[Bf-python] Re: tp_getset conversion update
Ken Hughes
khughes at pacific.edu
Thu Sep 29 19:36:53 CEST 2005
Stephen Swaney wrote:
> I am not claiming that the solution is poorly docuemnted or that it
> is not cleverly coded. My position is that it is misleading.
> Doing a dir() on World will show 4 attributes: mistype, qua, lin
> and sqr. That does not coorespond to the actual data model, either
> internally or from the user's perspective.
As I said earlier, my goal was to maintain backwards compatibility AND
to make the API resemble the UI. I'd happily vote for removing
"mistype" and making script writers rewrite everything, but I know
better :-)
> The fundamental fact here is that qua, lin and sqa are not attributes.
> They are names for the values an attribute can take.
OK, consider this:
A user does a dir on a texture. They will see:
'imageFlags', .....'interpol', 'useAlpha',....
'imageFlags' is the entire bitfield variable, while 'interpol',
'useAlpha' and others are bits within the same bitfield.
Now either:
(1) I slipped one by you in my previous tp_getset commit and we
should have had this discussion earlier
(2) You're not being consistent, or
(3) We disagree on what "value an attribute can take" means.
???
Ken
More information about the Bf-python
mailing list