Jim Mussared

> You re-wrote my PR so that I can no longer pass the polygon coordinates a list of lists/list of tuples 👎 > > I'm trying to be very responsive...

Just a quick note on the efficiency of the `array.array` approach, further to what @dpgeorge said in #8987. Because this uses `mp_get_buffer`, it also works with anything else that supports...

> As an outside contributor I am asked to sign off on submissions, so it seemed against the spirit of that policy to merge something different to what I signed...

@mbooth101 thanks for the repro! The problem is this line: ``` if (py1 != py2 && MIN(py1, py2) = row) { ``` which I changed from your PR, where it...

> in the meantime I will have a think about the bottom edge problem @mbooth101 I don't know if you came across [this page](http://alienryderflex.com/polygon_fill/) while writing this code, but it...

@mbooth101 Based on the idea that drawing the outline can solve this, I have implemented a more optimised fix for this. The idea is to detect the "lost" pixels during...

FWIW, this PR is now at +1204 bytes. > I don't know if this is of interest. The following script runs a "fuzz test" on polygon fill. Thanks @peterhinch !...

Thanks @peterhinch I think it's valuable to be able to support round rect, although we need to consider the trade off against doing it in Python. Here's a Python version...

> Ingenious, but odd, not least specifying a rectangle by its centre. As an example of the implications, in [micro-gui](https://github.com/peterhinch/micropython-micro-gui) a user can specify an on-screen Pushbutton object. This can...

OK, one final alternative to add to the mix...! This is basically identical to @peterhinch's round rect implementation, except it's implemented by adding an additional `radius` argument to `rect()` instead...