pcp
[Top] [All Lists]

Re: [pcp] rpm migrate_tempdirs() problem on SuSE

To: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Subject: Re: [pcp] rpm migrate_tempdirs() problem on SuSE
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Mon, 1 Apr 2013 21:12:36 -0400 (EDT)
Cc: David Disseldorp <ddiss@xxxxxxx>, PCP Mailing List <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <5154052F.70302@xxxxxxxxxxxxxxxx>
References: <51528C9C.5070403@xxxxxxxxxxxxxxxx> <1955776917.25463227.1364365934207.JavaMail.root@xxxxxxxxxx> <20130327120037.7d6455bb@xxxxxxxxxx> <5154052F.70302@xxxxxxxxxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: UyhFZsJ7cBHDCQqDJR1wLWZht1PIeA==
Thread-topic: rpm migrate_tempdirs() problem on SuSE

----- Original Message -----
> On 27/03/13 22:00, David Disseldorp wrote:
> > ...
> > It looks like --no-clobber was added in 2009 with coreutils 7.1 - it's
> > present on SLE11SP2 but missing from SLE11SP1.
> > ...
> > Sounds fine to me.
> 
> I've pushed a commit into my try after extensive testing of the proposed
> change on SLES11SP1.
> 
> Nathan, in the testing I was tricked by this predicate
> 
>       test -d "$d" -a -k "$d" || continue
> 
> I don't understand why the migration is only done for old dirs that have
> the "sticky" bit set ... so what's the reason behind the -a -k part of
> the predicate?
> 

We should only be moving tmpdirs we created, which always have this set,
so figured I'd check for that (wouldn't want anyone playing games with
symlinks, and getting us to move arbitrary files around, etc).

cheers.

--
Nathan

<Prev in Thread] Current Thread [Next in Thread>
  • Re: [pcp] rpm migrate_tempdirs() problem on SuSE, Nathan Scott <=