You are not logged in.
Yesterday a "pacman -Syu" updated the package "less" from 563-1 to 580-1 and now it completely breaks paging. When running the following example, using the pager garbles the output to the point of being useless:
ag --color test /etc | less -RA little searching shows that less went from using upstream v5.63 to v5.80 and also from using pcre to pcre2:
25ca18cf730d (heftig 2021-03-18 16:41:43 +0000 6) pkgver=580
25ca18cf730d (heftig 2021-03-18 16:41:43 +0000 21) sh configure --prefix=/usr --sysconfdir=/etc --with-regex=pcre2Also, less 5.80 is still in beta, as the Less homepage indicates: "less-580 has been released for BETA testing."
I could be me, but Arch packaging and releasing a beta version of such a much used tool sounds like a pretty bad idea.
My pcre versions are currently:
$ pacman -Qs pcre
local/pcre 8.44-1
A library that implements Perl 5-style regular expressions
local/pcre2 10.36-1
A library that implements Perl 5-style regular expressions. 2nd versionWhen I re-install the old version of less (563-1), everything's fine again.
Anybody else experiencing this?
Last edited by jhalfmoon (2021-03-27 11:48:00)
Offline
https://github.com/archlinux/svntogit-p … 65c93f3d66 notes it was a required change for compatibility with systemd 248 which is currently in testing.
https://github.com/gwsw/less/issues/140 appears to be the upstream bug report.
Offline
Ah yes, that upstream bug report is spot on. Thanks for the quick reply; Much appreciated.
But how in the world have we come to the point that systemd is now *requiring* updates of packages that it should have completely no dependency on whatsoever? In other words: How come systemd is now dictating which version of less we get to use? IMHO those two packages should be completely de-coupled from each other.
Maybe I'm missing some important piece of information here, that's an option I don't rule out. But requiring an update of "less" because systemd requires it sounds very wrong. Especially because it is forcing us to now have a broken package in Arch.
What's the point of releasing a package that is known to contain such a huge bug, just because it satisfies a systemd requirement? How I see it, 5.80 should be retracted, at least until this most serious bug is fixed.
NOTE: This post may sound angry. It's not meant that way. I am very interested in the rational behind the decision to release 5.80 in its beta state and the rational to have systemd require a specific version of less.
Offline
But how in the world have we come to the point that systemd is now *requiring* updates of packages that it should have completely no dependency on whatsoever? In other words: How come systemd is now dictating which version of less we get to use? IMHO those two packages should be completely de-coupled from each other.
Maybe I'm missing some important piece of information here, that's an option I don't rule out. But requiring an update of "less" because systemd requires it sounds very wrong. Especially because it is forcing us to now have a broken package in Arch.
Offline
Tnx mate. I've filed a bug report, FWIW:
Offline