I see, thanks both.
https://slack.com/intl/en-gb/blog/news/pricing-and-plan-updates > We’re also simplifying subscription limits. Instead of a 10,000-message limit and 5 GB of storage, we are giving full access to the past 90 days of message history and file storage, so you’ll never have to guess when your team will hit your limit. The majority of our active free teams will have access to more of their message history with the new 90-day limit compared to the previous limit. No matter how much your free team uses Slack, you’ll always have access to 90 days of history.
The <https://racket-slack.herokuapp.com|slack signup link> is currently reporting an application error
The snip displayed in the REPL is a copy of the snip you pass in. This means that the snip needs to have a proper copy method, otherwise any values set on the initial snip will be lost. Also, once displayed in the REPL, you won’t be able to change the snip from the program. One of my map snip videos showed how to work around that.
At some point I need to change this behavior (but in a way that lets @alexharsanyi’s code continue to work). It is on my very long term todo list!
Ugh. I think @jsyeo set that up?
I noticed <https://twitter.com/fngeekery/status/1549333276350971904|this tweet> about an episode with Shriram on the <https://twitter.com/fngeekery|Functional Geekery> podcast. Good episode, then I found there are others with a number of Racket people, so I subscribed. Looking forward to the Rackety episodes :)
I gained some appreciation for the Rhombus effort via that episode. Hearing Shriram mention both his love for s-expressions and the difficulty of fighting resistance to them over decades was helpful (I know some of you have done that very thing, but sometimes it takes just one more data point :) ). If both can successfully flourish, that would be nice. But if for some unforeseen reason we discover they’re mutually exclusive, than naturally ( ) should win :)
how can I run raco test with errortrace?
Maybe: #lang errortrace racket ?
Suggestion from: https://github.com/racket/rackunit/issues/97
You can’t get away with using racket instead of raco?
Sorry
@badkins things like #lang, reader extensions like 2d syntax and syntax/parse macros give me hope that we are moving to a world where you freely use the right syntax for the job.
(Even if I haven’t got round to making a #lang of my own yet- I will make one this summer no matter how silly/simple it is)
Pleasantly surprised to hear that Elixir has real hygienic macros. IIUC adding modules syntax/parse and #lang would give us a racket equivalent on beam
A workaround I came up with is: racket -l errortrace -i -e "(require (submod <fn> test))" which annoyingly forced me to exit the repl every time. (-i is required in this case)
I might take a look at whether I can enable this feature for raco test
I think the problem with #lang errortrace racket is that errortrace, as far as I know, adds more code that slows down execution.
So when I have a module that contains both production code and tests, I either have to make the production code unnecessarily slow or don’t see stacktraces. At the moment, my workaround is to add errortrace to #lang while I’m developing most of the code and tests, and remove errortrace later. Of course that’s error-prone.
> I might take a look at whether I can enable this feature for raco test So I guess it would make sense (if that’s possible) to enable errortrace in test code only, without having to modify the #lang line.
@greg is there a config to tweak so that racket-mode shows only error traces when running racket-test? Most of the time, I don’t need coverage.
@capfredf I’d suggest racket -l errortrace -l racket/base -e '(require (submod <fn> test))'
Oh, this is much better! @sorawee thanks