Could someone explain to me the rationale of the Git repository
organization? Over the past few months, I have submitted a few bug
fixes and improvements for pgbouncer. I wanted to check into whether
they had been committed or released.
There are 58 pgbouncer.spec files in the git repository!
rpm/$OS/$PGVERSION/pgbouncer/$OSVERSION/pgbouncer.spec
This doesn't make any sense, because pgbouncer doesn't care about the
PostgreSQL version, and a well-written spec file also shouldn't care
that much about the OS version.
I think my changes have only been merged into the 9.4 section (and
presumably later "branched" into 9.5). Why? It's not clear why some
commits but not others go into only the latest $PGVERSION subdirectory.
If I want to submit changes, where are they supposed to go?
What's totally crazy, however, is that when I download an RPM from the
web site, the file names don't include the $PGVERSION. So there are RPM
files out there that are named identically, have similar timestamps, but
different features and bug fixes.
I don't know what the build infrastructure is or what file layout it
requires, but I think for non-extension modules, the layout ought to be
simply
rpm/redhat/pgbouncer/pgbouncer.spec