03:35 <andrewrk> hey everybody, an update on copy elision stuff. I'm about to push a commit where I backported all the non-copy-elision changes to master branch. And then I'm going to branch off master again, with a better design in mind. So this represents tossing out some work that I've been doing, but still gaining a few benefits from it

03:36 <andrewrk> before branching off master, however, I'm going to do other stuff. so it should feel a lot more productive around here for a few weeks until I start doing the final copy elision design

03:37 <andrewrk> the new design breaks copy elision into even smaller steps, so the code doesn't get messy like it started doing again, and only the smallest step is necessary to continue on with coroutine rewrite

03:37 <andrewrk> it's unfortunate that I had to put so much effort before realizing the final design, but at least now I understand all the problems and how to overcome them in a maintainable way

03:38 <andrewrk> I'll probably make a blog post to talk about some of this stuff. I'm overdue for giving an update

04:02 <emekankurumeh[m]> daurnimator: with `comptime_int`s and `comptime_float`s to runtime types mostly.

04:05 <emekankurumeh[m]> daurnimator: for example, zig can't infer that the type of the `comptime_int`s in the switch arms are supposed to be `i32`s.

04:19 <diltsman> Can anyone point me to how/where to add compiler_rt functions to Zig? I would be willing to put in a significant effort to add some of this functionality...especially for ARM, which I am developing for.

04:24 <emekankurumeh[m]> compiler_rt stuff goes in `std/std/special/compiler_rt`.

04:25 <emekankurumeh[m]> *`std/special/compiler_rt`

04:37 <diltsman> Any suggestions of where in the repository history I should use as a porting source? I would assume head.

05:13 <emekankurumeh[m]> porting what?

05:34 <diltsman> The different compiler_rt functions from LLVM to Zig.

05:34 <emekankurumeh[m]> i think HEAD would be a good place to start.

05:35 <emekankurumeh[m]> i also suggest trying to compile your project and just porting compiler_rt functions as the need arises.

08:27 <kristate> andrewrk: thanks for all of your hard work. Looking forward to the blog post and the brain dump. Src/Gen makes sense

20:01 <MajorLag> huh, apparently the order of operator precidence changed at some point... and actually I think it is a bug: `value = (value << 6) | bytes[3] & 0b00111111;` != `value = (value << 6) | (bytes[3] & 0b00111111);`

20:01 <MajorLag> but the docs list bitwise & as preceeding bitwise |

20:15 <MajorLag> see: #1902