Arch Linux User Repository

thanks @voidzero and @drzee for the fixes, they are implemented now. I omitted a pkgrel bump to avoid unnecessary rebuilds for people with working packages.

For people like me who have an issue with the tests (like in my below comment), the trick is to add -DTESTS_WITH_ASAN=OFF to the cmake line in build function in the pkgbuild.

Digging a bit more into this.

If I go to /postsrsd/src/postsrsd-2.0.10/build/_deps/confuse-src/ and then run autoreconf --force --install to regenerate the files they are correctly configured for automake version 1.17.

Going into the build dir and manually executing make all then succeeds without error.

But when then I then try to build with makepkg -s in the PKGBUILD folder it again complains about version 1.16 missing and the files in /postsrsd/src/postsrsd-2.0.10/build/_deps/confuse-src/ are again generated for automake 1.16

Guess CMAKE files there must be a hardcoded reference to automake 1.16 - I have tried finding it but no luck.

Crudely adding these lines to PKGBUILD after the CMake comand and before make all will allow the package to build:

cd _deps/confuse-src/ 
autoreconf --force --install
cd ../..

Build error. Expects aclocal 1.16 but aclocal was updated with core/automake to 1.17 recently (20th of July).

Or is it actually libConfuse that needs to be build with the newest automake version?

Downgrading to automake 1.16 with: pacman -U /var/cache/pacman/pkg/automake-1.16.5-2-any.pkg.tar.zst allows for successful build.

/home/arch/.cache/yay/postsrsd/src/postsrsd-2.0.10/build/_deps/confuse-src support/missing: line 81: aclocal-1.16: command not found
WARNING: 'aclocal-1.16' is missing on your system.
    You should only need it if you modified 'acinclude.m4' or
    'configure.ac' or m4 files included by 'configure.ac'.
    The 'aclocal' program is part of the GNU Automake package:
    <https://www.gnu.org/software/automake>
    It also requires GNU Autoconf, GNU m4 and Perl in order to run:
    <https://www.gnu.org/software/autoconf>
    <https://www.gnu.org/software/m4/>
    <https://www.perl.org/>
make[3]: *** [Makefile:440: /home/arch/.cache/yay/postsrsd/src/postsrsd-2.0.10/build/_deps/confuse-src/aclocal.m4] Error 127
make[2]: *** [CMakeFiles/ExtConfuse.dir/build.make:87: ExtConfuse-prefix/src/ExtConfuse-stamp/ExtConfuse-build] Error 2
make[1]: *** [CMakeFiles/Makefile2:970: CMakeFiles/ExtConfuse.dir/all] Error 2
make: *** [Makefile:166: all] Error 2`

@voidzero: looks like an upstream issue to me, but might also be related to the limited ressources in the container. Try to build in a normal environment and see if the error persists. I will also try to reproduce when I find time for it.

This package does not build anymore for me. I'm running my mail server in an lxc container. When trying to build this, it fails with tests:

[100%] Linking C executable postsrsd
[100%] Built target postsrsd
==> Starting check()...
Running tests...
Test project «dir»/.cache/paru/clone/postsrsd/src/postsrsd-2.0.9/build
Connected to MAKE jobserver
    Start 1: test_netstring
1/6 Test #1: test_netstring ...................***Failed    0.01 sec
    Start 2: test_sha1
2/6 Test #2: test_sha1 ........................***Failed    0.01 sec
    Start 3: test_util
3/6 Test #3: test_util ........................***Failed    0.01 sec
    Start 4: test_database
4/6 Test #4: test_database ....................***Failed    0.01 sec
    Start 5: test_srs2
5/6 Test #5: test_srs2 ........................***Failed    0.01 sec
    Start 6: test_socketmap
6/6 Test #6: test_socketmap ...................   Passed    0.28 sec

17% tests passed, 5 tests failed out of 6

Total Test time (real) =   0.33 sec

The following tests FAILED:
          1 - test_netstring (Failed)
          2 - test_sha1 (Failed)
          3 - test_util (Failed)
          4 - test_database (Failed)
          5 - test_srs2 (Failed)
Errors while running CTest
Output from these tests are in: «dir»/.cache/paru/clone/postsrsd/src/postsrsd-2.0.9/build/Testing/Temporary/LastTest.log

I've added the output from those tests in this gist

Is this something that can be fixed on the PKGBUILD or should I report it upstream? Thanks

No worries, mate. Such things slip by easily!

That is fair. Built the package in a arch container and it struggled to find make. But yes that makes sense. Apologies for the noise.

@Torxed: Nitpicking back: IIRC, both are part of base-devel, prerequisite of building from AUR at all.

Nitpicking, but I believe gcc and make are also makedepends:

diff --git a/PKGBUILD b/PKGBUILD
index d66b401..df20808 100644
--- a/PKGBUILD
+++ b/PKGBUILD
@@ -4,7 +4,7 @@ pkgrel=1
 pkgdesc="Provides the Sender Rewriting Scheme (SRS) via TCP-based lookup tables for Postfix"
 arch=('i686' 'x86_64' 'armv7h')
 depends=('glibc' 'systemd')
-makedepends=('cmake' 'help2man')
+makedepends=('cmake' 'gcc' 'make' 'help2man')
 backup=("etc/$pkgname/$pkgname.conf" "etc/$pkgname/$pkgname.secret" "etc/$pkgname/$pkgname")    #last entry is legacy from version 1
 url="https://github.com/roehling/$pkgname"
 install=$pkgname.install