Dan Kolsoi
Or maybe the name is valid for certain orderings only?
Travis is basically toast now for OSS, so I've started working on a GH actions build which should get us OSx and Windows coverage in the long run: https://github.com/TheDan64/inkwell/tree/new_ci
That's odd. What about `position_at_end`? https://thedan64.github.io/inkwell/inkwell/builder/struct.Builder.html#method.position_at_end
That's the one I usually write tests with - and sounds like it does what `SetInsertPoint` does - but I'll have to look into `position_at`
@djozis Does my feedback make sense? Is there anything I can expand upon?
> I wanted my use of LLVM to be as efficient as possible Panicking isn't any more efficient because it still has to make a check and you could achieve...
Static typing would be ideal, but I don't think there's enough Rust support for a sufficient amount of it yet
> The efficiency I was referring to would come from keeping the check to debug mode only, as I was thinking it's a code path you'd encounter in a static...
Is it because you provided `null_terminated` as `true` in your rust version, when it is actually not? ```rust let const_string = ctx.const_string(b"test", true); ``` I suppose we could check for...
We don't currently support using Inkwell types across multiple threads as it would be a pretty big undertaking to get right. Thanks for pointing it out, though. https://github.com/TheDan64/inkwell/issues/28 is related...