[Soc-2014-dev] Weekly Report #2 Shape Key Enhancements

[Lukas Tönne]

PIL profiling is sufficient when you want to examine just a single
function. For more complete profiling you could use gprof, which works fine
with gcc (not sure about clang, think it works there as well). I've been
using it in combination with a nifty little py tool called gprof2dot [1],
which turns gprof output into a fancy graphviz image, so following
convoluted callstacks becomes easier.

[1] https://code.google.com/p/jrfonseca/wiki/Gprof2Dot


On Fri, May 30, 2014 at 7:21 PM, Bastien Montagne <montagne29 at wanadoo.fr>
wrote:

>  Hi Grigory,
>
> About profiling, we generally simply use helpers from
> PIL_time_utildefines.h (in blenlib)…
>
> As for lib data, I’m not a great expert in this area, but I think since
> skeys are just mesh sub-data, it should not be hard to access them in
> readonly mode? Other coders should know better though, imho.
>
> Btw, did you see the comments I put on your proposal?
> http://wiki.blender.org/index.php/User:Revzin (UI topic is still a bit
> fuzzy…).
>
> Cheers,
> Bastien
>
>
> On 30/05/2014 17:14, Grigory Revzin wrote:
>
> Hello everyone
>
>  This was an academically tough week, but I've managed to clear the next
> week for half-time work and two weeks after that for (almost) full time
> gsoc!
>
>  This week:
> - Still struggling with making editmode mixing work, didn't have time to
> finish it, but almost there now.
> - Caught a crash in hashing if an edge didn't have any faces
>  - From a disussion with an artist a decision was made to implement
> compression of shape key data on disk save by storing only the vertices
> that have changed their positions in a given keyblock
>
>  Problems:
> - Shape keys on linked data aren't accessible from the UI and this needs
> to be sovled. Is this a deep problem that I can add into gsoc targets?
>  - Is there a standard way to profile Blender code for time?
>
>  Thanks,
> Grigory
> --
> Г. Г. Ревзин
>
>
> _______________________________________________
> Soc-2014-dev mailing listSoc-2014-dev at blender.orghttp://lists.blender.org/mailman/listinfo/soc-2014-dev
>
>
>
> _______________________________________________
> Soc-2014-dev mailing list
> Soc-2014-dev at blender.org
> http://lists.blender.org/mailman/listinfo/soc-2014-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/soc-2014-dev/attachments/20140530/aa7d19dd/attachment.htm