Re: pg_upgrade regression test litters the source tree with log files

On Tue, Jan 8, 2013 at 04:11:41PM -0500, Peter Eisentraut wrote:
> On 1/8/13 4:04 PM, Bruce Momjian wrote:
> > On Tue, Jan 8, 2013 at 01:08:44PM -0500, Tom Lane wrote:
> >> In a tree in which I previously ran "make check" in contrib/pg_upgrade:
> >>
> >> $ make -s distclean
> >> $ git status
> >> # On branch master
> >> # Untracked files:
> >> # (use "git add <file>..." to include in what will be committed)
> >> #
> >> # contrib/pg_upgrade/pg_upgrade_dump_1.log
> >> # contrib/pg_upgrade/pg_upgrade_dump_12912.log
> >> # contrib/pg_upgrade/pg_upgrade_dump_16384.log
> >> nothing added to commit but untracked files present (use "git add" to track)
> >>
> >> Not sure how long this has been happening.
> >
> > Those look like files left over from a failed upgrade, or you used
> > --retain. Does that make sense? Because they are tracked by oid, it
> > is possible a later successful upgrade would not remove all those files,
> > bit it should remove contrib/pg_upgrade/pg_upgrade_dump_1.log because it
> > is "1".
>
> I think this came in with the pg_upgrade --jobs option.

Yes, it was part of the split to allow creation of per-database SQL
files, but pg_upgrade always created files in the current directory ---
there are just more of them now.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +