IRC channel logs

2023-07-27.log

back to list of logs

<sevan>ok, so the prebuilt file are for linux, big endian is for linux & mips, and the little endian ones are for x86
<rlb>My impression is that they're supposed to work on any platform that matches the criteria.
<rlb>(in theory)
<rlb> https://buildd.debian.org/status/package.php?p=guile%2d3%2e0
<rlb>e.g.
<sevan>does that work with guile on x86 windows via wsl or cygwin? (something not running linux)
<sevan>rlb: it looks like it's doing the bootstrap itself on Debian https://buildd.debian.org/status/fetch.php?pkg=guile-3.0&arch=powerpc&ver=3.0.8-2&stamp=1646535997&raw=0
<sevan>(stage0 generating ice-9/eval.go psyntax-pp.go boot-9.go)
<rlb>sevan: oh, I sorry, I misinterpreted. I just meant that I believe guile only intends for there to be 4 flavors, and so "the right one" should work on any compatible arch, but yeah, debian doesn't use the ones in the release archive (or use the release archives at all -- debian's packaing repo is derived directly from guile git).
<sevan>rlb: right, and the issue wont crop up on Debian because if you're regenerating things from scratch on the target host, it's all going to work ok, it's just going to take a lot of time. The iMac G5 took around 3-4hours to do the 3 stage0 .go files which are normally in the bootstrapped kit. Looking through the bug/list emails, a Debian developer was testing on m68k! no idea how long that would take :)
<rlb>sevan: depending on what you mean, format-wise I'd think that the matching-flavor packaged files are supposed to work on the g5, but whether they do or not... And of course if you needed to patch things, you might have to rebuild some of the slow bits anyway.
<sevan>rlb: this potentially explains the issue https://debbugs.gnu.org/cgi/bugreport.cgi?bug=26854#11