PR is here: https://github.com/racket/data/pull/14\|https://github.com/racket/data/pull/14

[solved] Since this morning, when I make, I get the following error: standard-module-name-resolver: collection not found for module path: (lib "setup") collection: "setup" in collection directories: /home/laurent/.racket/7.6.0.17/collects /home/laurent/build/racket/racket/collects context...: show-collection-err standard-module-name-resolver module-path-index-resolve [repeats 1 more time] module-declared? Something I forgot?

Great

Is the directory “/home/laurent/build/racket/racket/collects/setup” missing, or is Racket somehow having trouble finding it?

The directory exists

Oh, now I remember: I moved some the base directory down one level the other day, so I guess I need to remake all. There’s no make clean, what’s the preferred way to rebuild from scratch?

I use git clean -x -d -f .

This is a FAQ, so probably I should add a clean target that echos that suggestion.

Probably you really only need to rebuild the executable, so a make clean in the racket/src/build directory is another option.

make clean in src/build didn’t work, so back to git clean. Thanks!

(cleaning worked.)

Pro tip: don’t try echo "Trygit clean -x -d -f ." unless you’re ready to run git clean. :slightly_smiling_face:

haha :smile:

PR submitted: https://github.com/racket/racket/pull/3076

@greg what’s the best way to try your check-syntax branch?

@samth Probably M-x package-remove racket-mode, then add to Emacs load-path where you checked out that branch.

ah

I’m worried my emacs-fu isn’t up to that

If you use use-package, keep that but remove the :ensure t and replace with :load-path /path/to/wherever.

I’ve been on the fence, close to merging that branch to master, soon, anyway

For awhile, dogfooding I’d discover a problem every day or two. Less frequently now. Probably just need other people to try it.

Well also, last 10 days I was mentally distracted. But focused again, last few days.

I would definitely use it right now if it didn’t require writing elisp

OK well let me look at merging to master within a day or two, and you could get it the normal way from MELPA.

Some days ago, it was probably ready, I just wasn’t sure I was ready to deal with any onslaught of bug reports etc, then. OK now.

I would be happy with “paste this code into your .emacs”

Ah OK. Do you use use-package?

no, my .emacs has no mention of racket-mode

OK, then, let me double-check what the thing to add, would be ….

I think simply (add-to-list 'load-path "/path/to/racket-mode-repo")? i.e. Wherever you did git clone racket-mode and git checkout check-syntax, append that path string to the load-path list.

Also, M-x package-delete racket-mode so it stops looking for that one.

@samth ^

Ok, I’ll try that. It would be cool to have something where racket-mode maintained itself with a git repository that it could auto-update, for those of us who want to live on the bleeding edge

Yep. I mean, MELPA is already that, where master is the bleeding edge. (MELPA Stable is supposed to be for “stable” packages, but I don’t use that personally or supply racket-mode for that.)

There’s also something called straight.el, which is a pkg manager a bit more like Racket’s, in that it just gets things directly from e.g. git repos. But it’s new-ish and I don’t have any mileage, and it’s one more thing you’d have to install and learn.

> Some days ago, it was probably ready, I just wasn’t sure I was ready to deal with any onslaught of bug reports etc, then. OK now. FWIW, it’s been extremely stable for me over the last couple of weeks and I use it for at least an hour a day every day. The last issue that I had was the macroexpansion problem that you promptly fixed. The only other thing I can think of is it doesn’t seem to track generics well (i.e. it’ll say an import isn’t being used, when in reality it’s being used to define generics for a struct). I hit that a couple times but forgot to report it. Definitely not a blocker in my mind, though.

Here’s a minimal example that reproduces that issue:

#lang racket/base (require component) (struct foo () #:methods gen:component [(define (component-start a-foo) a-foo) (define (component-stop a-foo) a-foo)]) (you might need to raco pkg install component )

@popa.bogdanp Thanks for letting me know! Sounds like mostly good news. As for generics, I’ll take a look but that sounds 99% like something that would need to be fixed down in drracket/check-syntax. If so, I’m game to try to fix that, there, and make a PR. At the same time, that means it probably would not be a blocker to merging the check-syntax branch in Racket Mode.

Is there a way to have ERROR tests in rackunit report locations?