jeapostrophe
2017-4-27 12:53:35

greg - Not a frog user, but I think it’s the Racket way


cowbs
2017-4-27 17:48:57

Is there a way to have DrRacket use the —compiled option?


danl-ndi
2017-4-27 17:54:04

alternately, is there way to run DrRacket via plain Racket.exe ?


samth
2017-4-27 17:58:44

@danl-ndi if you just run racket, and then do (require drracket) at the repl, that will start dracket


samth
2017-4-27 17:58:51

alternatively, racket -l drracket


danl-ndi
2017-4-27 18:00:18

huh… that will help quite a bit, thank you


danl-ndi
2017-4-27 19:18:42

what is the racket code to reproduce this -W command line option: racket -W “debug@setup/parallel-build error” ?


mflatt
2017-4-27 19:26:02

@danl-ndi You can set up a new receiver with make-log-receiver and start a thread to write any received messages to (current-error-port). There’s no way to get exactly the same effect as -W, though, which configures an initial log receiver.


danl-ndi
2017-4-27 19:31:33

is there an example of using a log receiver to get those messages?


danl-ndi
2017-4-27 19:32:18

I assume i’ll start with (make-log-receiver logger ’info ’setup/parallel-build)


danl-ndi
2017-4-27 19:33:53

mflatt
2017-4-27 19:47:51

Right


dan
2017-4-27 22:32:21

@asumu @samth The add-annotations macro in the implementation of typed classes disarms the syntax object it receives, should it also be calling syntax-rearm on the syntax objects it returns?


asumu
2017-4-27 22:38:48

@dan probably, but I’m not sure. I don’t remember explicitly not rearming it for any reason.


dan
2017-4-27 22:39:51

On a second look, it doesn’t look like anything in typed racket ever rearms syntax


lexi.lambda
2017-4-27 22:53:21

@dan: The new “Remove Unused Requires” feature is really neat. However, it doesn’t work properly for things like multi-in due to the way it expands to require specs. Is there some way that either multi-in or the DrRacket refactoring could be changed to accommodate?


dan
2017-4-27 22:55:50

@lexi.lambda: probably, I’ll try to look into it later tonight/tomorrow


lexi.lambda
2017-4-27 22:56:48

Cool, thanks! My bet is that racket-mode doesn’t handle it properly, either, since it seems pretty tricky to solve in general, but I haven’t tried it. Maybe @greg would know. :)


greg
2017-4-27 23:02:02

Modulo reformatting stuff, racket-mode is a pretty thin wrapper around macro-debugger/analysis/check-requires


greg
2017-4-27 23:02:25

remembers Slack doesn’t handle markdown style links :disappointed:


greg
2017-4-27 23:03:47

greg
2017-4-27 23:04:41

I started to look at the code back then and discovered it was way above my pay-grade, then


lexi.lambda
2017-4-27 23:12:40

I would try it myself to see what it does, but I don’t use emacs, and I definitely don’t have it set up on this machine.


dan
2017-4-27 23:14:51

I think I need to look more into what the macro-debugger does to see if check-syntax can use that


dan
2017-4-27 23:15:17

@lexi.lambda I filed an issue for the multi-in problem, thanks for pointing it out



lexi.lambda
2017-4-27 23:17:12

Require transformers are generally tricky. There’s a bug of sorts I’ve tried to fix twice before, which is that require transformers don’t leave behind 'origin properties when nested. As far as I can tell, fixing this is impossible with the current API for require transformers.


lexi.lambda
2017-4-27 23:18:20

For example, this means that in (require (combine-in (combine-in))), DrRacket understands the use of the first combine-in but not the second.