Re: [racket-users] Why is get-impure-port* slower than a system call to curl?

Curl supports http/2 (and even has experimental support for http/3 which is orders of magnitude faster than http/1.1. Depending on what protocols the server you're talking to supports, that could be part of it. On Monday, September 14, 2020 at 3:42:37 PM UTC-7 samdph...@gmail.com wrote: > Hi

Re: [racket-users] [racket users] describe variant issue?

This is a difference in behavior between Racket BC and Racket CS, and not something in the describe library: [samth@homer:~/work/teaching/c211 (master) racket-7.8] racket Welcome to Racket v7.8. > (struct->vector 5) '#(struct:fixnum-integer ...) > ^D [samth@homer:~/work/teaching/c211 (master)

[racket-users] [racket users] describe variant issue?

Hi guys, I’m not sure why the describe library’s variant is always returning ‘simple regardless of numeric value. For instance, the docmentation says: (variant 1) ->

Re: [racket-users] Writing make-sized-byte-string alternative on CS

Thank you for the rapid response! ~slg ‐‐‐ Original Message ‐‐‐ On Tuesday, September 15, 2020 5:15 PM, Matthew Flatt wrote: > You use `make-bytes` and `memcpy`, instead of writing a new loop. (The > non-copying part of `make-sized-byte-string` is what CS can't support.) > > Matthew

Re: [racket-users] Writing make-sized-byte-string alternative on CS

You use `make-bytes` and `memcpy`, instead of writing a new loop. (The non-copying part of `make-sized-byte-string` is what CS can't support.) Matthew At Tue, 15 Sep 2020 21:12:22 +, Sage Gerard wrote: > The docs for >

[racket-users] Writing make-sized-byte-string alternative on CS

The docs for [make-sized-byte-string](https://docs.racket-lang.org/foreign/foreign_pointer-funcs.html?q=free#%28def._%28%28quote._~23~25foreign%29._make-sized-byte-string%29%29) indicate that CS does not support this operation. I'm guessing I just need to fall back to building a loop using

Re: [racket-users] package manager woes on Windows 10?

The change I made is in the gui-pkg-manager repo (well there may be one more word in the name). Yes, it does just strip. I didn't see any code that was doing what you describe but I also didn't look for it! Robby On Tue, Sep 15, 2020 at 8:39 AM John Clements wrote: > I have a question about

Re: [racket-users] package manager woes on Windows 10?

I have a question about the new behavior. (ObResearch: actually, I checked the drracket, racket, and gui repos, and I couldn’t find any new push, so I couldn’t check the code myself.) Does it simply strip newlines, as Jack suggested, or does it signal an error? The latter seems less likely to

Re: [racket-users] package manager woes on Windows 10?

I just worry about backwards compatibility. There are probably places that already do something about this problem woutside of the control-- how will they interact? Maybe if someone were to audit existing code on the pkg server then we would know that changing the behavior in a certain way would