mozilla.org/nss

[and-ri]

closes #279

[and-ri]

I'm having trouble building this for darwin/aarch64.
I managed to make some progress using llvm clang but still run into the error aes-armv8.c:14:2: error: "Compiler option is invalid"
My mac architecture is x86-64 and I can't reproduce the error locally.
Perhaps you have some ideas?

[jhheider]

Something makes me this -march=armv8-a -mfpu=crypto-neon-fp-armv8 isn't right.

For gnutls we had to do this: https://github.com/teaxyz/pantry/blob/577ffe507465e5b26411c39c951177e5e0b2c382/projects/gnutls.org/package.yml#L28-L34

Possibly force-settings arch/cpu/fpu options is the right move. Interesting that it doesn't build out of the box on aarch64, since Apple doesn't make any x86_64 devices any more.

[jhheider]

Homebrew's recipe is against 3.87. Maybe see if that builds? This appears to be a mirror of it.

If that works, the release notes might be instructive for building later releases (versions: can be an array).

[jhheider]

homebrew's comments make me think that firefox did some seriously non-standard stuff here. might have to branch out for other formulae for clues.

[and-ri]

@jhheider maybe we can package it for linux only for now? This commit: a211990

chromedriver needs nss on linux

what do you think?

[jhheider]

Releasing for Darwin for now with a comment seems fine.