready for rigorous testing, which should be done
before midterm evaluations. More tangible targets will be decided during the
meeting.

The deadline for midterm evaluations is Monday 14th July, so there one full
week to test all the features and documentation.

Issues
=====
I dont have too many native files as of now, but the mechanism is ready for
demonstration with whichever I have.

Schedule
=======

The work has very much deviated from the original proposal and the idea, but
have completed the targets as decided with the mentor for this week. And
will continue to do so.


Cheers ALL;

Happy Independence Day.

--
Dhanannjay (DJ) Deo

(Vidyaya Amrutam Ashnute, Let us go beyond life and death with the knowledge
of truth)


On Fri, Jul 4, 2008 at 4:19 PM, André Pinto <andresusanopinto at gmail.com>
wrote:
> Week 5
> ======
>
> Last week
> ---------
> This week I worked only on simple deform.. (it turned out I spend more
> time on UI things.. and I didnt had time for touch BVH tree)
>
> So on simple deform:
> *Added modes Squash and Strech to SimpleDeform (need testing.. and I
> need to find the right math function XD)
> *Changed the add modifier menu to show a submenu for selecting the
> type of simpledeform.
> *Added lower and upper options on all modes of simple deform
> *Fixed bend
> *Added vertex groups on SimpleDeform
>
>
> Next week
> ---------
> I thought I would have add time to start addressing the bvhtree
> (spherecasting) but it turned out wrong.
> So thats my target for the coming week.
>
> Issues
> ------
> I spent more time than I initially planned on UI code, because I had
> an hard time reading, understanding, and debugging it, since that was
> a part of blender code I hadn't touched yet. (and also, pardon my
> expression, because it looks a big mess).
>
> Also it seems a general agree on #blendercoders (or well to Briggs and
> broken) that SimpleDeform should actually be several separated
> modifiers (both on UI level as well on code level).. so that needs a
> bit of review from my part.. to find an efficient way (in code terms)
> to do it without code duplication and without bloating the code with
> preprocessor code.
>
> Also doing that after the modifier refactor (proposed by Artificer)
> goes effective would be nice.. so that I could make some design  that
> fits that model better.
>
>
> Schedule
> --------
> Blender UI made me lost quite a few time :S
> Altought I still think I am on schedule.. but I need to be careful on
> future not to waste time on UI
> _______________________________________________
> Soc-2008-dev mailing list
> Soc-2008-dev at blender.org
> http://lists.blender.org/mailman/listinfo/soc-2008-dev
>



-- 
--
Dhanannjay (DJ) Deo

(Vidyaya Amrutam Ashnute, Let us go beyond life and death with the knowledge
of truth)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/soc-2008-dev/attachments/20080704/26070d77/attachment.htm