I got Resyntax working in github code reviews!

@laurent.orseau I’d really like to figure out how to fix the problems with quickscript-test.rkt on DrDr. It’s of course possible to disable them, but that seems less good. Have you had a chance to look at it at all?

Yes, I soon as I have time I’ll try a fix. I had a discussion with Robby (in #drracket) and it seems the problem could be. I was just either pretty busy or away these past 2 weeks.

My remaining concern is that I’ve never observed this bug on my machine, and I don’t see how to reproduce it. So that may be a blind fix unfortunately.

I just ran it and it failed first-time

on your machine or on DrDr?

my machine

Ok, I hope I can reproduce it somehow then. I should have time tomorrow, hopefully.

I ran it 5 times, failed 4 of them.

huh, weird. I’ve ran it countless times on my machine, it never failed

what OS are you on?

Linux

(finishing my sentence above:) It seems the problem could be that it doesn’t really wait for the tab to be ready, thus it could write in the wrong place.

> Note that “0.5.0” isn’t technically a valid package version (see Package Concepts). I wouldn’t have thought that using <https://semver.org|semantic versioning> , used by many, would be a problem. This feels really weird.

Personally, ideally I think that Racket should allow everything that semantic versioning allows, but at least version numbers of the form “x.y.z”, where x, y, and z are integers >= 0.

> (it’s probably also too late to enforce this in raco setup at this point since probably lots of packages (incl. many of my own) are non-compliant) Somehow I find this relieving. :smile:

That said, I can understand the preference for a canonical version number format and I guess Racket’s preferred format is older than Semantic Versioning?

By the way, as the name implies, Semantic Versioning not only defines valid version numbers, but also under which conditions they should change in which ways.

I don’t think Racket’s versioning scheme prevents you from applying [the subset of] semver [commonly used in practice] to your projects, though. 1.2 and 1.2.3 are both valid versions, it’s just that instead of 1.2.0, you’d write 1.2; so it doesn’t feel like a big deal to me. That said, I don’t see a problem with extending the scheme to allow a 0 "<sub>" when there’s no "<rel>", but I don’t know the exact reasons why it doesn’t presently. Calver is similarly compatible, in that you can have versions like 20220529.0 or 2022.5.29.