Herbert Xu <[EMAIL PROTECTED]> writes: > Brian Nelson <[EMAIL PROTECTED]> wrote: >> >> It's pointless to go through this again. Instead, I'll offer a concrete >> example of the confusion this can create (the original submitter asks >> for clarification of how the bug was fixed): >> >> http://bugs.debian.org/188740 > > That's a documentation issue. debian/changelog is not the place for > documenting random features.
Huh? The bug report was a feature request with a patch. The bug was
closed with the description of "New Upstream Release". No indication
was given whether the patch was integrated upstream, or implemented
differently (with a different interface). I don't consider this
information "documentation of random features".
>>> As I have said before, this is incomplete: only bugs that were reported
>>> and identified are listed, and redundant: these changes should be in
>>> the upstream changelog already.
>>
>> I don't see anything particularly helpful in the upstream changelog for
>> the above example.
>
> Please do not generalise the practice of individual upstream authors.
Uhh, I didn't. In fact, *you* are the one trying to dictate what goes
in upstream changelogs, which is utterly pointless. Every upstream is
different, and Debian has absolutely no control what upstream decides to
put in their changelogs. That's why we must standardize[1] our changelog
entries, so that the pertinent information will be available regardless
of what upstream does.
If the upstream authors for your packages already include accurate
descriptions of bug fixes in their changelogs, then great. So you have
to spend a couple extra minutes duplicating the descriptions. Tough.
[1] By which I mean each separate (i.e. not merged) bug closed in a
Debian changelog must get its own changelog entry with a
description. Separate bugs must get separate descriptions.
Unrelated bugs should never share a changelog entry. Or something
like that.
--
Poems... always a sign of pretentious inner turmoil.
pgpf12QOZu7Tt.pgp
Description: PGP signature
