Re: Issue #3048 and r29189

[Karl Fogel <kf...@red-bean.com>]

"David Glasser" <gl...@davidglasser.net> writes:
> On Wed, Feb 27, 2008 at 7:49 AM, Justin Erenkrantz
>>  As I said when it was suggested (and was promptly ignored), I must
>>  protest that the UI changes to the interactive conflict resolution to
>>  support #3048 make the entire system useless.
>>
>>  First-off, case-sensitivity should simply not matter.  The removal of
>>  'd' and 't' options are incredibly short-sighted - this is not an
>>  acceptable situation and certainly not for 1.5 to be 'not implemented'
>>  if that is the plan.  I don't believe the problem that was raised
>>  (that the diff is not as optimal as it could be) is really a
>>  legitimate problem.  I think it's actually more useful to have the 'd'
>>  and 't' be as they were - with (at best) new options to show a reduced
>>  merge.  I find it way more useful to show me what changed overall
>>  since my update (ie base rev) *and* where the conflicts are - because
>>  without what merged 'cleanly', it's not easy to understand the
>>  conflicts!  So, even if/when 'd' and 't' are implemented as described
>>  in #3048, the default options are going to be not that helpful because
>>  we're going to throw away the remaining merge information that *led*
>>  to the conflict and we'll just get only part of what changed.  This is
>>  not enough information to be useful to me as a developer when I'm
>>  evaluating the change.
>
> I think this ends up being a matter of opinion.  I've certainly
> already found on a regular basis that when a small number of conflicts
> turn up during a huge merge, the diff options are pretty worthless to
> me.

All addressed in r29648, and I'll nominate it for backport now.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org