software-properties Build failure

Compare your log with the buildlog of the package in the Ubuntu repository, https://launchpadlibrarian.net/462031231/buildlog_ubuntu-bionic-amd64.software-properties_0.96.24.32.12+horizon-os12_BUILDING.txt.gz

test_ppa_info_from_lp (tests.test_lp.LaunchpadPPATestCase) ... skipped 'skipping online tests unless TEST_ONLINE environment variable is set'

vs.

test_ppa_info_from_lp (tests.test_lp.LaunchpadPPATestCase) ... E: Caught signal ‘Terminated’: terminating immediately

It seems that you have to disable certain tests (I have no idea how to do that).

Oops, wrong link. The link to the Ubuntu version should have been

https://launchpad.net/ubuntu/+source/software-properties/0.96.24.32.12/+build/18213453/+files/buildlog_ubuntu-bionic-amd64.software-properties_0.96.24.32.12_BUILDING.txt.gz

That certainly isn't the original software without any modifications (the packaging is part of the software for this purpose). I downloaded both software-properties 0.96.24.32.12 and software-properties 0.96.24.32.12+horizon-os12, ran debdiff (from the devscripts package) on both .dsc files; this showed a large number of differences, including things like a build-dependency on dh-virtualenv which seems rather unlikely to be appropriate, and all the checkinstall stuff. You should go through the diff and clean it up, I think - it's always best to exclude your local changes before asking for support.

That said, I did some debugging. Here's what I found:

 * Due to output buffering somewhere, the test output that you see in the build log is extremely misleading about where the hang is. Although it looks like the hang is in tests.test_lp.LaunchpadPPATestCase.test_ppa_info_from_lp, that test has actually finished (or in fact been skipped), and it's the next test along that hangs. (I discovered this by inserting a call to strace.)
 * Your change in 0.96.24.32.12+horizon-os13 to remove dbus-run-session is definitely wrong. Without dbus-run-session, the tests hang trying to talk to D-Bus.
 * Your 0.96.24.32.12+horizon-os12 version built cleanly when I copied that source package version into a separate PPA configured to install build-dependencies from your PPA.

So I think this may just be some unreliability in software-properties' test suite. If you're interested in chasing it down then you should be able to do that in a local sbuild(1) instance; but failing that, retrying the failed build of 0.96.24.32.12+horizon-os12 would probably have worked, and you should revert the change you made in 0.96.24.32.12+horizon-os13. I can't see any problem with Launchpad as such here; this all seems to be at most some problems with the package itself.

Note that Manfred's comment above is definitely *incorrect*, although I went down the same line of thought at first, having been misled by the confusing test output buffering.

Sorry if I created confusion with my attempt to help.
I will refrain from commenting questions where my level of expertise isn't sufficient.

My problem has been solved, the packages have wrong modifications that I didn't accidentally notice!