Re: [Emacs-diffs] master 5515625: ; ChangeLog.2 fixes

[Juanma Barranquero <lekktu@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 2792 bytes --]

On Sun, Oct 25, 2015 at 8:21 PM, Eli Zaretskii <eliz@gnu.org> wrote:

> That's correct.  Personally, I think the commit messages themselves
> could have the period in this special case, but that's just MO.

I agree.

> This is borderline, and you might as well leave those untouched.

The ones I change tend to be like

        Use `unless' to have one fewer `not'
        * lisp/vc/vc-git.el (vc-git-resolve-when-done): Use `unless' to
        have one fewer `not'.

or

        Fix customization type of `even-window-sizes'.
        * lisp/window.el (even-window-sizes): Fix customization type.

where, really, the header doesn't add anything of value (to the ChangeLog).

> These are issues of style, so there are no hard rules.  Back when we
> maintained ChangeLog files by hand the style was not really uniform,
> either.

Yes, I know. I didn't try then, and do not try now, to have absolute
consistency, just to increment it a little ;-)

>   +       * lisp/character-fold.el: Many improvements.
>           (character-fold-search-forward, character-fold-search-backward):
>   -       New command

> The "many improvements" part could simply go away, it doesn't add any
> useful information.

I'd tend to agree, but I'm conservative in removing other people's words
unless they are really redundant (as in the cases above).

> I think the commit message should have been as corrected to begin
> with, and a header line should say something else, without being
> formatted as an entry.

Certainly. The best messages IMHO are those who do *not* format the header
line as an entry. They are usually more informative and less redundant.

> This is okay, but when I see such changes, I always ask myself whether
> it's worth the trouble.  Your call.

Some of these are because I use column markers to see which entries go over
75 or 80 columns, and some because I agree with Stefan that having dangling
words is ugly. Which doesn't mean that I don't leave some of these, when
the alternative is equally ugly, of course...

>    2015-10-20  Dmitry Gutov  <dgutov@yandex.ru>
>
>   -       Don't declare vc-exec-after anymore
>   -
>           * lisp/vc/vc-svn.el:
>           * lisp/vc/vc-mtn.el:
>           * lisp/vc/vc-hg.el:
>           * lisp/vc/vc-cvs.el:
>           * lisp/vc/vc-git.el:
>   -       * lisp/vc/vc-bzr.el: Don't declare vc-exec-after anymore.  Its
>   -       usages have been replaced with vc-run-delayed.
>   +       * lisp/vc/vc-bzr.el: Don't declare vc-exec-after anymore.
>   +       Its usages have been replaced with vc-run-delayed.
>
> Not sure why the header line was deleted, it looks OK to me.

Borderline, but the header is duplicated word-for-word in the entry. I
think I would've left it if the file list were quite long.

Thanks for your comments.

     J

[-- Attachment #2: Type: text/html, Size: 3784 bytes --]