Re: [bug-grep] [PATCH] --initial-tab, 3 newly colorized items, and GREP_COLORS

[Charles Levert]

* On Sunday 2005-01-16 at 14:31:18 +0000, Julian Foad wrote:
> 
> ("Please _test_ it out"? :-)

":-)" indeed.  This happens.  I meant "check it out".

> Thanks for the patch.  Both the initial-tab and the extended colours look 
> like good ideas.

I would like to solicit the following feedback.

  -- Should the new parse_grep_colors() print a warning like this:

        fprintf(stderr, _("%s: GREP_COLORS is ill-formed at or just before 
remaining substring \"%s\".\n"),
                program_name, q);

  -- Should uses of sgr_end in fputs() statements be replaced by SGR_END
     for uniformity with SGR_ARG (and a "#define SGR_END sgr_end" added)?

> Since they are independent of each other, and each need 
> their own review and discussion and testing, they really should be separate 
> patches.  If you would be willing to separate them, I'm sure that would 
> increase the chances of one or both of them being accepted.

I was thinking about applying this sound principle when I was preparing
the patch (I did write both features concurrently).  The problem is,
they touch a good portion of common code, particularly in the prline()
function.  This means I would have to either write one to be applied
on top of the other, or write two patches to the original while loosing
the ability to apply both.

Given this, I would prefer to exceptionaly leave both features in
the same patch.  Otherwise, please help me choose among "initial-tab
over color over original", "color over initial-tab over original" or
"initial-tab over original, and color over original".

Both features are also related in that the idea for the field separator
(":" or "-") to have its own color (instead of that of the preceding
field) grew out of the fact that initial-tab can make them separate
substrings in the output line.

> We probably will need a copyright assignment or disclaimer to cover this 
> work.

Can you guide me through this?  I've never done this with the FSF before.
Is there a boilerplate or step-by-step instructions somewhere?  Can this
be done once and take effect anytime I would explicitly invoke it in
the future (for any software whose copyright is held by the FSF)?

Thanks.