Sorry, I have just been prompted to look what happened here and it seems that my message to 432893@ was misaddressed and bounced.
I wrote: Following discussions on debian-dpkg and debian-policy and in #443334, we seem to have concluded that the correct behaviour for dpkg is to not run the postinst in the confusing case. That is, if the package was halfconfigured beforehand, then the postinst should not be run in an attempt to fix it. I think it's still correct that if the package wasn't broken to start with (ie, installed or trig*), a failing dpkg --remove should put the package back the way it started - ie, it should run the postinst and if that's successful the usual rules about the meaning of successful postinst should apply (basically, the package will then be installed or perhaps trig*). Ian. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
