David Nadlinger

> But, there are .di files...? `.di` is just a convention for the name of files which have been stripped of function bodies where possible. Those implementations still exist elsewhere....

> It will be parsed by each compilation unit either way, and not emitting the code (unless it's referenced) is logically less work, and should be faster. Agreed – and...

I also bumped into this as a first-time user of `zxlive`; lowering 2-ary H boxes sounds like a sensible pass to implement (and I agree, a warning for "invalid" inputs...

Which use case are you most concerned about, performance-wise (that isn't covered by the nthreads() fast path)? I don't see an obvious way to avoid locking altogether, other than avoiding...

I've honestly not used ProgressMeter.jl enough yet to be able to judge what the correct design tradeoffs are. One option to lower the overhead would be to use lock-free atomics...

I'm not a lawyer either, but as long as all contributors approve a license change, you can re-issue the code under whatever license you want. Of course, you cannot retroactively...

@ViViDboarder: I am not a lawyer, but… Yes, you can, if you _own the rights_ to the code (so you can't just »grab any GPL code« and do that). You...

@ViViDboarder: Which part exactly of »but as long as all contributors approve a license change« above was unclear?

Just noticed that this is still open. You probably want to merge this at some point to avoid accidentally making elements with elaborate conversions `@trusted`. The existing documentation was also...

Hmm – isn't mecca's fibre implementation is Posix/x86_64 only anyway, where the floating point registers are caller-save in the SysV ABI anyway? (`core.thread.Fibre`s don't save XMM registers either, for the...