02:51 <emekoi> do we have computed goto's in zig?

02:54 <andrewrk> emekoi, no

02:55 <andrewrk> will consider adding them if we can't get switch statements optimizing into equivalent performance

03:50 <wilsonk> andrewrk: is there a way to output --verbose-ir only for the file I am compiling and not the builtins/std functions? I know in D there was a way to exclude the stdlib

03:59 <wilsonk> I guess that last question is open for anyone! :)

04:12 <andrewrk> wilsonk, no, there's not. here's a clue though: https://gist.github.com/andrewrk/5684434a2a8d4cbb08bb0d855c4f2ada

04:21 <wilsonk> exporting and overriding the panic handler works very well...thanks! :)

04:25 <wilsonk> I get 4.1k output with both tricks, 219k if I just import debug.warn (and use it, of course) and 4.7M if I don't use the panic override!! Wow

04:33 <andrewrk> wilsonk, what about in --release-fast or --release-small mode?

04:34 <andrewrk> a large portion of the size is stack traces

04:34 <andrewrk> which I believe are not included in release fast / release small modes

04:37 <wilsonk> 3.8k, 201k and 4.3M for release-fast and release-small

04:38 <andrewrk> which one is which?

04:48 <daurnimator> I assume that is for "both tricks" vs "debug.warn" vs "no panic override"

04:51 <wilsonk> yes daurnimator has it correct...the sizes don't change between fast/small modes strangely

04:55 <andrewrk> that's not surprising - but it would be surprising if the size didn't change between debug and release (non-safe)