On 18/11/2022 19.47, Sean Whitton wrote:
I can't reproduce this locally, and looking at the package tracker for a
few of those packages you've listed, piuparts is now passing if it's
been re-run recently. So I believe this one is fixed.
... except that it's still showing up in failing autopkgtests, e.g. for
emacs-deferred.
Since most extensions don't have a hard dependency on emacs (which is
good, otherwise it would have to be an or-ed list of all supported
alternatives), this bug does not show up at regular piuparts tests.
I have set of local piuparts tests running on packages matching
Depends: *emacs* # usually matches emacsen-common
that installs emacs (or xemacs or both) prior to the package to be tested.
There were enough errors in the past only showing up in bigger
installations including some emacs frontent. (Often --install-recommends
tests). That were the times when we still had a bunch of emacsXX in the
archive and I took the liberty to test all their combinations :-)
Now that we have an installable emacs in sid again, I'll try to look
into the dependency difference between sid and testing to see what
changed to cuase this bug.
Andreas
