Bug 295381

Summary:	dev-db/postgresql-server-8.3.8 HAVE_INT64_TIMESTAMP problem
Product:	Gentoo Linux	Reporter:	Andrey Korolyov <andrey>
Component:	New packages	Assignee:	PgSQL Bugs <pgsql-bugs>
Status:	RESOLVED FIXED
Severity:	minor
Priority:	High
Version:	unspecified
Hardware:	AMD64
OS:	Linux
Whiteboard:
Package list:		Runtime testing required:	---

Description Andrey Korolyov 2009-12-01 22:22:48 UTC

There is one small issue: when i am started to use PITR to slave host using skytools postgresql with Debian Lenny amd64 on slave. Unfortunately replication fails on HAVE_INT64_TIMESTAMP because Gentoo master have been compiled without --enable-integer-datetimes. I think that people who maintains this ebuild subtree should modify ebuilds for new useflag or simply include configure flag to src_compile() section for further compability. I do not have tested newer ebuilds in 8.4 tree but seems that datetimes option also disabled on them.

Reproducible: Always

Steps to Reproduce:
1. Install dev-db/postgresql-8.3.8 on amd64 host
2. Try to do online replication to debian lenny amd64 slave

Actual Results:  
At least PITR fails w/ lenny64 slave.

Comment 1 Patrick Lauer gentoo-dev

2010-01-07 00:31:22 UTC

+  07 Jan 2010; Patrick Lauer <patrick@gentoo.org>                                                                                        
+  +postgresql-base-8.4.2-r1.ebuild, metadata.xml:                                                                                        
+  Changing pg-intdatetime to pg_legacytimestamp useflag since upstream                                                                   
+  changed defaults and deprecated the old behaviour. See #285475 #290535 

This should be fixed in 8.4.2-r1, do we need to fix 8.3 too?

Comment 2 Johan Bergström 2010-01-07 00:38:26 UTC

I don't know what the correct way of fixing this would be, but a USE-flag at least gives the option of fixing it yourself :-)

Comment 3 Patrick Lauer gentoo-dev

2010-01-26 20:35:02 UTC

(In reply to comment #2)
> I don't know what the correct way of fixing this would be, but a USE-flag at
> least gives the option of fixing it yourself :-)
> 
Yeah, useflag added to 8.4.2 to make it foolproof. Hope that fixes it all :)