Bug#544046: [Pkg-mailman-hackers] Bug#544046: mailman: postinst find takes far too long

2009-09-25 Thread Thijs Kinkhorst
Hi Paul,

 p...@mail:~$ ps -fe | grep find
 root      6846  5473  9 12:50 pts/0    00:00:39 find /var/lib/mailman/
 -type d -exec chmod g+s {} ; paul     11875 31817  0 13:01 pts/48  
 00:00:00 grep find

 It's still running. Ah, at 13:03 it's done.

 is more efficient, is there an expectation that the permissions may have
 become wrong somehow during the months mailman was running? If so, isn't
 a cron job to take care of this more appropriate? If not, why do this
 upon a simple upgrade from 1:2.1.11-2 - 1:2.1.11-11 ?

Thanks for reporting this. The find is indeed superfluous, if something has 
changed the permissions meanwhile than that should be fixed, but I'm not 
aware of any such processes. Fixing it upon upgrade is quite arbitrary.

The problem is that current mailman packaging is full of legacy that 
accumulated over many years. This find has been in there since the beginning 
of our SVN history. We're removing such legacy bit by bit and have been 
making significant progress since sarge. However, there's still more to do 
and this kind of bug reports really helps to highlight items that can be done 
better.

 I;m just happy that apache was not also upgraded at the same time, as it
 probably would have been stopped during this whole time. That would have
 sucked.

It would - but I wonder if it's actually necessary at all for Apache to be 
down during an upgrade of unrelated packages. Will investigate.


cheers,
Thijs


signature.asc
Description: This is a digitally signed message part.


Bug#544046: [Pkg-mailman-hackers] Bug#544046: mailman: postinst find takes far too long

2009-09-25 Thread Paul Slootman
On Fri 25 Sep 2009, Thijs Kinkhorst wrote:
 
  I;m just happy that apache was not also upgraded at the same time, as it
  probably would have been stopped during this whole time. That would have
  sucked.
 
 It would - but I wonder if it's actually necessary at all for Apache to be 
 down during an upgrade of unrelated packages. Will investigate.

If you're upgrading multipla packages at once, it's not uncommon for a
number of daemons belonging to those packages to be stopped, then those
packages get configured, and then all the daemons are started again...
So, yes, it's entirely possible that apache may be down during the
upgrade of an unrelated package.

That's the point I was trying to make :-)



Paul



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org