Antony Lee

@tfpf I think best would be to perhaps not encode the entire 64 internoad spaces for now (unless you want to go for it) and just specifically encode an inter-fraction...

> Does this mean: add the subscript and superscript changes here? This would result in one less change for some images, because there are some tests which have fractions as...

Sorry, I slightly messed up the "svg lightweight tests" change. Can you additionally include the two lines at https://github.com/matplotlib/matplotlib/pull/23270/files#diff-17f9692b77108a62b59c71825d8461488729ef280f56785d3ad55ce3a86044d8R210 (as that doesn't look like it's getting merged any time soon)...

My (very not guaranteed...) understanding is the following: Looking at the hlists passed to ship(), the old version was (after stripping out Hlist sizes) ``` (A) H[H[], H[H[H[`s`, k1.07, `i`,...

Previously, the two tests ```python r'$xyz^kx_kx^py^{p-2} d_i^jb_jc_kd x^j_i E^0 E^0_u$', # github issue #4873 r'${xyz}^k{x}_{k}{x}^{p}{y}^{p-2} {d}_{i}^{j}{b}_{j}{c}_{k}{d} {x}^{j}_{i}{E}^{0}{E}^0_u$', ``` rendered the same. With this PR, some of the exponents are shifted...

Oops, thanks for trying. Hopefully fixed now?

FWIW, simple handling of rst to html conversion can be implemented as in the `_parse_doc` and `_itertext_ansi` functions at https://github.com/evanunderscore/defopt/blob/master/defopt.py#L285 and https://github.com/evanunderscore/defopt/blob/master/defopt.py#L342.

Fair enough, feel free to close if you're not going to implement the requested feature then (I agree 2>/dev/null may be enough).

I renamed USE_CBLAS to USE_BLAS throughout. > do all cblas implementations also propose a FORTRAN API? I'll have to check that :(

Based on https://github.com/Kitware/CMake/blob/6c01bbe180df8771fa1e93776f0b716ac146218f/Modules/FindBLAS.cmake#L344 (where `_name` has typically been set to something like sgemm), I *believe* that cmake's FindBLAS.cmake ensures the presence of `segmm_`, i.e. a FORTRAN API. (The only exception...