Shubhamkar Ayare

Results 239 comments of Shubhamkar Ayare

Would you be adding better compile-time reporting any time? The current reporting I know if includes things like: - the compiler macro being incomplete and therefore not being able to...

I could imagine there being another parameter `values` to `cons-specifier-ctype` that is either of type `null` or `t`. But I don't feel having a separate `specifier-ctype` and `values-specifier-ctype` is a...

Okay, TIL about `sb-kernel:values-specifier-type`, and yes given the following behavior, things make sense: ```lisp CL-USER> (sb-kernel:values-specifier-type '(values &rest t)) # CL-USER> (sb-kernel:specifier-type '(values &rest t)) ; Evaluation aborted on #....

> Did you try building with [:coalton-release](https://github.com/coalton-lang/coalton/blob/main/docs/coalton-lisp-interop.md#interaction-mode) on your `*features*`? No, I hadn't - thanks for the pointer! Although, even with this, I don't see much performance improvement. So, that...

> The development/release mode divide isn't about type safety, it's about allowing both interactive development and performance. Oh, I see. Or, erm, no, I don't. I had mistook the structures...

> Inlining call sites recursively is a difficult problem. It's easy to write a compiler pass to do it, but figuring out heuristics for which functions should be inlined is...

As I understand, the issue above is that `+` itself gets inlined, but its instances do not get inlined. I am looking at the coalton code, and I haven't yet...

Thank you for the pre-review! I'm still thinking what the right thing to do would be. But yes, I definitely want to see inlining in coalton. One main concern (EDIT:...

Thanks for the extensive feedback! I will rework this over the next week or two and push the commits for further review. Yes, now, I'm in favour of attributes over...

> Wether a function is marked inline should be tracked in the [function-environment](https://github.com/coalton-lang/coalton/blob/main/src/typechecker/environment.lisp#L623-L625) instead storing a flag on each ast node. The function environment is set for functions [here](https://github.com/coalton-lang/coalton/blob/main/src/typechecker/define.lisp#L230-L236). It...