IRC channel logs

2019-11-23.log

back to list of logs

<dsmith-work>What would be a better name for "bytevector"? It just feels clumsy as a name. "string" is nice. No one says characterarray or charactervector.
<RhodiumToad>string is a sequence of characters though.
<dsmith-work>Yes. But back in the day, that was also a sequence of bytes.
<RhodiumToad>bytevector is probably as good a name as any, unless you want to go with "blob"
<RhodiumToad>in postgresql we call them "bytea", which is probably not a good name
<dsmith-work>Something like blob, but different. Something along those lines. Single syllable.
<dsmith-work>"poab" Plain Old Array of Bytes. But I don't like that at all.
<jcowan>In Algol 68, and in early drafts of R6RS, the term was "bytes". One bytes, two byteses, ...
<jcowan>I did push for blob, but the R7RS-small committee thought it was gratuitously different from R6RS
<dsmith-work>In some ways, blob is not bad at all. But is suppposely for "large" things. (Firmware images, videos, ..)
<dsmith-work>Bytearray is ok. Just wishing there was something nicer that's all.
<dsmith-work>jcowan: Did you actualy ever use algol?
<jcowan>Algol 60 in one college class (with punch cards)
<jcowan>I studied A68 closely but never used it in anger.
<jcowan>I have a preso on A60 ....
<jcowan> http://vrici.lojban.org/~cowan/Algol60.ppt
<jcowan>I started out by playing the William Tell Overture
<jcowan>"Return with me to those thrilling days of yesteryear, when coders wrote on coding forms, I/O was on punch cards and line printers, and computers were used to ... compute!"
<RhodiumToad>year or two ago I installed a copy of the mainframe OS I used at university just out of nostalgia. there was an algol-w compiler in there somewhere
<RhodiumToad>also a fairly early xlisp
<dsmith-work>Ahh xlisp.
<RhodiumToad>not all that many compilers unfortunately due to licensing issues
<dsmith-work>The first lispy thing I ever touched was advsys, an interactive fiction thing by David Betz, of xlisp fame.
<dsmith-work>Didn't they put xlisp into AutoCAD ?
<RhodiumToad>dunno.
<RhodiumToad>lisp was I think the third high-level language I learned as a kid
<RhodiumToad>after basic and forth
<dsmith-work>Yey Forth.
<dsmith-work>Did several embedded systems with Forth.
***logicmoo is now known as dmiles
<jcowan>Algol W was halfway from A60 to Pascal
<jcowan>there are about five A60 compilers/interpreters available now
<RhodiumToad>I never did algol 60, but I did encounter Simula
<jcowan>RhodiumToad: Indeed, the Simula compiler can handle A60 programs just fine: it's one of the five I mentioned:
<jcowan>GNU Marst (compiler to C)
<jcowan>Nase A60 (interpreter)
<jcowan>Algol60i (interpreter written in Python)
<jcowan>Racket Algol 60 (compiler to Scheme)
<jcowan>Cim (Simula 67 compiler, near-superset)
<lispmacs>ok, I'm about to take the plunge and convert my desktop PC from Debian to Guix System
<lispmacs>here it goes, see ya on the other side, hopefully!
*lispmacs emerges on other side of his Debian -> Guix desktop conversion
<zig>o/
<lispmacs>now, I've got Guix system and emacs
*lispmacs goes to bed
<zig>fwiw, I pushed a definition for guile-2.9.5 (and guile-arew (previously known as guile-r7rs)) in my guix channel at https://git.sr.ht/~amz3/guix-amz3-channel/
<civodul>zig: i've also updated "guile-next" in guix.git :-)
<zig>neat
<zig>gnutls requires a patch to be build with guile-next
<civodul>zig: yeah, the changes were committed upstream but we're using an older version
<civodul>i think we should add a "guile3.0-gnutls" package based on that newer version
<civodul>wingo: there are a few "guile3.0-" packages and they're happy with 2.9.5: http://ci.guix.gnu.org/eval/8882?status=succeeded
<civodul>gdb support in Guile 3.0 is broken
<zig>bumping gnutls to 3.6.10 is not enough, the build segfault on a call to guild command.
<retropikzel>How can I get parameter from POST request body easily?
<retropikzel>No, not from body but from place they are supposed to be. Are they supposed to be in the body?
<ng0>wasn't a new gdb released? I've just seen gdb being merged for my system, but this might be unrelated to gdb support..
<zig>retropikzel: yes POST params are in body. use https://github.com/a-guile-mind/guile-web/blob/master/src/web/decode.scm
<zig>retropikzel: the default encoding is the same as query string.
<retropikzel>zig, thank you very much :)
<zig>retropikzel: file uploads use something else.
<zig>the code of decode could be better
<retropikzel>I have version 2.0.11, on the website it says that to install 2.2 in Trisquel but there is no package guile-2.2, only guile-2.0. I need to compile to update.
<zig>retropikzel: or install guix on top of trisquel
<zig>retropikzel: yes, it says 2.2 because that is what was stable on ubuntu / guix at the time.
<zig>retropikzel: you can use those files with guile 2.0 I think.
<zig>or check the NEWS file in git master
<zig>retropikzel: mind the fact that the above link is just slapped together procedures that I had will *experimenting* webdev.
<zig>retropikzel: hence there is prolly better implementation and missing stuff (like file upload
<retropikzel>Ok
<zig>retropikzel: I made it work for my usecase, but once you have a good enough grasp of scheme or guile, look also at artanis.
<zig>I hesitated to delete the repo, but... promised myself to never delete anything.
<zig>I hesitated because it is far from perfect.
<retropikzel>Dont delete anything
<retropikzel>I have deleted so much stuff over the years and I regret it
<zig>yes
<zig>I learned that the hardway
<retropikzel>I get the value from body like this now: https://paste.artemix.org/s/HRkIYN was fun to do and fits my case for now
<eiro>hello people
<eiro>is there a way for a lambda to call itself? something like the perl __SUB__ symbol ?
<eiro> http://mc.aude.re/pub/tmp/1574434800/tcrguile.txt < i would like to replace the inner function but maybe it's better to write that way ?
<civodul>hi eiro!
<civodul>eiro: you can do: (define (fib x) (if (zero? x) 1 (fib (- x 1))))
<civodul>it's a recursive function
<eiro>civodul: i know but in the case of factorial, you need 2 arguments to get a tcr
<eiro>so what i did is a 1 arg function based on the definition then the call of a 2 args function
<eiro>i used a function instead of a lambda just because i don't know how to make a lambda recursive to itself
<eiro>(maybe it's just impossible to do so)
<mwette> eiro: you need a name, but you can use a let also (let ((f (lambda ...))) (f 1 2 3))
<mwette>eiro: check out named-let (let foo ((a 1) (b 2) ...) xxx (foo 4 5 ..))