Package Details: efl-git 1.26.99.66907.g7743b17dba-1

Git Clone URL: https://aur.archlinux.org/efl-git.git (read-only, click to copy)
Package Base: efl-git
Description: Enlightenment Foundation Libraries - Development version
Upstream URL: http://www.enlightenment.org
Licenses: GPL2, custom, BSD, MIT, LGPL2.1
Conflicts: efl, elementary, elementary-git, elementary_test, elementary_test-git, evas_generic_loaders, evas_generic_loaders-git
Provides: efl, elementary, elementary-git, evas_generic_loaders
Submitter: Scimmia
Maintainer: raster
Last Packager: raster
Votes: 87
Popularity: 0.000000
First Submitted: 2013-02-15 10:00 (UTC)
Last Updated: 2022-06-13 11:54 (UTC)

Required by (52)

Sources (1)

Latest Comments

1 2 3 4 5 6 .. 37 Next › Last »

raster commented on 2022-11-09 11:18 (UTC)

so this is something makepkg itself does - it runs the pkrvger func if there and modifies the PKGBUILD. i can't help that it loses your pkgrel :( i think you need to pass this to makepkg devs. :)

maderios commented on 2022-11-09 11:10 (UTC)

I could build efl-git-1.26.99.66959.g3267070da9-2 after removing 

pkgver() {
  cd $_pkgname

  local v_ver=$(grep version meson.build | head -1 | sed s/version//g | tr ":'," "   " | awk '{print $1}')

  printf "%s.%s.g%s" "$v_ver" "$(git rev-list --count HEAD)" "$(git rev-parse --short HEAD)"
}

maderios commented on 2022-11-08 18:20 (UTC) (edited on 2022-11-08 18:22 (UTC) by maderios)

It's pkgrel problem. Strangely, after editing PKGBUILD, building process changes 'pkgrel=2' to 'pkgrel=1' My PKGBUILD 

_pkgname=efl
pkgname=$_pkgname-git
pkgver=1.26.99.66959.g3267070da9
pkgrel=2


updating files: 100% (9863/9863), done.
==> Starting pkgver()
==> Updated version: efl-git 1.26.99.66959.g3267070da9-1

raster commented on 2022-11-08 16:47 (UTC)

oh wait - pkgrel or pkgver? pkgver there seems right - 66959 so pkgrel not pkgver? this may have to do with the pkgbuild modifying itself to store the pkgver?

maderios commented on 2022-11-08 14:26 (UTC) (edited on 2022-11-08 14:27 (UTC) by maderios)

What i only do: git stash && git pull --rebase then, I edited PKGBUILD 'pkgrel=2' and save it then 

makepkg -sicC
==> Making package: efl-git 1.26.99.66907.g7743b17dba-2 (Tue 08 Nov 2022 03:23:10 PM CET)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> Retrieving sources...
  -> Updating efl git repo...
==> Validating source files with sha256sums...
    efl ... Skipped
==> Removing existing $srcdir/ directory...
==> Extracting sources...
  -> Creating working copy of efl git repo...
Cloning into 'efl'...
done.
Updating files: 100% (9863/9863), done.
==> Starting pkgver()
==> Updated version: efl-git 1.26.99.66959.g3267070da9-1
==> Removing existing $pkgdir/ directory...
==> Starting build()...
Free Mem: 3055 M, using 4 threads for build
The Meson build system
Version: 0.63.3

But I get version 1 at the end

raster commented on 2022-11-08 13:07 (UTC)

pkgver gets the version from meson.build with

grep version meson.build | head -1 | sed s/version//g | tr ":'," " " | awk '{print $1}'

which would e.g. get you 1.26.99

then it ADDS the number of git commits ever in history so you get an always increasing number ever time a commit goes in so a newer git repo with newer commits will always have a higher number. so it runs

git rev-list --count HEAD

which produces e.g. 66959

unless you've done something odd with the git repo like done a shallow clone of the repo (this is certainly not a default) which gets rid of history then this can't calculate history commit count because you don't have any thus it can't ensure the generated package version of a newer clone is higher than an older one. have you done something like this? force shallow clones?

maderios commented on 2022-11-08 10:42 (UTC)

When building efl-git, enlightenment-git and terminology-get, I get problem with 'pkgver' in PKGBUILD. If I change version 1 -> 2 before building, it always goes back to 1. Is there a clean way to change version?

raster commented on 2022-01-03 18:08 (UTC)

Yes - ccache is already listed as a dependency. I'm not sure what I can do for you to fix that - I suspect your problem is somewhere else...

1 2 3 4 5 6 .. 37 Next › Last »