Antony Lee

> I would want to give the current state of the API (what is used in 3.10) an ID. - I don't really care on the exact spelling 1.0 or...

Actually the warning shows up even if the figure is shown (e.g. if you type @timhoffm's lines one after the other at an IPython prompt). AFAICT the root cause is...

@ayshih I think the plots you produce are quite convincing to show that this fix is reasonable, but I still have trouble understanding what's happening here and keeping track of...

Thank you for the super-detailed writeup! > That function's purpose – which admittedly took me some time to understand! – is to adjust the _integer_ array `m_weight_array` such that each...

> I'll point out that this PR does not change the normalization approach itself, but rather only deletes the code for post-normalization mirroring that uses the incorrect pivot. Your alternative...

re: normalization vs. symmetry: I guess this depends on what is your exact domain of application, but I can see why one would care more about exact normalization. I suspect...

I likely won't be able to review properly the changes since Oct. 8 before the holiday break. Sorry about that, this looks like great work.

In the usetex case, the origin of the shift is, I believe, the following. Remember that usetex+agg is currently implemented by compiling the tex string to a dvi, converting that...

This is because alpha is actually controlled separately by `grid(..., alpha=)`. I agree that this behavior is not optimal: probably the default grid_alpha should be None (in which case it...

As a side point, I find it rather surprising that mathtext respects rcParams["font.weight"] at all (how is that even supposed to interact with rcParams["mathtext.default"]?) Looking at the implementation I'm not...