Обсуждение: contrib/earthdistance missing regression test files
The Makefile for contrib/earthdistance indicates that there should be a regression test, but the files seem to be missing from CVS. The change to the Makefile was made here: http://developer.postgresql.org/cvsweb.cgi/contrib/earthdistance/Makefile.diff?r1=1.11&r2=1.12 Was the Makefile change a mistake, or are there files missing? Joe
On Tue, Sep 24, 2002 at 10:43:51 -0700, Joe Conway <mail@joeconway.com> wrote: > The Makefile for contrib/earthdistance indicates that there should be a > regression test, but the files seem to be missing from CVS. The change to > the > Makefile was made here: > > http://developer.postgresql.org/cvsweb.cgi/contrib/earthdistance/Makefile.diff?r1=1.11&r2=1.12 > > Was the Makefile change a mistake, or are there files missing? There is supposed to be a regression test. I may have forgotten to use -N or -r on the diff. If it is confirmed that the files needed for the regression test didn't make it into the submitted diff file, I can send in a diff versus current cvs.
On Tue, Sep 24, 2002 at 15:02:20 -0500, Bruno Wolff III <bruno@wolff.to> wrote: > > There is supposed to be a regression test. I may have forgotten to use > -N or -r on the diff. If it is confirmed that the files needed for the > regression test didn't make it into the submitted diff file, I can send > in a diff versus current cvs. I still have a copy of the diff file (at least I think it is the one I sent in) and it has the regression sql and output files defined.
Bruno Wolff III wrote: > On Tue, Sep 24, 2002 at 15:02:20 -0500, > Bruno Wolff III <bruno@wolff.to> wrote: > > > > There is supposed to be a regression test. I may have forgotten to use > > -N or -r on the diff. If it is confirmed that the files needed for the > > regression test didn't make it into the submitted diff file, I can send > > in a diff versus current cvs. > > I still have a copy of the diff file (at least I think it is the one I > sent in) and it has the regression sql and output files defined. Yep, I missed adding earthdistance.out. Is that the only file. I usually do a 'gmake distclean' and 'cvs update' after a batch of patches to see that there aren't any new files in my CVS tree. I missed it this time. Does that fix the problem? -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001+ If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania19073
On Tue, Sep 24, 2002 at 16:10:19 -0400, Bruce Momjian <pgman@candle.pha.pa.us> wrote: > Bruno Wolff III wrote: > > On Tue, Sep 24, 2002 at 15:02:20 -0500, > > Bruno Wolff III <bruno@wolff.to> wrote: > > > > > > There is supposed to be a regression test. I may have forgotten to use > > > -N or -r on the diff. If it is confirmed that the files needed for the > > > regression test didn't make it into the submitted diff file, I can send > > > in a diff versus current cvs. > > > > I still have a copy of the diff file (at least I think it is the one I > > sent in) and it has the regression sql and output files defined. > > Yep, I missed adding earthdistance.out. Is that the only file. No. The new files are: expected/earthdistance.out (which you metion above) sql/earthdistance.sql
Bruno Wolff III wrote:
> On Tue, Sep 24, 2002 at 16:10:19 -0400,
> Bruce Momjian <pgman@candle.pha.pa.us> wrote:
> > Bruno Wolff III wrote:
> > > On Tue, Sep 24, 2002 at 15:02:20 -0500,
> > > Bruno Wolff III <bruno@wolff.to> wrote:
> > > >
> > > > There is supposed to be a regression test. I may have forgotten to use
> > > > -N or -r on the diff. If it is confirmed that the files needed for the
> > > > regression test didn't make it into the submitted diff file, I can send
> > > > in a diff versus current cvs.
> > >
> > > I still have a copy of the diff file (at least I think it is the one I
> > > sent in) and it has the regression sql and output files defined.
> >
> > Yep, I missed adding earthdistance.out. Is that the only file.
>
> No. The new files are:
> expected/earthdistance.out (which you metion above)
> sql/earthdistance.sql
OK, here's what I see now in CVS:
#$ pwd/pgtop/contrib/earthdistance#$ lfCVS/ README.earthdistance earthdistance.outMakefile
earthdistance.c earthdistance.sql.in
What should be changed?
-- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610)
359-1001+ If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square,
Pennsylvania19073
Bruce Momjian <pgman@candle.pha.pa.us> writes:
>> No. The new files are:
>> expected/earthdistance.out (which you metion above)
>> sql/earthdistance.sql
> OK, here's what I see now in CVS:
> #$ pwd
> /pgtop/contrib/earthdistance
> #$ lf
> CVS/ README.earthdistance earthdistance.out
> Makefile earthdistance.c earthdistance.sql.in
> What should be changed?
The earthdistance.out file should be in an expected/ subdirectory, not
directly in the contrib/earthdistance directory. Also, there is a
missing regression input script file earthdistance.sql (this is not
related to earthdistance.sql.in), which should be in a sql/
subdirectory.
regards, tom lane
Tom Lane wrote: > Bruce Momjian <pgman@candle.pha.pa.us> writes: > >> No. The new files are: > >> expected/earthdistance.out (which you metion above) > >> sql/earthdistance.sql > > > OK, here's what I see now in CVS: > > > #$ pwd > > /pgtop/contrib/earthdistance > > #$ lf > > CVS/ README.earthdistance earthdistance.out > > Makefile earthdistance.c earthdistance.sql.in > > > What should be changed? > > The earthdistance.out file should be in an expected/ subdirectory, not > directly in the contrib/earthdistance directory. Also, there is a > missing regression input script file earthdistance.sql (this is not > related to earthdistance.sql.in), which should be in a sql/ > subdirectory. OK, done. I thought the earthdistance.sql file was derived from earthdistance.sql.in, and I thought the out was just a test file. I got them fixed now, in their proper directory. How do I run the regression tests for /contrib stuff? -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001+ If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania19073
Bruce Momjian <pgman@candle.pha.pa.us> writes:
> How do I run the regression tests for /contrib stuff?
make
make install
make installcheck
AFAICT, earthdistance is nowhere near passing yet :-(. It looks to
me like the regression test is depending on the cube-based features
that we decided to hold off for 7.4. Bruno, is that right?
regards, tom lane
On Tue, Sep 24, 2002 at 23:57:29 -0400, Tom Lane <tgl@sss.pgh.pa.us> wrote: > Bruce Momjian <pgman@candle.pha.pa.us> writes: > > How do I run the regression tests for /contrib stuff? > > make > make install > make installcheck > > AFAICT, earthdistance is nowhere near passing yet :-(. It looks to > me like the regression test is depending on the cube-based features > that we decided to hold off for 7.4. Bruno, is that right? It shouldn't be. When I resubmitted the patch I intended to take out all of the cube related tests. If there is a reference to cube in there it is by mistake.
> > AFAICT, earthdistance is nowhere near passing yet :-(. It looks to > > me like the regression test is depending on the cube-based features > > that we decided to hold off for 7.4. Bruno, is that right? > > It shouldn't be. When I resubmitted the patch I intended to take out > all of the cube related tests. If there is a reference to cube in there > it is by mistake. I took a look at the diff file I submitted and the only reference to cube in the regression test was in a comment I didn't change after removing the tests for the cube based distance stuff.
OK, I reinstalled the proper earthdistance.out/sql files and it passes regession now. Sorry for the mistake. --------------------------------------------------------------------------- Bruno Wolff III wrote: > > > AFAICT, earthdistance is nowhere near passing yet :-(. It looks to > > > me like the regression test is depending on the cube-based features > > > that we decided to hold off for 7.4. Bruno, is that right? > > > > It shouldn't be. When I resubmitted the patch I intended to take out > > all of the cube related tests. If there is a reference to cube in there > > it is by mistake. > > I took a look at the diff file I submitted and the only reference to > cube in the regression test was in a comment I didn't change after > removing the tests for the cube based distance stuff. > -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001+ If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania19073
Bruce Momjian <pgman@candle.pha.pa.us> writes:
> OK, I reinstalled the proper earthdistance.out/sql files and it passes
> regession now. Sorry for the mistake.
Looks good here too. Thanks.
regards, tom lane