Re: [PATCH 2/3] add-patch: add diff.context command line overrides

[Junio C Hamano <gitster@pobox.com>]

Phillip Wood <phillip.wood123@gmail.com> writes:

>
>> +`-U<n>`::
>> +`--unified=<n>`::
>> +	Generate diffs with _<n>_ lines of context. Defaults to `diff.context`
>> +	or 3 if the config option is unset. Implies `--interactive/--patch`.
>> +
>> +`--inter-hunk-context=<n>`::
>> +	Show the context between diff hunks, up to the specified _<number>_
>> +	of lines, thereby fusing hunks that are close to each other.
>> +	Defaults to `diff.interHunkContext` or 0 if the config option
>> +	is unset. Implies `--interactive/--patch`.
>
> This documentation is repeated for each command. I think it would be
> better to put this in separate file that is then included where it is
> needed. That way if we need to update the documentation in the future
> we only have one copy to worry about. The syntax to include a file
> called diff-context-options.adoc is
>
> include::diff-context-options.adoc[]

Excellent suggestion.

I however think "-U implies -i/-p" makes no sense at all.  What if
the user said "git add -U6"?  Do we run "git add -i -U6" or go
directly to "git add -p -U6"?

I think it is better to encourage the user to be more explicit to
require them to say "git revert -U6 -p" or "git add -U6 -i".

If -U<n> is given and there is no "-i" (for "add" only) or "-p" (for
everything), just error out *until* other modes of these commands
find a good use of context width information, at which point we can
use the context length information to drive these commands in their
non-"patch" mode.

Even if we cannot think of a way to use the diff output (and with
configurable context length) in modes that are not "--patch" for
these commands RIGHT NOW, that does not mean -U<n> would never
become useful in these modes for the commands.  Let's not close the
door prematurely from our future developers by making -U<n> imply
anything, and make it an error to use -U<N> without -p/-i for now.