Raph Levien

I should do more research on Windows 7, but at the absolute worst, hard-coding a list of system fonts along with metadata (so it knows to put Meiryo at the...

Research on Windows 7 below (it was bothering me, obviously). Qt has a [hardcoded list of fonts](https://github.com/qt/qtbase/blob/5.12/src/platformsupport/fontdatabases/windows/qwindowsfontdatabase.cpp#L1686) and includes the Han unification logic in there (there's a separate font list...

Looking more closely at DirectWrite, I think another possibility for Win7+ is to create a [TextLayout](https://docs.microsoft.com/en-us/windows/desktop/api/dwrite/nn-dwrite-idwritetextlayout) object, then call pass a custom [TextRenderer](https://docs.microsoft.com/en-us/windows/desktop/api/dwrite/nn-dwrite-idwritetextrenderer) to the [Draw](https://docs.microsoft.com/en-us/windows/desktop/api/dwrite/nf-dwrite-idwritetextlayout-draw) method, gathering the font...

@jfkthame This seems likely to me, yes. The CJK stuff would have already been covered by pref, so this seems most effective in the long tail fonts like [Ebrima](https://docs.microsoft.com/en-us/typography/font-list/ebrima). I...

Thanks for the input. And to be clear, I don't consider this an issue of optimization (though that would be nice) so much as avoiding footguns from either memory leaks...

Another data point: if I replace the do-while with a regular while loop, naga translates the file without complaint and the output looks reasonable from a quick eyeball scan, though...

Thanks for the response. I've started prototyping on my end, but still not sure whether it's better to push that to release or work here. The problem with `std::sync::mpsc` is...

I released my prototype: https://github.com/google/synthesizer-io. I'm not sure how much I'm going to continue developing it, whether it's primarily useful as a source of ideas or as a real codebase....

The original motivation is mostly that they types they were replacing (in fluent-locale-rs) was `Option`. And tags in locales can never be empty strings. So the idea was that by...

This is likely to be an issue in piet-gpu as well. I'm linking a [gist](https://gist.github.com/raphlinus/1406c2a41589ea1a616cfc2552d6dc1c) of a mostly automated translation of [coarse.comp](https://github.com/linebender/piet-gpu/blob/947a85f9cfb97f946b9bbd32ca9c2c1e5246b482/piet-gpu/shader/coarse.comp), not intended to be correct but intended to...