Bug#923825: multiverse-core: FTBFS randomly (failing tests)

[Santiago Vila]

Package: src:multiverse-core
Version: 0.7.0-4
Severity: important
Tags: ftbfs

Dear maintainer:

I tried to build this package in buster but it failed:

--------------------------------------------------------------------------------
[...]
 debian/rules build-indep
dh build-indep --with javahelper
   dh_update_autotools_config -i
   jh_linkjars -i
   debian/rules override_dh_auto_build
make[1]: Entering directory '/<<PKGBUILDDIR>>'
mkdir /<<PKGBUILDDIR>>/.gradlehome
mkdir: cannot create directory '/<<PKGBUILDDIR>>/.gradlehome': File exists
make[1]: *** [debian/rules:9: override_dh_auto_build] Error 1
make[1]: Leaving directory '/<<PKGBUILDDIR>>'
make: *** [debian/rules:6: build-indep] Error 2
dpkg-buildpackage: error: debian/rules build-indep subprocess returned exit status 2
--------------------------------------------------------------------------------

(Yes, this is funny, please see the full build logs below, my automatic build log
cut script assumes the interesting stuff is always after "debian/rules build-indep",
but not in this case)

The build was made in one of my autobuilders with "dpkg-buildpackage -A".

The failure happens randomly, i.e. sometimes it fails, sometimes it does not,
so I don't have a "recipe" to reproduce it as such, but the "randomness"
is reproducible in my autobuilders (i.e. when I build it a lot of times,
there are always some failed builds).

I've put several failed build logs here for reference:

https://people.debian.org/~sanvila/build-logs/multiverse-core/

If you need a test machine to reproduce this, please contact me
privately and I could provide ssh access to a machine where it happens
(caveat: randomly and maybe with low probability).

If this is really a bug in one of the build-depends, please use reassign and affects,
so that this is still visible in the BTS web page for this package.

Thanks.