IRC channel logs

2024-09-19.log

back to list of logs

<dsmith88>Howdy Guilers!
<dsmith88>Has the bot been behaving?  (I've been in Utah last week and this)
<ArneBab>dsmith: sneek did,, I think.
<ArneBab>cwebber: doesn’t seem like GraalWasm has Tail calls: https://github.com/appcypher/awesome-wasm-runtimes?tab=readme-ov-file#graalwasm — that would have been too much fun ☺
<cpli>dpk exact complexes don't seem to exist in guile
<dpk>yeah, that’s not R6RS compliant afaik
<haugh>is this? (map (λ(x) (magnitude (make-polar 1 x))) '(0 1)) => (1 1.0)
<dpk>make-polar … will probably usually return an inexact complex, as mentioned, because it has to feed its arguments through the trig functions
<haugh>we should put the "probably usually" language in the spec, I like it
<dpk>but i think, like sqrt and co, they’re allowed to return exact results for exact inputs where that’s possible
<cwebber>ArneBab: unfortunate but I am not shocked
<retropikzel>I have a R7RS library named (retropikzel r7rs-pffi-srfi-170), which include file "r7rs-pffi-srfi-170.main.scm". They are both in directory called retropikzel. It works when I test it, but when I move the libary and the file to somewhere else and then add that someplace else to load path with -L it can not include it anymore
<retropikzel>It says: No such file or directory: "retropikzel/./r7rs-pffi-srfi-170.main.scm"
<ArneBab>cwebber: yes — there are only three implementations that support tail calls. Maybe support will grow with popularity.
<retropikzel>But the manual says: If file-name is a relative path, it is searched for relative to the path that contains the file that the include form appears in.
<retropikzel>I was informed that this is a bug, I will try with latest guile from git repo
<cwebber> https://spritely.institute/news/spritely-goblins-v0140-libp2p-and-improved-persistence.html
<cwebber>New Goblins!
<civodul>wo0t!
<civodul>cwebber: congrats to y’all!!
<civodul>this team never stops, woow
<dthompson>:)
<mwette>I love goblin's nomenclature: many uses for vats, syrup for persistent storage.
<cwebber>thanks civodul !
<cwebber>mwette: heh, vats isn't our nomenclature
<cwebber>they were called "hives" in 8sync
<cwebber>and were at the beginning of Goblins
<cwebber>but when I started studying ocap stuff more I found out the paradigm already existed, and more fleshed out too, as "vats"
<cwebber>and Mark Miller convinced me to switch from "hives" to "vats"
<cwebber>I'm still not sure that was the right move :)
<mwette>thanks for the history
<ArneBab>cwebber: nice! That’s awesome!
<cwebber>:)
<dpk>cwebber: apropos Mark Miller, i have a note that when we were talking about ‘safe’ macros in ocap systems, you said he said Scheme macros have ‘single hygiene’ but you need ‘double hygiene’ for security. do you have a reference for that?
<cwebber>dpk: http://erights.org/data/irrelevance.html
<cwebber>however I'm not sure I understand or agree with the proposed design
<cwebber> http://www.erights.org/elang/quasi/terms/like-ellipses.html more is here I htink but I'm not sure about that.
<cwebber>I wrote up my own doubly hygienic macro "solution" at one point but it has an ugly syntax that makes the Kernel language look pretty, partly because it builds on its ideas :P
<dpk>> Who should we acknowledge for Scheme's "..." system?
<cwebber>the core idea Mark is expressing is that it should be clear from glancing at how data may be manipulated, which capabilities are being handed to what
<dpk>answer: Eugene E. Kohlbecker
<cwebber>oh interesting
<cwebber>that page is 20+ years old and I doubt it will be updated tho ;)
<cwebber>but neat!
<dpk>and Dybvig and Hieb, but more probably Hieb, for working out how ellipses inside ellipses should work
<cwebber>> Oleg Kiselyov has also pointed out the wonderful/terrible story of the boy that ended up with a police record because he accidentally put an extra pair of parens in a Perl program. This is an example of the kind of accident we hope to make less likely.
<dpk>what
<cwebber>from the bottom of The Power of Irrelevance
<dpk>ACTION goes to the WayBack machine …
<dpk> https://web.archive.org/web/20060621023818/http://www.unixreview.com/documents/s=1235/urm0105l/0105l.htm
<cwebber>of course a certain dissertation preface by Olin Shivers already beats out that story in terms of ability to make school administrators nervous
<dpk>haha
<dpk>i noticed that Olin removed some of the spicier content (such as the ‘grad student’s guide to automatic weapons’) from his page at some point
<cwebber>not shocking, probably for the best.
<cwebber>and that's probably as much as we should discuss it here ;)
<dpk> http://web.archive.org/web/20100223043252/http://www.ccs.neu.edu/home/shivers/advisor-stmt-original.txt also disappeared
<dpk>cwebber: okay, this is interesting. i need to re-read Dave Herman’s PhD dissertation because it was about ‘statically typing’ the binding structure of a macro, which i think is basically what Miller is on about here
<dpk>except argh, now i’m thinking about my Algol M/Algol 21 idea again. (my approach to hygienic macros in a language with Algol-like syntax … in this case i would actually try to implement most of the syntax of Algol 60 with macros)
<dpk>ACTION pushes brain back on topic