Re: rdiff -s

[Paul Edwards]

"Derek Robert Price" <derek@ximbiot.com> wrote in message 
news:mailman.392.1058969547.8231.bug-cvs@gnu.org...
> >>Why not just regex for each pattern or
> >>something predictably more consistent like \d\.[\d.]*\d?
> >
> >and requires a much higher level of sophistication or tool in
> >order to convert CVS's summary into... a summary!
>
> A summary "suitable for easy processing by other utilities".  I still
> don't think reducing a four-case case statement in your shell script to

3 I think.

> a single line justifies fixing the line length of CVS output until the
> end of time.  I'd like to see CVS using gettext or the like for multiple
> language support someday and getting fixed line length to translate into
> several languages would be a bear.

Actually, that's the best argument you could possibly have
made in favour of fixing it.  The summary shouldn't have
anything that is subject to change in translation.  Because
people should be able to write and distribute scripts that
don't require modification when the user users a different
language.

The words should be changed from "removed" to "R" etc.

The output of rdiff (no -s) is a very fixed/standardized
"patch" format.

rdiff -s should be likewise standardized.  You could argue
that it is already standardized, but English verbiage is a very
unusual standard.  It is more common to see a "computer
format" with subsequent English explanation in the manual.

BFN.  Paul.