tags 672158 + pending thanks On Tue, 2012-05-08 at 23:48 +0200, Andreas Beckmann wrote: > Looking at the postrm script, I think it is an error to have a condition > on db_input (which may 'fail' if the question won't be presented). > db_input should be allowed to fail (at least with an error code that > indicated 'won't be shown', iirc that is a return value of 30) and > processing should continue in that case (db_go, db_get, ...).
Thanks for reporting this and sorry for not getting back to you sooner. I've grown a bit scared of preseeding because in other packages I've seen a lot of troubles with it in combination with normal handling of configuration changes. The approach in the postrm was recommended (in 2003) by Joey Hess in a review he did of the debconf handling: https://lists.debian.org/debian-devel/2003/04/msg01296.html The idea was only to remove the user's data if the user saw the message. However, since the default value is false and at least MySQL seems to do the same I'm not too worried about changing it back. -- -- arthur - adej...@debian.org - http://people.debian.org/~adejong --
signature.asc
Description: This is a digitally signed message part
