Fix issue with bytes_to_string in library.

[cristianoc]

Fixes #5573

[cristianoc]

@TheSpyder take a look at this fix -- does it work for you?

[cristianoc]

The situation of bytes_of_string seems even worse in that it looks like it's not defined anywhere. Nor is seems to be used anywhere. Probably better to just remove it.

[TheSpyder]

That looks like it'll work, yes. The Char library file no longer references a method that doesn't exist :)

[TheSpyder]

I wasn't paying enough attention. RC1 has the same issue - the code I said replaced Caml_bytes isn't Bytes.bytes_to_string, it's Bytes.unsafe_to_string.
https://github.com/rescript-lang/rescript-compiler/blob/375cc68473d08d8357531b7cb48a91fec27a34f5/jscomp/stdlib-406/bytes.ml#L158

[cristianoc]

Would you create a PR?

[TheSpyder]

I'm trying, but figuring out how to run a locally built compiler in my project is turning into a hassle. I got it working, but after changing the reference and rebuilding it was completely broken 🤷‍♂️ I guess I could still make the PR, the code changes are correct now.

[cristianoc]

I'm trying, but figuring out how to run a locally built compiler in my project is turning into a hassle. I got it working, but after changing the reference and rebuilding it was completely broken 🤷‍♂️ I guess I could still make the PR, the code changes are correct now.

Thank you.
Did you figure out what the issue with building the compiler was? Any suggestions to improve the contributors guide?

(there's a makefile too though it's ongoing. For your case: "make clean && make test" should do.
Plus, installing the locally built compiler into a project)

[TheSpyder]

Ah I was just following the contribution guide and didn’t notice the Makefile. Thanks for the tip, I’ll see if that helps later this afternoon.

[TheSpyder]

Didn't help. I'm at a loss as to why sometimes the local build works in my project and other times it fails with a syntax error.

[cristianoc]

If it can "help", the build is no deterministic. Still haven't figured out where but definitely there are race conditions.
To be confirmed if telling ninja "-j 1" solves it.

[TheSpyder]

The build completes successfully, it's the compiler executable it builds that is behaving differently. It's acting like it has a different version of the syntax that isn't quite right.

[cristianoc]

Is the same executable behaving differently or is there a recompile in between?

[TheSpyder]

A recompile. On two macs it worked once and then I switched to my branch and never again, even after switching back to master and cleaning everything.