Re: Re: SUG: Automatic RPM database verification and repair

[Tony Nelson <tonynelson georgeanelson com>]

At 8:40 AM -0500 11/29/06, Jeff Johnson wrote:
>On 11/28/06, Tony Nelson <tonynelson georgeanelson com> wrote:
>>
>> Rpm does something odd when it is given a non-existent or non-full path to
>> a RPM database: it creates a new RPM database there.  If it is given a full
>> path (or a proper relative path) to a broken RPM database it will report
>> the errors.  For proper operation, the script requires a full path to a
>> real RPM database.  Arguably this is a deficiency in rpm's "--verifydb"
>> option.
>
>RPM creates the dbpath lazily if it does not exist. Whether that is a feature
>or a bug depends on who you are talking to.
>
>The current behavior (creating the path lazily) makes creating chroots
>easier, one does not have to do the explicit
>    mkdir -p /path/to/chroot/var/lib/rpm
>
>OTOH, invocations of rpm with random dbpaths can/will litter the end-point
>with lazily created empty directories.

Yes, as will doing a rebuild on a missing database.

RPM should mention this creation when doing a "--verifydb" or a "--rebuilddb".
-- 
____________________________________________________________________
TonyN.:'    The Great Writ     <mailto:tonynelson georgeanelson com>
      '      is no more.             <http://www.georgeanelson.com/>