Adrian Sampson

Yes, forcing the linking is unfortunate! I couldn't find a way around it, though, to install both packages under the same prefix. If someone better with the Homebrew API is...

Hello! As the README says, this was written for LLVM 3.8. (Any help with the port would be appreciated.)

OK! I’d accept a pull request if you can offer one. :+1:

Aha, got it! In fact, [the post is on GitHub too](https://github.com/sampsyo/home/blob/master/_posts/2015-08-03-llvm.md). :smiley: So if you have a moment, please do consider sending a PR there.

Ah, that’s tricky—thanks for pointing that out. I’m not currently using this recipe, but I’d gladly accept a pull request if you know how to fix this.

In the mean time, at least, it would probably better to emit [ordinary commit statuses](https://developer.github.com/v3/repos/statuses/), which should be very simple and actually useful (since there's a UI).

Also, just because it came up on HN today, here's a nice explanation of a [somewhat bonkers optimized fixed-length reduction](https://software.intel.com/content/www/us/en/develop/blogs/can-you-write-a-vectorized-reduction-operation.html) that's worth pondering.

Interesting discussion; thanks for the ping! This is a good list from @lemaitre. To break things down a bit, I sort of think there are two competing goals here: -...

FWIW, I think CPU fingerprinting is inevitable, even in fixed-length SIMD (https://github.com/WebAssembly/simd). I agree that's not great, but maybe we should try to focus here on potential problems relative to...

Yes indeed—thanks for opening the issue. We should switch sooner rather than later. (It's too bad about these deprecations, to be honest! I know they were rarely used, but they...