<farkuhar>
Looks like some libtool files got corrupted. Try `grep -lr libuuid.la /usr/lib`, identify the packages that own those files, and rebuild them.
<surface-vermin>
okay, I'll try that
<farkuhar>
Long-time CRUX users have a standard recipe they use to automate this task, something like: for f in $(grep -lr libuuid.la /usr/lib 2>/dev/null); do pkginfo -o $f; done | awk '{ rebuildMe[$1] } END { for(i in rebuildMe) {print i} }'
<farkuhar>
The task of identifying the ports, I mean. For the actual rebuilding you pass the resulting list to `prt-get update -fr`
<surface-vermin>
okay, thank you
<farkuhar>
No problem. Stick around, someone might chime in with other suggestions specific to consolekit.
ukky has joined #crux
<ukky>
the reason for failure is missing /usr/lib/libuuid.la
<ukky>
After switching to meson to build core/util-linux, libuuid.la is not created anymore.
<surface-vermin>
Following your advice farkuhar i was able to complete the installation of consolekit
<surface-vermin>
So, i'll deffinitly be keeping a note of the for f in $(grep -lr libuuid.la /usr/lib 2>/dev/null); do pkginfo -o $f; done | awk '{ rebuildMe[$1] } END { for(i in rebuildMe) {print i} }'
<surface-vermin>
Yeah, before asking I googled the missing .la, I was presented with a ton of stackoverflow posts of how annoying libtool is.
ukky has left #crux [#crux]
<SiFuh>
*.la files should rarely exist anymore these day.
<SiFuh>
Glad it is working now surface-vermin. Was very busy this morning so had no time.
<SiFuh>
farkuhar to the rescue ;-)
<r0ni>
.la files strike again
<SiFuh>
r0ni: Why is it still a thing?
<r0ni>
cuz most of the ports don't strip out the la files during the builds, so when updating ports, if they weren't there, the fresh builds put them there.
<SiFuh>
Yeah, but why are *.la files still a thing?
<r0ni>
at least partially, and then something comes along and it says I can't find la files for _xxx_ package and it poops itself
<r0ni>
autotools i think
<r0ni>
meson and such don't make em, that i've seen anyway
<SiFuh>
I just don't understand why most programs still create it. It's mostly just junk
<SiFuh>
There are exceptions of course but still..
<r0ni>
just a part of that build system, likely why many are moving away from it?
<surface-vermin>
No problem SiFuh, I was fixed
<surface-vermin>
*It was fixed
<r0ni>
that one liner on the crux-musl page will fix issues with la files forever though, tbh it should come preinstalled as such
<surface-vermin>
I've made a note of that, and have it close by
_moth_ has quit [Ping timeout: 255 seconds]
<SiFuh>
r0ni: emmett1 has moved away from his MUSL system and is now using CRUX-MUSL as his main driver.
<SiFuh>
He took the ISO I had made with his help and said that it is better.
<SiFuh>
So he should be producing many ports as he goes.
<SiFuh>
I plan to make a CRUX-MUSl 3.8 as well when 3.8 becomes ready. But as I said to jaeger. I will only do the bare essentials around core as the basic building blocks for anyone who chooses to run it and build upon.
zorz has quit [Quit: leaving]
<surface-vermin>
Thank you for the help today, I'll cya around
surface-vermin has left #crux [#crux]
SiFuh has quit [Remote host closed the connection]
SiFuh has joined #crux
<r0ni>
SiFuh: i'd love to try it yet, haven't had the time to invest in it as of now but i'd be going arm64 anyway so i'm totally on my own i'd think
SiFuh has quit [Remote host closed the connection]
SiFuh has joined #crux
ppetrov^ has joined #crux
_moth_ has joined #crux
lavaball has joined #crux
<cruxbot>
[contrib/3.7]: assimp: fixed build with ccache
<ppetrov^>
libblockdev, shouldn't the download link be https://github.com/storaged-project/$name/releases/download/$version/$name-$version.tar.gz
<ppetrov^>
$version-1 -> $version
<SiFuh>
What about $version-2 ?
