"C. Scott Ananian" <cananian@lesser-magoo.lcs.mit.edu> writes:
> On 6 Sep 2001, Ben Collins-Sussman wrote:
>
> > Very true. I believe we had always planned to extend the DTD for
> > this format to include inline (xml-escaped?) unified diffs. If this
> > problem interests you or others, please add this feature!
>
> it does interest me, not sure i'm going to have to work on it, though.
> help me out on one thing, though: what does 'xml-escaped' mean? I would
> assume the DTD would mark <diff-delta> as CDATA, so only the magic string
> ']]>' is prohibited from occurring.
I really know very little about XML. I believe that you're correct --
although maybe there are other sequences to protect against.
> http://www.w3.org/TR/2000/REC-xml-20001006#sec-cdata-sect
> It seems logical to fall back to <text-delta> if the output of GNU diff
> includes that (fairly uncommon) string.
Our DTD should keep <text-delta> to mean exactly that: a delta on a
file's textual contents. If we decide to add unidiff format, we
should allow arguments to <text-delta>:
<text-delta>, or
<text-delta kind='svndiff'> -- the default: svndiff data.
<text-delta kind='unidiff'> -- contains unidiff data instead.
[Hope you don't mind my cc'ing this back to the dev list. There are
many more on the list who actually understand xml and DTDs.]
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:36:40 2006