IRC channel logs

<jackdk>Is anyone else having trouble acquiring gnulib? I'm trying the `live-bootstrap/rootfs.py` script and getting `requests.exceptions.RequestException: [Errno Unable to download _ from ] any mirror or original` when trying to fetch gnulib-8e128r.tar.gz

<fossy>stikonas: earlier

<fossy>stikonas: the new chain I am trying is 5.6 -> 5.8 -> 5.16 -> 5.32 -> 5.40 or something like that

<fossy>lanodan: no :-\

<fossy>lanodan: it is very tedious work unfortunately, a centralized effort would be good, but we are still a ways from that

<fossy>stikonas: will revisit mpfr in the near future hopefully. i think there is something wrong with my regenerated strtofr.c. it is definitely not identical with the original table, i remember that much

<fossy>but it was fairly similar, and tests were passing iirc

<lanodan>Well I think only way it could be centralised is upstream and otherwise something which can carry patches, and I think that would end up being a duplicate of live-bootstrap

<fossy>jackdk: ah, i will update the README. this was introduced a little while ago, you need to use a mirror or repackage git repositories yourself. there are a few ways to do this https://github.com/fosslinux/live-bootstrap/issues/519#issuecomment-2746486189

<fossy>lanodan: for sure, but live-bootstrap is (a lot) more limited in scope than a distro is

<fossy>i think a lot of the regeneration work can be pushed upstream

<fossy>(for sufficiently new versions)

<fossy>but actually a lot of the story of regeneration has improved materially in the last decade or so, e.g. with perl you have to call a ton of random scripts in the older versions, but in the new versions it's just one regen.pl script

<fossy>same with python

<lanodan>true, which then is like a cross-distro effort which is stuck to being decentralised due to how different they can be (like having carried patches for musl compatibility from gentoo to alpine and vice-versa, and had to tweak them few times)

<lanodan>Oh nice to hear, even though there's still bootstrapping issues with both

<fossy>yeah, _bootstrapping_ work in particular is not improved much, but the actual "pregenerated files in repositories" has become a bit easier to deal with

<fossy>perhaps decentralised is inevitable... but the part that would be pretty similar is any steps of recreating pregenerated files

<vasi>actually, what's the status of the wiki? The one at bootstrapping.miraheze.org is marked as inactive, and I see that oriansj's wiki on man.sr.ht has more Google-juice right now. Should I be sending merge requests against that repo, instead of editing the wiki?

<lanodan>Oh, https://man.sr.ht/~oriansj/bootstrappable which I guess could be preferable as it's easier to clone (and depending on familiarity with git send-email could be easier or harder to do modifications)

<blockhead> "google-juice?"

<vasi>blockhead: it appears higher in search results, is more discoverable

IRC channel logs

2025-07-14.log