Re: [racket-users] Dumb question about Windows paths

2019-01-08 Thread Philippe Mechaï
Hi, If I understand the racket code correctly (*), you cannot generate windows paths using strings on unix. But you can do it using bytes->path-element which allows you to specify the path-convention (windows here): (build-path/convention-type 'windows (bytes->path-element (string->bytes/utf-8

Re: [racket-users] How to get libsodium-factory on OSX?

2018-12-14 Thread Philippe Mechaï
dylib is the OSX equivalent of Linux .so, so nothing wrong here. Looking at the ffi documentation it seems that libraries are looked into a list of paths reported by (get-lib-search-dirs) See

Re: [racket-users] Difficulty using subprocess

2018-11-15 Thread Philippe Mechaï
Hi Brian I think you need to add a call to flush-output after sending data to outp Regards Philippe Le jeu. 15 nov. 2018 à 19:51, Brian Adkins a écrit : > I'm porting a Ruby application to Racket (will be deployed on Linux, but > I'm developing on OSX). It uses the s3270 terminal emulator to

Re: [racket-users] How to download image using net/http-client?

2018-10-16 Thread Philippe Mechaï
Indeed the issue is with your call to pipe-content-length which is to be used with pipes and is not the Content-Length header returned when doing HTTP requests. The Content-Length can be found in the list of headers returned by the call to http-sendrecv (second value or b in your let-values call).

Re: [racket-users] How to download image using net/http-client?

2018-10-16 Thread Philippe Mechaï
Hi You could simply do a (copy-port ip op) and remove the content-length and buffer parts i think. Le mar. 16 oct. 2018 à 21:17, Chansey a écrit : > > Dear all, > > I wrote some experiment code for net/http-client. > > The purpose is downloading some images from web. > > This is my code (run

Re: [racket-users] Should #; comments be colored as comments or code in Dr Racket?

2018-10-09 Thread Philippe Mechaï
Just for the record, the bug seems to be known to someone :) https://github.com/racket/gui/blob/master/gui-lib/framework/private/color.rkt#L989 On Mon, Oct 8, 2018 at 8:15 PM Robby Findler wrote: > I like the idea that you suggested where the parens keep their paren > color but the rest of