Bug#277652: Confirmed in postgresql-common, too

[Martin Pitt]

Hi all,

this has been confirmed in postgresql-common as well. So this is a
serious regression compared to Etch which ought to be fixed for Lenny.

Moving missingok above the glob is *not* the solution, it is at most
a temporary local workaround. We are not going to change a million
conffiles spread over all Debian packages and break backwards
compatibility just because. It worked fine in earlier logrotate
versions, so it should somehow work today, too?

Thanks,

Martin

-- 
Martin Pitt| http://www.piware.de
Ubuntu Developer (www.ubuntu.com)  | Debian Developer  (www.debian.org)


signature.asc
Description: Digital signature

Bug#277652: Confirmed in postgresql-common, too

[Paul Martin]

On Mon, Aug 11, 2008 at 12:48:42PM +0200, Martin Pitt wrote:
 this has been confirmed in postgresql-common as well. So this is a
 serious regression compared to Etch which ought to be fixed for Lenny.

It's not a regression compared with Etch as the bug predates Sarge, and
persists through *all* previous versions of logrotate.

Each individual file in /etc/logrotate.d inherits the settings from
/etc/logrotate.conf, and nothing else. Once the included file has been
parsed, the settings revert to those in /etc/logrotate.conf.

In other words, putting missingok at the start of a file in
/etc/logrotate.d/ won't affect the behaviour of logrotate for any other
package's logs.

This problem *may* have been addressed upstream, but I'm loath to introduce
changes this close to release.

-- 
Paul Martin [EMAIL PROTECTED]


signature.asc
Description: Digital signature