Re: pg_upgrade regression test litters the source tree with log files - Mailing list pgsql-hackers

From Bruce Momjian
Subject Re: pg_upgrade regression test litters the source tree with log files
Date
Msg-id 20130108211253.GC8789@momjian.us
Whole thread Raw
In response to Re: pg_upgrade regression test litters the source tree with log files  (Peter Eisentraut <peter_e@gmx.net>)
List pgsql-hackers
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@momjian.us>        http://momjian.us EnterpriseDB
http://enterprisedb.com
 + It's impossible for everything to be true. +



pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: pg_upgrade regression test litters the source tree with log files
Next
From: Andres Freund
Date:
Subject: Re: [PATCH 2/5] Make relpathbackend return a statically result instead of palloc()'ing it