Daniel Rall wrote:
> On Wed, 20 Sep 2006, Mathias Weinert wrote:
> ...
> > If you want to use the diffs as input for a patch program we should
> > provide two diffs. One is the copy as described above and one is the
> > succeeding modify.
>
> Please, no...
Okay.
>
> > On the other hand as a human reader I will always prefer only one diff
> > which only contains the changes after the copy action beacuse the
> > information that this is a new file as a result of a copy I can find
> > in the enhanced texts in brackets (or in the header of the diff or at
> > the beginning of the message, as Daniel pointed out).
>
> +1
>
> > So what do we have until now?
> > - Some people like the more verbose texts and some don't.
> > - Most people only use the diffs for human consumption while several
> > use it as input for a patch program.
> >
> > This again makes me think about an option diff_style or something
> > similar, althoguh I am not totally convinced that this is necessary.
> > And BTW it won't help people who cannot set personal diff options as
> > it is toady for recipients of *the* Subversion repository changes.
>
> The primary use cases for the change notification emails generated by
> this script are, in order of priority:
>
> 1) Code review.
> 2) Patch repository.
>
> #1 is the primary use case, with #2 being more of a bonus that we
> should try hard not to lose.
And #1 being in focus of my patch.
So I will provide another version of my patch which
- includes current discussion results
- doesn't make automatic patching worse than it is today
Mathias
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Sep 21 11:18:14 2006