On Sat, Jun 5, 2010 at 8:34 PM, Dan Eicher <[email protected]> wrote:
> On Sat, Jun 5, 2010 at 10:28 AM, Campbell Barton <[email protected]>wrote:
>
>> Mentioned some of the problems we have with the API as I see it in a
>> mail a while back, not many things have changed since then.
>>
> A lot of that stuff just needs someone to make a decision.
yep, but even to list changes, give some justifications, allow for
some discussion - it takes time/energy.

IMHO you dont have to look that far to find issues with the existing
rna/api, anyone who's interested can evaluate some area, list the most
obvious/simple problems
Mail the list with suggested amendments, agree on what to change, then
write up a patch and submit it.

> Things like "all add functions will be called foo_add()" or "all collections
> will override add() and remove() instead of using new() and delete()". Could
> also have controlled breakage this way, all add_foo() calls (and the bundled
> scripts that use them) could be fixed on one commit.
Editing bundled scripts to update with the changes is not so hard,
sometimes I batch process this, another reason I like having addons
within our svn repo.

> Then someone writing a script will know that if foo.collection.add() doesn't
> work it's a bug that needs to be reported and worked around. Also help with
> finding the shortcomings in makesrna -- for example nodetree.nodes.new()
> won't work because it needs a pointer to the nodetree but there's no flag
> like USE_REPORTS that can pass it to the new() function.
fairly sure USE_REPORTS could be enabled for this.

> Not that I don't appreciate how busy y'all are these days.
There's always time but I prefer to work on this stuff back home with
less distraction.



-- 
- Campbell
_______________________________________________
Bf-committers mailing list
[email protected]
http://lists.blender.org/mailman/listinfo/bf-committers

Reply via email to