jaeger changed the topic of #crux to: CRUX 3.7 | Homepage: https://crux.nu/ | Ports: https://crux.nu/portdb/ https://crux.ninja/portdb/ | Logs: https://libera.irclog.whitequark.org/crux/
ppetrov^ has joined #crux
jason123santaoni has joined #crux
jason123onirc has quit [Ping timeout: 272 seconds]
<cruxbot> [core.git/3.7]: tar: update to 1.35
ppetrov^ has quit [Quit: Leaving]
ppetrov^ has joined #crux
<cruxbot> [core.git/3.7]: libarchive: update to 3.7.0
jue has joined #crux
jue has quit [Ping timeout: 240 seconds]
<farkuhar> SiFuh: after further testing with both the previous and the latest libarchive, it appears to be a bug in pkgutils, not bsdtar. The command that reports different contents for the two tarballs is 'pkginfo -f', whereas 'tar tvJf' reports identical contents.
<SiFuh> Hmm
<farkuhar> looks like we might have to modernize the pkgutil.cc code, to accommodate any changes in the libarchive API.
<SiFuh> Or use Winrar with WINE :-P
<SiFuh> Des libarchive need to be bleedin' edge?
<SiFuh> Why not have one built specifically for pkgutils?
<farkuhar> It's probably too much work to backport any security fixes from upstream, just to keep a crux-specific version of libarchive. Why not follow the upstream development and just use the latest function calls (not whatever was current when pkgutils was first written)?
<SiFuh> But was it a security fix?
<farkuhar> No, today's release notes (libarchive 3.7.0) do not mention any CVE.
<SiFuh> That's how it usually is.
<farkuhar> It's not critical to support both ways of creating a tarball, 'bsdtar -a -cf name#version-release.pkg.tar.xz' versus 'bsdtar -J -cf name#version-release.pkg.tar.xz' ... But I'm dying to know why 'pkginfo -f' digests the output of those two commands differently.
<farkuhar> What's even weirder is that the discrepancy hasn't manifested itself in all ports. So far I've only seen the divergent 'pkginfo -f' results for contrib/go.
<cruxbot> [contrib.git/3.7]: gitea: 1.19.4 -> 1.20.0
<cruxbot> [contrib.git/3.7]: highlight: 4.6 -> 4.7
<cruxbot> [contrib.git/3.7]: prometheus-node-exporter: 1.6.0 -> 1.6.1
<cruxbot> [contrib.git/3.7]: matrix-synapse: 1.87.0 -> 1.88.0
<cruxbot> [contrib.git/3.7]: p5-libwww: 6.71 -> 6.72
<cruxbot> [contrib.git/3.7]: python3-distlib: 0.3.6 -> 0.3.7
<cruxbot> [contrib.git/3.7]: python3-matrix-nio: 0.21.0 -> 0.21.1
<cruxbot> [contrib.git/3.7]: dysk: 2.6.0 -> 2.7.1
<cruxbot> [contrib.git/3.7]: lfs: renamed to dysk
<jaeger> go is often a pain for packaging... currently fighting with go and docker about a new issue that they claim to have fixed upstream but I can't get their fixes to work locally
<SiFuh> I don't like go, I think they should rename it to halt or even stop.
<ppetrov^> SiFuh, "gone"?
<SiFuh> Speedbump, boom gate, tool booth ;-)
<SiFuh> Toll*
<SiFuh> Anchor, wheel clamp, flat tyre... so many better alternatives
<SiFuh> I thought of the perfect name 'Reverse'
<r0ni> i was wondering what was wrong with go, as the latest 1.20.6 built just fine on arm64 even for me, but then I re-read the statement and it says "docker" and it all makes sense
<SiFuh> go/reverse*
<ardo> how do I get to create a login for crux flyspray? I don't see any 'register' option on the bugs page on crux.nu.
<SiFuh> I doubt the hidden register options works ;-)
<ardo> valgrind complains about missing symbols in /lib/ld-linux-x86-64.so.2, glibc's .nostrip related line does not match the .2 at the end of the filename.
<cruxbot> [core.git/3.7]: cmake: 3.26.4 -> 3.27.0
<cruxbot> [contrib.git/3.7]: kodi-addon-inputstream-adaptive: 20.3.9 -> 20.3.11
<cruxbot> [contrib.git/3.7]: p5-business-isbn-data: 20230714.001 -> 20230718.001
<cruxbot> [contrib.git/3.7]: python3-click: 8.1.5 -> 8.1.6
ppetrov^ has quit [Quit: Leaving]
tilman has quit [Ping timeout: 264 seconds]
tilman has joined #crux