Package Details: earlyoom 1.3-1

Git Clone URL: (read-only, click to copy)
Package Base: earlyoom
Description: Early OOM Daemon for Linux
Upstream URL:
Licenses: MIT
Submitter: FlorianJacob
Maintainer: Jake
Last Packager: Jake
Votes: 49
Popularity: 3.02
First Submitted: 2016-04-15 15:13
Last Updated: 2020-01-11 16:59

Dependencies (0)

Required by (0)

Sources (2)

Latest Comments

1 2 3 Next › Last »

yurikoles commented on 2020-01-11 17:14

Sorry for misunderstanding it, you are correct.

Martchus commented on 2020-01-11 17:02

@yurikoles That is not how the archs specified in the PKGBUILD are supposed to be interpreted. See (although the restriction to mention only x86_64 doesn't count for AUR packages).

@Jake Thanks for fixing.

Jake commented on 2020-01-11 17:02

@Martchus: You are correct, any should only be used if it is architecture independent after compilation and that is of course not the case here. I have changed it now.

yurikoles commented on 2020-01-11 15:38


This package is built from source, but you are looking on concrete binary on concrete x86_64 system. You will have different result if you build it on different machine with different CPU arch.

Martchus commented on 2020-01-11 13:35

The architecture must not be any:

file /usr/bin/earlyoom 
/usr/bin/earlyoom: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV), dynamically linked, interpreter /lib64/, BuildID[sha1]=62e918f0b97ae9f0191035de45b59541054762ba, for GNU/Linux 3.2.0, stripped

Jake commented on 2019-06-09 10:59

Yes, i have it updated now. Sorry for the delay.

push_sla commented on 2019-06-08 09:40

Hi! Could you please update this package for the 1.3 version? There are very important fixes in this version

Jake commented on 2018-12-26 12:02

Thank you, FlorianJacob. Yes, optdepends is also not optimal, i decided for B) now and pushed the changes.

FlorianJacob commented on 2018-12-12 01:19

@Jake sounds like sensible plans and ideas. Strictly, A) would need something like an "optmakedepends", wouldn't it, as the package won't generate manpages at runtime when pandoc is installed later, and pandoc is only required during build. This is why I added it to makedepends, but I guess the other options are fine as well.

This is actually my last post from ArchLinux, signing off for NixOS. Goodbye, and happy OOMkilling!

Jake commented on 2018-11-25 21:57

@FlorianJacob: I looked through the Makefile and the manpage generation gets already skipped if no pandoc executable is found. The man pages should also be installed with make install, i tested it and it works fine here, so this manual install call is probably not needed anymore. Some possible solutions for the dependencies:
A: Put pandoc in optdepends and just let it skip the manpage if it is not installed
B: Generate the manpage as maintainer a single time and put it into the repo/sources so everyone can install it without the large makedepends
(C: Use the split package feature and specify two packages earlyoom and earlyoom-man, where the man part only does make earlyoom.1 + make install-man, which depends on pandoc. But this does not really help... makedepends can still only be global, so it would be just a more complicated version of A)

Option A is fine i think, the manpage does not add much info to the output of earlyoom -h so it is not really required in my opinion.

BTW: Instead of defining additional env variables (which could conflict with the makepkg internal ones) it is usually done like make install DESTDIR="${pkgdir}" PREFIX="/usr" SYSTEMDDIR="/usr/lib/systemd". I am happy to help if you add me co-maintainer.