I think this question would be a great fit for the <https://racket.discourse.group/|Racket Discourse> ! :slightly_smiling_face: I hope to see answers there that will also help other Racket newcomers in the future. (If I can, I’ll write something myself.) On Slack, these answers get lost easily, despite the archive. :cry:

Since we have the Discourse instance, I use Slack only for simple questions or when I hope to get a quick answer, e.g. which function/macros may do something I want to do.

@bsilverstrim A very important question, by the way! :+1:

I gave it a try in a Windows 8 VM. Racket.exe crashes after printing the version prompt. Here’s a stack trace: https://pastebin.com/raw/Sj6ETLyc

I think the short answer is that it works basically the same as you describe for Go. You have some functions, they pass data around, sometimes you create threads and pass data to them over channels.

The biggest difference from your description is that Racket uses exceptions instead of extra return values for errors

Cc @alexknauth

I don’t think it can be related to Alex’s changes. I just tried creating a new snapshot locally, and it worked ok — but that’s a different machine, different MinGW version, and starting from a different build state. I think it’s most likely something going wrong with incremental builds, and I’ll keep investigating.

Maybe it’s related to equal-always? changes, after all. I had mixed up CS and BC in my new snapshot build, and a new BC build doesn’t run.

But a new BC build starting from scratch does run, so I’m back to thinking that it’s something wrong with incremental builds (like a missing dependency in the build script)

@sschwarzer Thanks for the feedback! I can copy and paste a copy of the question to the Discourse and see what others think there. I’ll get that done today. Also, thank you @samth for the information!

Hm

I think I figured out the missing dependency, and we’ll see whether the next snapshot works.