<guix-vits>Blackbeard: i'm just pressed something, and got "HELLO" buffer in emacs; can't reproduce it. It shows charset support in emacs. <Blackbeard>guix-vits: C-h ? Will give you all the help options <Blackbeard>I don't remember right now which one is for hello ***rekado_ is now known as rekado
<lloda>anyone seen a problem using srfi-64 where you (set! test-log-to-file #f) and it still writes logs anywayf <manumanumanu>lloda: is that expected to work? I thought you had to write your own test runner to not output anything <manumanumanu>lloda: maybe try (set! (@@ (srfi srfi-64) test-log-to-file) #f). That won't work in guile3 i suspect, since the module is declarative <manumanumanu>So, I want to host a real website using the guile webserver. How should I best do it? The guile server will be responsible for serving everything that ends in .gml (or whatever I'll call it), and the rest should be handled by a regular web server. <manumanumanu>I have avoided everything that targets the web since about 2002, so i feel quite lost. <rekado>manumanumanu: I always use a reverse proxy <rekado>I often use a pool of Guile web server processes and have the reverse proxy serve as a load balancer <rekado>with nginx as the reverse proxy you can match on locations ending on .gml (or whatever) and pass only those to the application server(s) <rekado>so other requests won’t be seen by the application <manumanumanu>Then I only need to figure out if I can splice raw HTML into sxml templates for those ugly files which I haven't made a proper markup... <manumanumanu>Why did I jump on to the markdown bandwagon instead of just continuing to use my old xml format? :D :D <lloda>thx manumanumanu I did try the @@ and it didn't work probably b/c of the declarativeness thing. It used to work before and maybe that's what broke it. <lloda>I can put that flag in a box I guess <manumanumanu>lloda: the upstream srfi-64 exports that binding, and even though it isn't part of the spec, it should really be counted as a regression that it stopped working <manumanumanu>or that variable should be boxed, if we want to deviate from upstream. <manumanumanu>you are more qualified to reason about that than I am, though, and in a much better situation to start a discussion about it. <manumanumanu>if there are any speed improvements of the declarativity of srfi-64, you could hack the variable to be identifier-syntax and do boxing and unboxing. <leoprikler>Hi, is there a way of running a hook until the first procedure succeeds without reinventing hooks themselves? <leoprikler>I guess probably something from SRFI-1 could help me if I used hook->list, but I'm looking for a "simpler" solution. <marmulak>is guile the best scheme to develop with, or are there other schemes you guys know about or use? <srandon111>guys fromm the manual how can i understand how to convert a bytevector vector to an hex string? <guix-vits>marmulak: idk, but emacs is aware of those: chez, chibi, chicken, gambit, guile, mit, racket <RhodiumToad>hm, there's no vector-fold or similar for bytevectors? *RhodiumToad curious about srandon111's question even though they left <brendyyn>(guix bytevectors) has that, although it pads the bytevector to an even number of zeroes, and errors if you give it an unpadded one <marmulak>I like how guile compiles my scripts before running them, although I kind of want to suppress the output of the compilation message. It's too verbose <leoprikler>you could also convert the bytevector to a list first and then fold <leoprikler>or to a string and then string-fold if you're really evil ;) <RhodiumToad>hm. array-map-in-order! and array-for-each only pass the element and not the index, while array-index-map only passes the indices and not the element <brendyyn>I found myself needing to translate pythons [::-1], but it works on both strings and bytes, so i would convert to bytevector, use a bytevector-reverse i made, and then back to string. <brendyyn>srandon111: there is bytevector->base16-string in (guix bytevectors), although it outputs hexadecimals with padded with zeros to an even length <jcowan>RhodiumToad: SRFI 160 is pretty portable and provides folding over u8vectors; these are not the same as bytevectors in Guile (uniquely so AFAIK) but could be easily adapted to them <RhodiumToad>(define (bytevector->hex bv) (call-with-output-string (lambda (port) (array-for-each (lambda (b) (put-string port (number->string (ash b -4) 16)) (put-string port (number->string (logand b 15) 16))) bv)))) <RhodiumToad>(define (bytevector->hex bv) (call-with-output-string (lambda (port) (array-for-each (lambda (b) (format port "~2,'0x" b)) bv)))) <dsmith-work>format actually has some looping things built in ( ~{ and ~} ) , but they are for lists. <jcowan>bring out your format combinators <jcowan>I would guess independent invention ***lfam_ is now known as lfam
***guix-vits is now known as guix-ci
***guix-ci is now known as Guest75614
<RhodiumToad>is it considered bad style to use letrec-syntax to define convenience abbreviations used by a small number of functions? <jcowan>I personally avoid let-syntax in general and just use define-syntax at the top level, but inside a library if I don't want to expose it. But I wouldn't call it bad style specifically. <jcowan>Note that there are issues around the interpretation of let-syntax <jcowan>as to whether it expands to a lexical countour (let () ...) as in R7RS or not as in R6RS. R5RS was ambiguous. <jcowan>I voted for the contour interpretation in the R7RS-small committee, but that's because I misunderstood the issue. <jcowan>It's easy to do a contour version given the splicing version, but not vice versa. <manumanumanu>RhodiumToad: I have a bytevector to hexstring if you want it. it uses a lookup table, though. <apfel>hi there, can i somehow redefine a inner function of a function from outside the function? Like, i want to redefine inner-fun from this example on the repl: https://pastebin.com/2k7AJRk9 ***dddddd_ is now known as dddddd
<mwette>or you have to pass the "inner function" as an argument to outer-fun <jcowan>manumanumanu: I'd like to reuse your hex library in SRFI 178, with a bit of renaming, are you OK with that? <jcowan>The license would need to change to MIT, but that's pretty trivial ***jonsger1 is now known as jonsger