[Soc-2014-dev] Weekly Report #02 - Viewport FX III

[Sergey Sharybin]

Hi,

Answers are inlined.


On Sun, Jun 1, 2014 at 2:59 AM, Jason Wilkins <jason.a.wilkins at gmail.com>
wrote:

>
>
> On Saturday, May 31, 2014, Sergey Sharybin <sergey.vfx at gmail.com> wrote:
>
>> I'm not sure why to blame merges. We couldn't postpone everything
>> actually. Not even sure why it might be an issue for the project. You might
>> easily ignore them for now.
>>
>
> I wouldn't suggest that anything be postponed.  I could put off
> merging, but I'm not convinced it would save work later.  Seeing as how
> this year was significantly less work than last year supports that idea
> (even though it is anecdotal).  The difference is merging after 2 months
> versus 6.  It has to be done eventually.


We're already trying to not do major changes in drawing area for quite a
while already. Changes there as minimal as we can make them without leaving
Blender in a stagnated state feature-wise.

Also, it'll help starting merges back to master to master to prevent
headache on merge in the future. Just mentioning.


> I actually feel that no gl calls should be made outside of the gpu source
> directory.  I'll continue this thought below...
>

This i'm not so much sure about actually. It has own cons and pros.


> That is why I've taken to calling the compatibility layer a "replacement".
>  It is not compatible with legacy OpenGL, but tries to capture how blender
> actually uses legacy OpenGL and only implements that functionality which
> needs to be replaced since it was removed from core.  It has a similar
> interface to make porting easier.
>

That's what'm talking about. Why to try mimmic OpenGL usage in Blender
rather than making it so Blender uses OpenGL in a way it was designed to.

Correct me if i'm wrong, but seems you've got immediate-mode-like API which
translates stuff to new GL calls (buffers-based). This is probably fine to
have for the first time, but making proper API migration is also somewhat
we'll need to do, IMO.

The porting has already been done.  I don't think you really want me to
> inline the OpenGL 3.2 code out of the replacement code.  My replacement
> might not have the best design, but I believe it's existence is the correct
> decision and can be improved as time goes on.
>

No, i don't want you to inline GL commands. But i'm not really considering
the immediate-mode-like-api something which is really ultimately finished.
It's for sure something like a must-have milestone, but IMO ideally Blender
should stop using immediate-mode-like api. Not sue in the scope of this
summer or in more long-term.


> A side note, I wouldn't use the word "fail" to describe my work.  I'm
> guessing you mean I've failed to contribute anything to master.  I'll agree
> with that, and I'm focused on rectifying that.
>

Yes, exactly, fail as in nothing was merged into the master yet.

-- 
With best regards, Sergey Sharybin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/soc-2014-dev/attachments/20140601/ecfe8f32/attachment.htm