[Bf-committers] BMOPs and multiple object editing

Campbell Barton ideasman42 at gmail.com
Sat Sep 8 09:01:38 CEST 2018


Short term, would do what ever seems most straightforward.
Longer term would go with 'C' since I don't think this operation
benefits much from being a bmesh-operator and this is currently using
inefficient logic.
See: https://developer.blender.org/D3674#84136 for details.
On Tue, Sep 4, 2018 at 10:47 PM Dalai Felinto <dfelinto at gmail.com> wrote:
>
> Hi,
>
> So far we were able to keep any multiple-object editing change away from
> the bmesh operators. However for "select similar" (D3674) I would like to
> check on which design to pursuit.
>
> We are taking into account all the selected elements of all the objects in
> edit mode. So in a way, a bmesh can't be processed entirely isolated from
> the rest. A few options:
>
> (A) Loop over all the objects in similar_vert_select_exec, populate an
> array with all the selected data (normals, directions, ...), pass this list
> to the BMOPs (as void *).
>
> (B) Handle all the loops as part of BMOPs, not passing any particular bmesh
> to it, but instead an array of all the bmeshes to interact among themselves.
>
> (C) Move the entire "select similar" from bmo_similar to editmesh_*.
>
> (D) Take only the active object as reference to what to select from, pass
> the active object's BMesh to all the other objects, to iterate over it.
>
> I find (A) quite sensible, and it is what I started to do in the
> forementioned patch.
>
> Thanks,
> Dalai
> --
> blendernetwork.org/dalai-felinto
> www.dalaifelinto.com
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at blender.org
> https://lists.blender.org/mailman/listinfo/bf-committers



-- 
- Campbell


More information about the Bf-committers mailing list