Confusion over t64 migration

[John Goerzen <jgoerzen@complete.org>]

Hi everyone,

Thanks to all that have put so much time and thought into the time_t
migration.  I am late to this party and am trying to figure my way
through it.

Quite a few of my packages are marked for removal from testing because
time_t migration bugs have been filed with severity serious.  Some of
these are filed against my own packages, but most are filed against
other packages which are dependencies of ones I maintain.

I am an uploader for gensio, which had a bug #1062097 reported for this
issue.  Like the others, it was reported as severity serious, tagged
found in the version in unstable.  This, of course, prevents migrations
to testing and also leads to removals from testing.

I uploaded a package with the enclosed patch quickly, but then was told
that was not the right thing to do.  I then uploaded a new version of
gensio with the change removed...  but again (and I'm not even sure how
this is the case, since the bug was resolved) it is marked for
autoremoval from testing.

So at the moment, I am unclear why there are bugs filed with severity
serious that apparently cannot be fixed.  Shouldn't they be normal with
a tag wontfix until the relevant dpkg changes are in unstable?

To put it another way, I'm not seeing why we are reporting RC bugs
against a bunch of packages before it is possible to fix them.

A key use case for me is uploading to bookworm-backports.  Of course,
that requires being in testing first.  What will happen with time_t in
bookworm-backports?

Thanks!

- John