[tuhopuu-devel] missing uv-edit propedit code

Brecht Van Lommel tuhopuu-devel@blender.org
19 Feb 2004 22:01:02 +0000


Hi,

On Thu, 2004-02-19 at 20:27, BjornMose wrote:
> hi brecht, hi all,
> the feature list and your post a elysiun says there was a proportional
> editing in UV edit
> but CVS mailing list does not show any commit neither tuhopuu checkout gives
> any code?
> is this a CVS server hickup or didn't you commit ?

Viewcvs seems to know it:

http://projects.blender.org/viewcvs/viewcvs.cgi/tuhopuu2/source/blender/src/editsima.c?rev=1.9&cvsroot=tuhopuu&content-type=text/vnd.viewcvs-markup

Did you maybe change some code? I don't think cvs will auto-merge the
changes when doing an update. Tuhopuu-cvs mailing list messages always
arrive late here btw.

> before you start face select, did you look at my new "active face" stuff?

I didn't start it, I didn't even know if I was going to do it or not.
Wasn't sure if you were working on it. I did check out the "active face"
stuff. Works nicely. Only criticism would be the drawing of an active
face in the UV editor. Seems to be a little inconsistent with the rest
of Blender with those thick lines. Also, the black white stippled lines
in the UV editor give strange effects here. If you scale/rotate/move
some texture vertices it's like there are crawling ants on my screen :).
Wasn't the case with the old code.

Anyway, I'm now working on finishing the undo, and I'll probably attempt
to do those transparent faces in both 3d view and the UV editor as in
William Reynish's proposal: http://www.shadeless.dk/ui/uveditor.htm. UV
editor should be trivial, and the 3d view is also possible because it
works with Draw Faces and Solid viewport shading in editmode.

> btw
> i did look at the "cow flatten etc.." papers:
> "BIG FISH !" but i think we can make it.
> there are more than one math issues to tackle.
> i am going to provide brief overview the bext days.

Yes, the algorithms are complicated, but wouldn't a full automatic cow
flattening implementation be awesome? We'll have an example
implementation, so if we're stuck we can always look at that. Bruno Lévy
offered help, so I think he'll answer our questions if we ask nicely :).
Too bad the ABF++/ABF code will not be available until May. ABF and the
algorithms floating around it are my personal favorite, but we'll need
to compare the algorithms further. Luckily we can use Graphite for that.

There are essentially 4 unwrapping steps:
- seam cutting
- unfolding
- parametrization optimization (relaxing)
- texture atlas packing

Each step on it's own is useful to the user, so we don't have to have
them all immediately. I find them all really interesting though.

Cheers,
Brecht