RPM corruption bug status?

[Colin Charles]

On Fri, 2004-06-11 at 10:00, Greg Trounson wrote:

> There was a bug in RPM which, upon trying to install packages when /usr 
> is mounted read only, results in a corrupted RPM database.

This was closed not a bug, because its really an RFE - the question
really lies is, why is there a point in installing a package if no files
are put on the filesystem if the /usr is RO ?

> Of course, it shouldn't be able to actually install any RPMS when /usr 
> is read only, but it should fail gracefully.

Granted, it should exit gracefully - but its a case of priorities. Lots
of other important bugs to fix

> It's marked in bugzilla (#119185) as 'closed' and 'notabug' although it 
> clearly is a rather serious bug.

If there's consensus as to why RPM should do RO with /usr, please add to
the bug report, and jbj will definitely take a look at it if there's
enough consensus
-- 
Colin Charles, byte at aeon.com.my
http://www.bytebot.net/