pgsql/src (Makefile.global.in)

Peter Eisentraut - PostgreSQL wrote:

Date: Sunday, October 8, 2000 @ 17:13:27
Author: petere

Update of /home/projects/pgsql/cvsroot/pgsql/src
from hub.org:/home/projects/pgsql/tmp/cvs-serv69058/src

Modified Files:
Makefile.global.in

----------------------------- Log Message -----------------------------

Append "/postgresql" to (certain) installation subdirectories when
installing into a shared location. Also Makefile.global organizational
cleanup.

make all runcheck
make -C ../../../contrib/spi REFINT_VERBOSE=1 refint.so autoinc.so
make[1]: Entering directory
`/usr/srn/cvs/workspace/tags/reindex/contrib/spi'
make[1]: `refint.so' is up to date.
make[1]: `autoinc.so' is up to date.
make[1]: Leaving directory
`/usr/srn/cvs/workspace/tags/reindex/contrib/spi'
MULTIBYTE=EUC_JP;export MULTIBYTE; \
MAKE=make;export MAKE; \
/bin/sh ./run_check.sh i686-pc-linux-gnu
=============== Removing old ./tmp_check directory ... ================
=============== Create ./tmp_check directory ================
=============== Installing new build into ./tmp_check ================
=============== Initializing check database instance ================

ERROR: Check initdb failed - cannot continue
Please examine
/usr/srn/cvs/workspace/tags/reindex/src/test/regress/tmp_check/log/initdb.log
for the reason.

[initdb.log]
Running with noclean mode on. Mistakes will not be cleaned up.
initdb does not find the file
'/usr/srn/cvs/workspace/tags/reindex/src/test/regr
ess/tmp_check/share/template1.bki'.
This means you have a corrupted installation or identified the
wrong directory with the -L invocation option.

Hiroshi Inoue

Hiroshi Inoue <Inoue@tpf.co.jp> writes:

ERROR: Check initdb failed - cannot continue
Please examine
/usr/srn/cvs/workspace/tags/reindex/src/test/regress/tmp_check/log/initdb.log
for the reason.

[initdb.log]
Running with noclean mode on. Mistakes will not be cleaned up.
initdb does not find the file
'/usr/srn/cvs/workspace/tags/reindex/src/test/regr
ess/tmp_check/share/template1.bki'.

FWIW, current sources (pulled about 23:30 GMT 10-19) pass regress tests
for me, either make runtest or make runcheck style. After fixing a
problem with not building libpq with -fPIC, that is. Not sure why you
might be seeing a problem with template1.bki...

regards, tom lane

Tom Lane wrote:

Hiroshi Inoue <Inoue@tpf.co.jp> writes:

ERROR: Check initdb failed - cannot continue
Please examine
/usr/srn/cvs/workspace/tags/reindex/src/test/regress/tmp_check/log/initdb.log
for the reason.

[initdb.log]
Running with noclean mode on. Mistakes will not be cleaned up.
initdb does not find the file
'/usr/srn/cvs/workspace/tags/reindex/src/test/regr
ess/tmp_check/share/template1.bki'.

FWIW, current sources (pulled about 23:30 GMT 10-19) pass regress tests
for me, either make runtest or make runcheck style. After fixing a
problem with not building libpq with -fPIC, that is. Not sure why you
might be seeing a problem with template1.bki...

Sorry to trouble you. I posted my previous mail to Peter.
Maybe I would be the first and last man who complains about
this. My working directory name doesn't contain neither pgsql
nor postgres. My DATALIB is .../test/regress/tmp_check/
postgresql/share due to Peter's change but runcheck.sh
^^^^^^^^^^^^
doesn't know it.
Unfortunately his bugs seems to have loved me.
I was the first man who met override CFLAGS bug also.

Regards.
Hiroshi Inoue

Hiroshi Inoue writes:

Sorry to trouble you. I posted my previous mail to Peter.
Maybe I would be the first and last man who complains about
this. My working directory name doesn't contain neither pgsql
nor postgres. My DATALIB is .../test/regress/tmp_check/
postgresql/share due to Peter's change but runcheck.sh
^^^^^^^^^^^^
doesn't know it.

Hmm... This is a twofold runcheck.sh bug which was just waiting to
happen:

1. If you override the installation directories, either using an explicit
--xxxdir option, or implicitly as done here, runcheck.sh's hardwired
installation layout knowlegde fails.

2. make install prefix=somewhere doesn't work either if you use configure
--xxxdir options.

That's why runcheck.sh has been replaced. You can run the regression
tests with "make check".

If you're attached to the old driver, perhaps we could trick it by
renaming "tmp_check" to "tmp_check_pgsql" or so.

--
Peter Eisentraut peter_e@gmx.net http://yi.org/peter-e/

Peter Eisentraut wrote:

Hiroshi Inoue writes:

Sorry to trouble you. I posted my previous mail to Peter.
Maybe I would be the first and last man who complains about
this. My working directory name doesn't contain neither pgsql
nor postgres. My DATALIB is .../test/regress/tmp_check/
postgresql/share due to Peter's change but runcheck.sh
^^^^^^^^^^^^
doesn't know it.

Hmm... This is a twofold runcheck.sh bug which was just waiting to
happen:

1. If you override the installation directories, either using an explicit
--xxxdir option, or implicitly as done here, runcheck.sh's hardwired
installation layout knowlegde fails.

2. make install prefix=somewhere doesn't work either if you use configure
--xxxdir options.

That's why runcheck.sh has been replaced. You can run the regression
tests with "make check".

OK,"make check" seems to work.
If runcheck.sh is not preferable any longer,shouldn't we
disable it ?

Regards.
Hiroshi Inoue