Re: [Emacs-diffs] trunk r116644: ChangeLog entries should be usable without the VCS

[Juanma Barranquero <lekktu@gmail.com>]

On Tue, Mar 4, 2014 at 6:28 PM, Glenn Morris <rgm@gnu.org> wrote:

> The ChangeLog should be understandable to someone sitting there with
> just the source tarball and no VCS.

As previously stated, I disagree because I don't think these people
are the real target. The ChangeLogs already do not match that
description. Some entries, particularly the more recent ones, if
unusually well written, *might* approach that goal, but most won't,
not without looking at the corresponding changesets. I know that for a
fact: I often look at the ChangeLog entry, then to the annotate
listing, and then to the diff to really see what was going on.

Our ChangeLogs, much as I like them and don't want to see them go, are
in that nebulous place between being a list of changes for outside
people (not really, too verbose and rambling*, that's what NEWS is for
and does a much better job), and a complete list of changes for the
source code tinkerer and developer (it isn't near complete enough, for
serious work you need the repo). So the goal that our ChangeLogs
satisfy right now is as a helpful tool to speed locating relevant
changes, or as a (incomplete) reminder of when was something done, and
by whom.

(*Like many of my posts ;-)

If you think that I'm exaggerating and our ChangeLogs are really
pretty complete, I posit that, as an example, while "Follow-up to
revno:XXXX" doesn't say much, "Fix last change." doesn't really say
much more (yes, you can locate the change it refers to, but, how was
it fixed, and what was the problem that required such fixing?). And
there are literally hundreds of such entries in our ChangeLogs.

> Git hashes versus bzr
> revisions, generated ChangeLogs, what gets written in *commit* logs, the
> various ways people access the Emacs sources, it's all irrelevant.

No, I don't think so.

> Anyway, none of this gets us any closer to 24.4.
> So I hope this thread will die, and that people with the time to comment
> will instead look at documenting the remaining issues in NEWS and fixing
> the numerous outstanding bugs.

Writing docs, no thanks. Unable to. As for the rest, my feature-freeze
patches are all bug fixes or typo fixes (which are bugs IMO); so I
tend to agree.

   J