IRC channel logs

2025-11-15.log

back to list of logs

<rlb>main appears fine on 390x, but it segfaults badly during the build on current debian sid (unstable) powerpc -- reminded me debian has a patch for that (iirc): https://debbugs.gnu.org/cgi/bugreport.cgi?bug=45214
<rlb>I wonder if doing something similar in main makes sense for now...
<rlb>similar/related
<rlb>Oh, and the current (debian specific) patch - https://salsa.debian.org/rlb/deb-guile/-/blob/deb/guile-3.0/d/sid/master/debian/patches/0007-Fix-32-bit-big-endian-builds-via-Oresolve-primitives.patch?ref_type=heads
<rlb>wingo: wondered if that meant anything to you offhand (the fact that the patch fixes big endian 32-bit) ^
<sam113101>guys
<sam113101>what is the closest thing to dotimes that we have?
<ekaitz>sam113101: a loop?
<sam113101>like a let loop?
<ekaitz>yep
<ekaitz>sam113101: you can implement dotimes in terms of it, and then use that
<ekaitz>or sam113101 you could do something cooler
<ekaitz>(for-each (lambda (i) (your thing with i)) (iota N))
<ekaitz>if you need more dimensions you can add more iotas
<identity>sam113101: you can use ‘do’ with a counter, like (do ((i 10 (- i 1))) ((zero? i)) (display i) (newline))
<identity>(iota n) will allocate a list of length n, which may be quite large
<sam113101>I wish there was a displayln function
<dsmith>Guile is not lazy, so (iota N) always makes a list, lots of consing for large N
<dsmith>sam113101, Like (define (displayln x) (display x) (newline)) ?
<listentolist>How should I send a patch to guile-lib if it is too big for the guile-devel mailing list? It includes all html character entities.
<listentolist>It is for the html-tokenizer in htmlprag.
<identity>listentolist: is guile-lib not on codeberg yet? you could put it somewhere on the internet where it will not expire and link to it from the mailing list
<listentolist>identity: Thanks.
<ekaitz>listentolist: i think apteryx wanted to add it to the guile org in codeberg
<euouae>Hello
<rlb>yeah, that patch still fixes powerpc (test-out-of-memory still fails, but it's known to be flaky)
<rlb>(it also "fixed" m68k and maybe hppa and/or alpha)
<mwette>I remember 68k machines. I had a (used) personal Sun 3/60 once.