[GUI] Fix issue with triangles not showing up when they are drawn clockwisely

[Rullec]

Hi all,

One of our classmate in GAMES201 from Wechat found that ONLY when [va, vb, vc] is in counterclockwise order can ti.gui.triangle(va, vb, vc) manage to display it. If the offered vertices are in clockwise, our ti.GUI simply ignore it.

For example, if I say

    triangle=[(0.3,0.5),(0.4,0.7),(0.5,0.55)]
    gui.triangle(*triangle,color=0xff0000)

This triangle won't be drawn.

If it is a designed feature, maybe we can try to doc it?
If not, here is one possible solution.

HDYT? Maybe we can have a better solution.

[Rullec]

@Eydcao proposed that, it can also be implented by barycentric coordinate. Also a good solution!

[codecov]

Codecov Report

Merging #1321 into master will decrease coverage by 0.40%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #1321      +/-   ##
==========================================
- Coverage   85.76%   85.35%   -0.41%     
==========================================
  Files          18       19       +1     
  Lines        3294     3374      +80     
  Branches      624      630       +6     
==========================================
+ Hits         2825     2880      +55     
- Misses        343      362      +19     
- Partials      126      132       +6     

Impacted Files	Coverage Δ
python/taichi/lang/impl.py	90.06% <0.00%> (-0.14%)	⬇️
python/taichi/lang/shell.py	40.67% <0.00%> (ø)
python/taichi/lang/kernel.py	80.82% <0.00%> (+0.04%)	⬆️
python/taichi/lang/__init__.py	75.35% <0.00%> (+0.35%)	⬆️
python/taichi/lang/ast_checker.py	72.22% <0.00%> (+0.39%)	⬆️
python/taichi/lang/matrix.py	90.67% <0.00%> (+0.50%)	⬆️
python/taichi/lang/expr.py	89.75% <0.00%> (+1.69%)	⬆️
python/taichi/lang/snode.py	86.02% <0.00%> (+8.68%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d9c9616...09d10b0. Read the comment docs.

[Eydcao]

@Eydcao proposed that, it can also be implented by barycentric coordinate. Also a good solution!

Thanks~ I was in a meeting just now. I have reviewed your code and find the solution good enough. It turns out it does not matter whether the cross product is positive or negative (or say the particle is at the left side or the right side of the edge), as long as the it is at the same side for all edges, it can only be inside.

So this is good enough already. Good fix!

[Rullec]

Thanks! You gave such a clear explanation. So let's see whether there is something more we can improve on in this issue... Yadi Cao <[email protected]> 于2020年6月25日周四 上午12:18写道：

…

@Eydcao <https://github.com/Eydcao> proposed that, it can also be implented by barycentric coordinate. Also a good solution! Thanks~ I was in a meeting just now. I have reviewed your code and find the solution good enough. It turns out it does not matter whether the cross product is positive or negative (or say the particle is at the left side or the right side of the edge), as long as the it is *at the same side* for all edges, it can only be inside. So this is good enough already. Good fix! — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#1321 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEAYDGTF5VWPDR5L6VU6GDTRYIRNXANCNFSM4OGSEB2Q> .

[yuanming-hu]

Thank you! Originally I designed the API to ignore clock-wise triangles, but since it is confusing the users, let's just take your proposal.