Tom Lane wrote:
> Bruce Momjian <pgman@candle.pha.pa.us> writes:
> >> Does anyone know when this problem was added. I don't see any
> >> significant changes in GNUmakefile.
>
> AFAICT, the results directory hasn't been touched in 21 months.
> Are you sure you haven't changed your own CVS setup or arguments?
I looked at that. My cvs hasn't changed, my OS hasn't changed. I
deleted my entire tree and did another checkout, but that didn't help.
> > The -d brings in any directories that may have been created since the
> > last checkout. That is causing problems when the regress/results
> > directory has been deleted and recreated.
>
> I use cvs update -d -P myself, and I do *not* see it creating the
> results directory.
The problem is that if the results directory exists, the update fails
because there is no /CVS directory in there.
Here is a sample session. At the start regress/results does not exist:
---------------------------------------------------------------------------
$ cvs update -d -P pgsql
? pgsql/config.log
...
$ cd /pg/test/regress/
$ mkdir results
$ cd -
/pgcvs
$ cvs update -d -P pgsql
? pgsql/config.log
...
? pgsql/src/test/regress/results
cvs update: in directory pgsql/src/test/regress/results:
cvs update: cannot open CVS/Entries for reading: No such file or directory
---------------------------------------------------------------------------
It seems to be failing because there is no CVS directory in results.
This is CVS:
Concurrent Versions System (CVS) 1.10.3 (client/server)
-- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610)
853-3000+ If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill,
Pennsylvania19026