<SiFuh>
Then you'd need to put another variable in unless it was $version only
<SiFuh>
n="-1" $version$n.tar.gz
<farkuhar>
ppetrov^, are you implying that https://crux.nu/Wiki/DealingWithGithub needs to be updated? Is their URL rewriting engine doing something differently these days?
<farkuhar>
If the Github URL rewriting engine has introduced new patterns, are they enabled globally, or does each individual Github repo have to opt in?
ivandi has quit [Quit: WeeChat 4.4.2]
ivandi has joined #crux
stratofall has joined #crux
<cruxbot>
[contrib/3.7]: libblockdev: fixed source download link
<cruxbot>
[opt/3.7]: strace: 6.10 -> 6.11
<cruxbot>
[opt/3.7]: gnupg: 2.5.0 -> 2.5.1
<cruxbot>
[opt/3.7]: git: 2.46.0 -> 2.46.1
<ppetrov^>
farkuhar, i just noted that the download link was wrong, it had /download/$version-1/ instead of /download/$version/
<ppetrov^>
it was fixed ^
ymas has joined #crux
ymas has quit [Changing host]
ymas has joined #crux
ymas has left #crux [#crux]
<farkuhar>
ppetrov^: oh, I thought the notation $version-1 -> $version was for a legacy rewrite rule that Github was no longer supporting. Nevermind the noise, then.
<ppetrov^>
well, i was not clear enough i guess
<ppetrov^>
i have no opinion of whether pkgmk should be able to handle git natively
<ppetrov^>
this would introduce a dependency on git
<ppetrov^>
and so far, although very annoying at times, download links from git are manageable
<farkuhar>
the way I see it, a port that opts into using any native git functionality in pkgmk would declare git as a dependency, and then it can expect all the git commands to be available. pkgutils itself does not need to declare a git dependency.
<ppetrov^>
heh, that's actually pretty elegand
<SiFuh>
elegant ?
<ppetrov^>
poor choice of a word or the spelling mistake?
<SiFuh>
Not sure, I was asking you.
<SiFuh>
I think you meant elegant.
<ppetrov^>
elegand as in "elegant solution"
<SiFuh>
All good
<SiFuh>
I don't think elegand is a word
<SiFuh>
Unless you want to add it to the dictionary.
<ppetrov^>
i can't spell
<SiFuh>
That's the second meaning of zorzing in the dictionary.
<ppetrov^>
you know what was embarrasing? i kept writing 'chromoZome' to some collaborators from Scotland..
<SiFuh>
Oh no
<ppetrov^>
heh
<SiFuh>
As long as they know you are not an English speaker they may push it under the rug and ignore it.
<SiFuh>
What was the outcome by the way?
<SiFuh>
I know in the US they often replace 'z' sounding 's' with a 'z' like recognize but it should be spelled regonise
<SiFuh>
Never figured out why cats was never spelled catz there.
<SiFuh>
As I mentioned before, a lot of the US spelling was because each letter cost money when using the printing press so they often shortened words or used more common letters. So gaol became jail and colour became color
<ppetrov^>
What was the outcome by the way? <- the corrected me
<SiFuh>
HAHAHA
<SiFuh>
Walk of shame dude, walk of shame :-P
<ppetrov^>
i had it in a preliminary figure
<ppetrov^>
yep
_0bitcount has joined #crux
<farkuhar>
ppetrov^: "preliminary figure", heh. The reviewers made the necessary changes before publication. Just like another instance of $version that will eventually be fixed by the time CRUX 3.8 is released: https://git.crux.nu/ports/core/src/branch/3.8/gcc/Pkgfile#L72
<farkuhar>
Thanks for pointing out how outdated the TODO3-8 wiki page has become, by the way. Toolchain updates are coming out faster than we can pull together a new distro release.
lavaball has quit [Remote host closed the connection]
<ppetrov^>
farkuhar, the 'preliminary figure' is before submission, when discussing the results with the other coauthors
lavaball has joined #crux
lavaball has quit [Remote host closed the connection]