[racket] updating users mailing list target

I'm about to change the target of the users@racket-lang.org email address to point to the Google Group. If this is the last message you get on the racket-users mailing list, it probably means you haven't yet signed up for the Google Group... or something else went terribly wrong. Best, John Cleme

Re: [racket] pkgs.racket-lang.org stuck?

A number of package builds earlier were failing to pull from GitHub as well. On Sat, Mar 21, 2015 at 10:58 PM, Jon Zeppieri wrote: > It seems like the package server isn't processing any updates. -J > > Racket Users list: > http://lists.racket-lang.org/users > __

[racket] pkgs.racket-lang.org stuck?

It seems like the package server isn't processing any updates. -J Racket Users list: http://lists.racket-lang.org/users

Re: [racket] TR: mystery behavior with optional-argument types

This feels eerily familiar to a very similar bug I found just yesterday, which I’ve reported here: https://github.com/racket/typed-racket/issues/56 Something about this is definitely very broken. For the first function, f1, it “typechecks” even

[racket] TR: mystery behavior with optional-argument types

I want to typecheck a function that looks like f1. But it doesn't work. The types of 'bool' and 'syms' are reported as Any. However, if I add a dummy keyword argument, as in f2, then everything typechecks fine: 'bool' is reported as a Boolean and 'syms' is (Listof Symbol). Feature? Bug? ;;

Re: [racket] Intercepting WebSocket connections to the Racket web server?

The Web server models applications as "servlet : request -> response" and there's no way to go from the request to the original HTTP connection. In contrast, the Web server has a lower level concept of a "dispatcher" that receives a parsed HTTP request AND the connection object (which has the I/O

[racket] Intercepting WebSocket connections to the Racket web server?

The WebSocket handshake is intentionally implemented in such a way so that WebSocket connections can be performed using the same port that the HTTP server is running on. This is implemented by making the handshake a standard HTTP 1.1 GET request with an “Upgrade: websocket” header. I’m interest

Re: [racket] making libraries work natively with both Racket & Typed Racket

On Mar 21, 2015, at 2:09 PM, Matthew Butterick wrote: > Is there an approved way of using #lang typed/racket/base/no-check (or maybe > `with-type`) to create libraries that have both a typed and untyped > interface? (The goal being to avoid use of `require/typed`) What I have done for measure

Re: [racket] making libraries work natively with both Racket & Typed Racket

1) You can't ordinarily use a macro from a typed module in untyped code. But with this technique, if you convert `fladd` to a macro, it will still work. 2) You can't ordinarily eliminate contract checks when using typed code in untyped code, because a contract is always erected at the boundary. Bu

Re: [racket] making libraries work natively with both Racket & Typed Racket

Sorry for my misunderstanding, but what’s the point of this? Why not just require the typed code in untyped code directly (as you’re doing in 'violator)? As far as I can tell, it looks like you’re just manually rewriting the generated contracts, so I’m not really sure what this achieves. > On M

Re: [racket] making libraries work natively with both Racket & Typed Racket

At RacketCon a couple years ago, Vincent and Neil Toronto worked out a crazy way to do this with macros and submodules. I think the math library uses the technique they came up with. One of them should write a blog post about the technique. -Ian - Original Message - From: "Matthew Butter

[racket] making libraries work natively with both Racket & Typed Racket

Is there an approved way of using #lang typed/racket/base/no-check (or maybe `with-type`) to create libraries that have both a typed and untyped interface? (The goal being to avoid use of `require/typed`) For instance, the following works, but maybe it's a bad idea for other reasons: ;;

Re: [racket] Unsafe version of require/typed?

As far as I know, yes. On Sat, Mar 21, 2015 at 11:02 AM, Robby Findler wrote: > Re-reading this, just to clarify: the only time when the optimizer > could cause problems is when the contracts would have signaled an > error, right? So, in other words, the difference between the fx+ from > racket/

Re: [racket] how to start a Web Application from a Server

The message that it is available on localhost is just informative during development. When you give "#:command-line? #t" it won't be shown. When you passed the argument "#:listen-ip #f" this means that it will be active on all IPs that the machine has. So, whatever the machine's IP is should work,

Re: [racket] Unsafe version of require/typed?

If you want the opposite thing, (require-from-typed-with-no-contracts ...), then you can use TR cheat: https://github.com/jeapostrophe/exp/tree/master/tr-cheat On Fri, Mar 20, 2015 at 2:20 PM, Eric Dong wrote: > It would be nice if we could have an unsafe version of require/typed, which > doesn'

Re: [racket] Unsafe version of require/typed?

Re-reading this, just to clarify: the only time when the optimizer could cause problems is when the contracts would have signaled an error, right? So, in other words, the difference between the fx+ from racket/base and the unsafe-fx+ from racket/unsafe/ops is exactly the same difference as using (r

Re: [racket] Unsafe version of require/typed?

If you look here at typed/syntax/stx.rkt: https://github.com/racket/typed-racket/blob/master/typed-racket-more/typed/syntax/stx.rkt You’ll see something that looks sort of similar to an unsafe require/typed. But if you look here: https://github.com/plt/racket/blob/f3dba3eb6bf44685cb2290d48204e81a