***heroux_ is now known as heroux
***_hanhart_ is now known as _hanhart
***adhoc_ is now known as adhoc
***wleslie_ is now known as wleslie
<sneek>I last saw ijp on Jul 08 at 10:14 am UTC, saying: probably negligible if any. <os_>ok, i will use the mailist ***petercom1and is now known as petercommand
<adhoc>"weak and ephemeral hashtables" <adhoc>taylanub: i've little experience with that in scheme/lisp <adhoc>is the aim that they can be lazy ? <taylanub>ephemeral is the correct implementation, non-ephemeral weak is the fast implementation. <amz3>I think it's the concepet of weak reference in python <amz3>but it well explained in the document <roelj>Is there an example to simply read a file using Guile? <taylanub>roelj: you can read it line-by-line via the (ice-9 rdelim) module <taylanub>(or by other delimiters. see read-line, read-delimited, etc.) <roelj>I think this is exactly what I was looking for. <amz3>I always look for that module <lloda>(help) searches all modules in the path if you use (use-modules (texinfo reflection)), e.g. in ~/.guile. For example (help read-line) works even when (ice-9 rdelim) isn't loaded. Would be great if ,a worked like that, too. <amz3>no, it's metacommand in guile REPL <amz3>there is even ,d which stands for ,describe <davexunit>nalaginrut: do you want a guile program to produce the image? <nalaginrut>davexunit: I'm drawing it with inkscape now, but yes, I would like to do that in the future, what's it? ;-) <nalaginrut>I think something like Scheme-postscript would be nice <davexunit>thanks to 'guix edit guile-rsvg' for opening it in emacs automagically :) <alezost>davexunit: you can also use "M-x guix-edit" <daviid>nalaginrut: see guile-charting as well <paroneayea>hopefully today will be a plesant day with guile and implementing json-ld <paroneayea>let's see if i can have running code by the end of the day this time, as opposed to a pile of untested algorithms ;) <paroneayea>you know what's great? writing code that doesn't have users yet ;) <paroneayea>I'm doing insane amounts of experimentation and refactoring that would be hard to get away with elsewhere <paroneayea>I really need to get a up an actual automake file and and an easy way to dump in tests before that bites me though <daviid>davexunit: fi, guile-srvg-2.18.1 compiles here <paroneayea>ACTION opens the robusthness paper on his other monitor to read idly while hacking <davexunit>paroneayea: 2 papers by sussman and stallman in there <davexunit>I hear tell that I can sit in on such classes <alezost>davexunit: it was added recently and should be available only if you use emacs ui from the guix git repo <alezost>davexunit: probably not, to check you may "M-x find-function guix-<anything>" and see if it is from your git repo <davexunit>alezost: yeah it's loading it from /gnu/store. I thought I had edited my load path to use my git repo... <paroneayea>arg, I need to write some stuff that folds over either vhashes but doesn't process a key twice if there are older values... I'm writing a fold-unique procedure that secretly keeps track of what keys it has seen in a hash table but hides that from the user <paroneayea>guess I'll leave myself a functional interface but secretly do this imperative thing <paroneayea>the alternative is to make the result being built up from the fold be a pair that contains both the real result and what's ben seen <davexunit>paroneayea: if a procedure only performs mutation on internal variables then it is still referentially transparent. <paroneayea>seems like kind of a combination of the actor model, capabilities, and reactive programming <davexunit>paroneayea: there's a video somewhere where sussman discusses his propagaters <davexunit>the implementation of reactive systems isn't functional, either. <davexunit>though I don't know how different propagaters are from the "system of constraints" in SICP <please_help>The following results in leaks but the hook still gets called while the leak is happening (so the problem isn't exactly the same): paste.lisp.org/+3BIF am I using scm-gc-register-allocation correctly? <taylanub>others feel free to chime in, but I especially trust mark_weaver's judgment :) <please_help>taylanub: exposing the hashing functions for the eq? and eqv? procedures is a hindrance for Scheme implementations with a moving garbage collector; what does that mean?