Some final remarks I want to make about this issue:
1) I do not trust libdb at all. It had too many annoying and surprising
bugs, like it just silently corrupting databases when multiple
subdatabases were opened for writing at the same time (an issue that
was never documented to not work, the only acknoledgement of it was
when it was fixed).
2) I did look at all documentation back when the bug was filed and I
know that there is no *documented* upgrading issues with those
versions for reprepro. The problem is what other issues are there
that are not documented. (Having a setup (subdatabases) that were
prone to break in the past does not make be very confident those
are properly tested. Not using any of the new-fangled stuff in
active developlement where documented upgrade issues are only
makes me more uneasy because of using modes not that actively looked
at). Given the issues of the past, I do not consider some test-suite
runs of reprepro to have a good chance to catch those issues, but
I'd really have liked to be able to do some tests in production
environment.
3) I do not think anyone will claim that upgrading my production systems
to testing a year before a release is even planed is something that
could have been requested from me. If the request to the library
maintainers to make a backport is "absurd" then I do not see why it
should be less absurd to require me to make it.
4) There was some misunderstanding about the time frame for those
library removals and the plans for the release process.
While every misunderstanding can be cured by the person
misunderstanding things asking more stuff, I really think there
are some lessons learned by future release teams how to avoid some
misunderstandings. Any of the following might have easily avoided the
misunderstanding:
- any hint about the bug submitter that this might be related to
release process (or if there was any mandate from release team).
- any hint from the release team in this bug report (or to me)
that this is a goal for lenny
- any hints in any freeze announcements, on the lists of release
goals or to a post to any mailinglist (including debian-release,
which I also subscribed to several months ago because of this issue),
that there is something planed about those libraries, or even only
that libraries not mentioned anywhere does not mean nothing is planed.
5) I'd really have liked a more descriptive way to contacting me than
just raising severity of some old bug with virtually no comment.
From my point of view that is just like suddenly raising severity
of an year old wishlist bug to package a upstream preview of the next
upstream version with some "we want feature <foo>" message.
Thanks in advance next time,
Bernhard R. Link
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
