On Thursday 16 February 2006 23:27, Julian Foad wrote:
> In languages like C, what we need is:
>
> Ignore spaces completely when they're between tokens that don't
> concatenate; ignore changes to the nonzero amount of space where tokens do
> concatenate.
>
> The sets of tokens that do and do not concatenate are language-specific.
I see lots of problems ahead.
1) In python the whitespace *at the start of the line* is significant.
2) In C, Perl, ..., the whitespace is often the thing which has to be repaired
*for visual reasons*, but should not necessarily show in diffs.
3) How are multiline-strings (valid eg in perl) and document-here handled?
4 ;-) What about whitespace (http://compsoc.dur.ac.uk/whitespace/) and
Acme::Bleach (http://search.cpan.org/dist/Acme-Bleach/lib/Acme/Bleach.pm)?
I believe that this should be solved like the diff-issue, ie. done by specific
diff-programs, which get chosen based on the mime-type or the file(1) result.
Note: I'm not sure how much work for blame is done on the server - but if the
comparisions are done there, all needed diff-programs would have to be
installed there, too.
Sure, you could expand the internal diff with various options, but that'll get
you only part of the way ... a full solution would have to know about the
file-type.
Regards,
Phil
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Feb 17 08:01:10 2006