David Nadlinger

@ChrisRackauckas: Your points are valid, but changing the API is different is a whole separate story from the transparent optimisations I did last year.

This has been a long time ago, so I'm a bit hazy on the details. IIRC you should see a massive reduction in the number of GC allocations at least...

I just had a quick glance over the benchmarks, and it seems like you don't have any for the precomputed case? It's been a year since I looked at the...

Ah, sure – just wanted to make sure it was clear enough in which scenario the "several times speedup" remark from my original post applies.

@WalterBright: So what is the point of the release branch then? Holding back on all merges for an indefinite amount of time can be an incredible demotivator for people working...

Just a debugging trick which might or might not be applicable here: Did you try `dump()`-ing the LLVM module in the debugger when the assertion is hit? You will see...

> (Note: I'm not sure what LDC's cmdline arg for -S is, but you get the idea) It's `-output-s`. > expect binary contains main() and lib.fun() (assuming unoptimised; no inline...

> C++ absolutely implements the semantics I describe. That's the entire point – C++'s and D's semantics as designed are different!

> And this is the only definition for inline which I find useful... Really? https://forum.dlang.org/thread/[email protected] > This concept is firmly established in other languages, it's called inline Language, maybe, singular...

> How can it not be conservative and emit the symbols where they're needed, just in case? Does it have knowledge of the contents of libPhobos? While doing the necessary...