M-n in diff mode: use to a better spot?

M-n in diff mode: use to a better spot?

[Kai Grossjohann]

For composing my ChangeLog entries, I like to do C-x v = on the
current file (or `=' from PCL-CVS), then I iterate over the hunks and
describe each change listed there.

However, it fairly often happens that the beginning of the hunk is
not in the method I would like to describe.  So what I do is to
manually move point down to the line starting with + or -, so that
C-x 4 a will automatically fill in the right method name for me.

WIBNI there was some automatic way of moving to the right spot in the
diff buffer, so that C-x 4 a can provide a better default?

Kai

Re: M-n in diff mode: use to a better spot?

[Richard Stallman]

    WIBNI there was some automatic way of moving to the right spot in the
    diff buffer, so that C-x 4 a can provide a better default?

Are you suggesting that if point is at the start of the hunk, C-x 4 a
should operate on the file position of first change in the hunk, rather
than on the file position corresponding to point?  That would have the
effect you want, and I think it would not cause problems.

Re: M-n in diff mode: use to a better spot?

[Kai Grossjohann]

Richard Stallman <rms@gnu.org> writes:

>     WIBNI there was some automatic way of moving to the right spot in the
>     diff buffer, so that C-x 4 a can provide a better default?
>
> Are you suggesting that if point is at the start of the hunk, C-x 4 a
> should operate on the file position of first change in the hunk, rather
> than on the file position corresponding to point?  That would have the
> effect you want, and I think it would not cause problems.

That would be fine.  Another possibility is for M-n to move to the
first +/-/! line inside the hunk instead of to the start of the
hunk.  Yet another possibility is to make a new command which moves
to the +/-/! line inside the hunk.  I would be happy with any one of
these.

Kai

Re: M-n in diff mode: use to a better spot?

[Richard Stallman]

    That would be fine.  Another possibility is for M-n to move to the
    first +/-/! line inside the hunk instead of to the start of the
    hunk.

That would look strange, I think.  Especially if it moves to the
"first" + line, in the second part of the hunk, because that comes
earlier in the source than the first - line in the first part of the
hunk.

So I suggest changing C-x 4 a's defaulting.

Re: M-n in diff mode: use to a better spot?

[Kai Grossjohann]

Richard Stallman <rms@gnu.org> writes:

>     WIBNI there was some automatic way of moving to the right spot in the
>     diff buffer, so that C-x 4 a can provide a better default?
>
> Are you suggesting that if point is at the start of the hunk, C-x 4 a
> should operate on the file position of first change in the hunk, rather
> than on the file position corresponding to point?  That would have the
> effect you want, and I think it would not cause problems.

Done.

I wonder, if point is at the start of a file, should it then operate
on the file position of the first change in the first hunk?
Currently, it doesn't.

Kai