Bug 122529

Summary:	rh-postgresql-test package has broken Makefiles
Product:	Red Hat Enterprise Linux 3	Reporter:	David Lawrence <dkl>
Component:	rh-postgresql	Assignee:	Tom Lane <tgl>
Status:	CLOSED NOTABUG	QA Contact:
Severity:	medium	Docs Contact:
Priority:	medium
Version:	3.0	CC:	hhorak
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2004-05-06 02:36:52 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description David Lawrence 2004-05-05 15:17:00 UTC

Description of problem:
The gmake check functionality in rh-postgresql-test does not work due
to improper paths in the makefiles. This is the error I get when I go
to the regress test directory installed by the test package.

.qa.[root@isqa3 regress]# gmake check
GNUmakefile:16: ../../../src/Makefile.global: No such file or directory
gmake: *** No rule to make target `../../../src/Makefile.global'.  Stop.

After some more playing I was able to get the gmake checks to work by
doing the following with the actual rh-postgresql source rpm after
performing a rpmbuild -bc:

rpm -ivh rh-postgresql-7.3.6-1.src.rpm
rpmbuild -bc rh-postgresql.spec
cd ../BUILD/postgresql-7.3.6/src/test
chown -R postgres.postgres regress/*
cd regress
su - postgres
gmake check

Version-Release number of selected component (if applicable):
rh-postgresql-7.3.6-1

Comment 1 Tom Lane 2004-05-05 18:23:54 UTC

I think the real problem here is that you are not performing the
regression tests in the intended way.  The BUILD tree cannot be used
for this because it is typically root-owned.  The recommended way to
run the regression tests with an RPM build is to install and start the
postmaster, then use the installed regression test tree.  See the
REGRESSION TESTING notes in the README.rpm-dist file.

Comment 2 Tom Lane 2004-05-06 02:36:52 UTC

Closing as pilot error.  (It *would* be nice if the standard PG
regression test procedure worked for the RPMs, but I think it'll take
upstream changes to make that workable at all.)