On Tue, 2008-06-10 at 15:30 +0200, Erik Huelsmann wrote:
> On 6/10/08, Jens Seidel <jensseidel_at_users.sf.net> wrote:
> > I also wonder why a temporary copy is provided to the diff command. This
> > is probably required if the comparision is not against BASE.
> > Nevertheless I like to edit the file in the diff viewer as well and have
> > to parse option 2 or 4.
I, too, fairly often want to edit my working file while it's in my diff
viewer, and I can't usefully do so when it's a temporary copy.
> Because you need to de-translate keywords and line-endings in order to
> get a clean diff between the working copy file and the base.
Well, you're right in as much as that we need to translate these things
in some direction to get a clean diff, but need we do it that way round?
Instead of giving a diff between (base or something like it) and (how
the working copy will be committed to the repository), I can see
arguments for giving a diff between (what the fresh working copy would
have been) and (what the working copy is now). That would provide the
user's own line endings, and his diff tool might work better with them
than with the repository-normal-format line endings that we currently
give it when eol-style=native (or do we? I'm not sure). It would also
show any changes he might have made accidentally or uncomprehendingly to
the keywords. No doubt there are opposing arguments too, but this seems
to have some merit.
- Julian
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-06-10 16:44:41 CEST