[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Accepted bwidget 1.6.0-1 (all source)

On Tue, May 27, 2003 at 07:41:34PM -0700, Brian Nelson wrote:

> > "Obvious" is a key word indicating that you need to check your
> > assumptions at the door.  While I will certainly concede that changelogs
> > that spell out the nature of relevant upstream changes are more useful
> > than those which do not, the only argument extended that persuades me
> > it's worth extra effort on my part is that it impacts the work of our
> > long-suffering Security Team.  It disappoints me that anyone would
> > consider this such a serious offense that it justifies prolonged
> > flogging on a public mailing list.

> > While using the changelog to close bugs without explaining what was done
> > or to close bugs that were not fixed by the upload in question should
> > not be tolerated, 

> That was the case in this thread.

In recent threads (I've lost track of which), it's happened more than
once that maintainers have been criticized simultaneously for both types
of changelog entries.  Indeed, the bulk of these threads have been
dedicated to arguing the legitimacy of 'New upstream release; closes: ...'
bugs, *not* to defending maintainers who use the changelog to fix bugs
unrelated to the upload.

> > you're getting upset about bug closings that absolutely *do* include a
> > description of what the maintainer changed in order to fix them:
> > incorporating a "new upstream release".  That this is too succinct for
> > some applications does not make it "changelog abuse".  

> I'll concede this point, though I still think we should encourage
> maintainers to spell out upstream fixes in the changelog.

Encouraging, yes; but, ...

> > Nor, BTW, is it changelog abuse to have forgotten to close a fixed bug
> > in the changelog: your hard-line stance invites maintainers to simply
> > close *all* bugs manually after upload, which makes the changelog a
> > much less useful tool on many levels.

> Hmm, I don't understand.  Every bug that has been fixed should be closed
> in the changelog when possible.  If a maintainer forgets, then close it
> manually with a note mentioning in which version it was fixed.  When did
> I say otherwise?

point being that public ridicule may simply encourage maintainers to
sidestep the issue by not closing bugs in the changelog at all --
particularly when there are differing opinions on what is or is not an
"acceptable" explanation in the changelog.

Steve Langasek
postmodern programmer

Attachment: pgpIskv80j0dG.pgp
Description: PGP signature

Reply to: