Re: svn commit: r27875 - in trunk/subversion: bindings/javahl/native bindings/javahl/src/org/tigris/subversion/javahl bindings/javahl/tests/org/tigris/subversion/javahl include libsvn_client libsvn_diff svn

On Fri, 16 Nov 2007, David Glasser wrote:

> On Nov 16, 2007 2:11 PM, <dlr@tigris.org> wrote:
> > Author: dlr
> > Date: Fri Nov 16 14:11:38 2007
> > New Revision: 27875
>
>
> > + * If @a relative_to_dir is not @c NULL, the @a original_path and
> > + * @a modified_path will have the @a relative_to_dir stripped from the
> > + * front of the respective paths. If @a relative_to_dir is @c NULL,
> > + * paths will be handled the pre-1.5 way. If @a relative_to_dir is not
>
> [...]
>
> > * Output a @a diff between @a original_path and @a modified_path in unified
> > * context diff format to @a output_stream. Optionally supply
> > * @a original_header and/or @a modified_header to be displayed in the header
> > * of the output. If @a original_header or @a modified_header is @c NULL, a
> > * default header will be displayed, consisting of path and last modified time.
> > - * Output all headers and markers in @a header_encoding.
> > + * Output all headers and markers in @a header_encoding. If @a relative_to_dir
> > + * is not @c NULL, the @a original_path and @a modified_path will have the
> > + * @a relative_to_dir stripped from the front of the respective paths. If
> > + * @a relative_to_dir is @c NULL, paths will be handled the pre-1.5 way. If
>
> Does "the pre-1.5 way" here just mean "they won't be stripped"? Just
> saying that (or saying nothing at all, since it's the absence of an
> action) seems more useful than requiring future developers to do
> archaeology to understand how an API works.

Yup, and yup. Jeremy, can you come up with a patch?

Thanks, Dan