I’m trying to build my own ‘#lang’ and have the binding arrows in DrRacket work for it. I found this page (https://docs.racket-lang.org/tools/Check_Syntax.html) which seems relevant, but I need help understanding it. The example of sub-range-binders (http://pasterack.org/pastes/74231) worked as expected for me, but the other examples didn’t (http://pasterack.org/pastes/32894 and http://pasterack.org/pastes/14696). I hovered over variables and expressions, clicked on ‘Check Syntax’, but no arrows or tooltips ever showed up. What am I doing wrong?

@leafac For the disappear-use/binding example, id1 has no binding position and id2 has no use position, so check-syntax can’t draw any arrows. If I change the call to (m a a) I get an error between the two a’s. I’ll try to clarify that in the docs.

I’m not sure what’s up with the mouse-over-tooltips example, sorry.

@florence: Thank you for your answer. In fact, a slightly modified example including (m a a) (http://pasterack.org/pastes/81777) does show arrows. I don’t get it, though. I thought the disappeared-use, disappeared-binding syntax properties had the intention of forcing an arrow to appear between the two. Clearly, I was wrong, the literal identifier under the syntax object also influences the result. What am I missing? How do I learn about these things? All I can find is the reference, but I need a guide :slightly_smiling_face: Regarding tooltips, I don’t mind, because I’m actually interested in the arrows. But it would be helpful to understand tooltips as well, I guess.

@leafac check-syntax associates bindings and reference that are equal as identifiers (free-identifier=?), more generally you can have multiple disappeared-binding``` anddisappeared-use``` properties on a single piece of syntax, so just putting on of each on a piece of syntax doesn’t have the effect of associating those identifiers

annoyingly our CI fails when it can’t contact http://unicode.org\|unicode.org, which I think had some downtime recently

Hello! Is there a way I regenerate .zo files for a minimal installation? We have 6.8 pushed out through perforce but the zo’s are missing.

Sorry for the noob’ish question, just trying to get caught up to speed helping @danl-ndi out.

@cowbs I think you want raco setup

Looks like it, thanks!

Great!

@dan: That makes sense. Is this condition mentioned in the documentation and I missed it? Or is this a potential documentation contribution?

robby: would it be hard to get a build-derivations that just gives me one derivation? (I need to find SOME typing and then do induction on it.) I’m sure my system has unique derivations, but it would be nice to not worry about constructing more than needed if that changes

Seems like a simple function that discards stuff?

If there are performance concerns maybe send me a@model?

So I’m running into an issue in 6.8 with parallel-compile-files where it seems to be stomping itself trying to update certain .zo and .dep files. This only seems to happen when it thinks it needs to update zo’s in rackets own collects directory (usually because of perforce trickery).

Here’s an example of the error, which will go away on subsequent runs: 0: Error precompiling 'C:\perforce\main\shared\src\plt\collects.new\dc\script.rkt'! rename-file-or-directory: cannot rename file or directory source path: c:\ndibin\racket-6.8-64\collects\racket\contract\private\compiled\6.8\compiled\tmp14931653101493165319184 dest path: c:\ndibin\racket-6.8-64\collects\racket\contract\private\compiled\6.8\compiled\arr-util_rkt.dep system error: Access is denied.; errno=5 compilation context...: C:\perforce\main\shared\src\plt\collects.new\dc\script.rkt context...: c:\ndibin\racket-6.8-64\collects\racket\file.rkt:231:5 c:\ndibin\racket-6.8-64\collects\racket\file.rkt:199:0: call-with-atomic-output-file20 c:\ndibin\racket-6.8-64\collects\racket\private\more-scheme.rkt:148:2: call-with-break-parameterization c:\ndibin\racket-6.8-64\collects\racket\file.rkt:220:5 c:\ndibin\racket-6.8-64\collects\racket\file.rkt:199:0: call-with-atomic-output-file20 c:\ndibin\racket-6.8-64\collects\compiler\cm.rkt:363:0: compile-zo* c:\ndibin\racket-6.8-64\collects\racket\private\more-scheme.rkt:206:2: call-handled-body

I can try working around the perforce tricky business, but I wanted to raise it here as well because we didn’t have this issue in 6.5 as far as I know.

@leafac It doesn’t look like that’s mentioned in the docs from a quick look, so adding a bit of text explaining that would be a welcome improvement.