hash mismatch.

I was wondering-- our use of outputHash style here seems .. fragile?

Not entirely surprised the hash is different, the two main dependencies here are fontforge and fonttools and I'm on much newer fontforge and slightly newer fonttools (former is chasing git, latter I will make a PR for soon oops).

Fixing the hash is easy enough, but that doesn't seem like the right answer.
If we're fetching the fonts maybe we should just grab the prebuild files instead?

Previously: #51331 and its successor #51956

cc @worldofpeace as involved with those (along with a certain @dtzWill character... :))

I was wondering-- our use of outputHash style here seems .. fragile?

Not entirely surprised the hash is different, the two main dependencies here are fontforge and fonttools and I'm on much newer fontforge and slightly newer fonttools (former is chasing git, latter I will make a PR for soon oops).

Fixing the hash is easy enough, but that doesn't seem like the right answer.
If we're fetching the fonts maybe we should just grab the prebuild files instead?

Previously: #51331 and its successor #51956

cc @worldofpeace as involved with those (along with a certain @dtzWill character... :))

I feel like we should prefer a source build, but if there's limitations that we can't work around easily it's not worth it.

(good to see that certain character back with an update btw 😄 )

I think having a source build is a generally preferred situation; committing to reproducibility via outputHash is a relatively bold move, although of course if reproducibility can be ensured it is nice.

But I don't keep up with making Fontforge output reproducible.

Latest release is 2.1.0. Will you update the PR?

Let's close it since it's very out of date.

@prusnak could you open an update request issue?

PR in #86343