Re: svn blame on trunk

On Fri, Nov 19, 2004 at 09:14:10AM -0600, Ben Collins-Sussman wrote:
> This whole issue lies in fuzzy territory: was it really a "silent UI
> change"? Silent, only because I've not documented it well yet. UI
> change? You could argue that we were simply fixing a long-standing
> class of bugs (issue 1093), and that the only solution was to "add some
> new UI syntaxes", just like we add new APIs.

We didn't just add new UI sytanx we changed some what some commands
meant when they didn't have peg revs specified. I realize that the
additions of peg revs was necessary and that some of the changes were
even probably the right thing to do. But that doesn't mean it wasn't
wrong per our compatability rules.

> But yeah, I agree that there's some danger here too. So far, after 7
> weeks since 1.1's release, I've not seen any complaints in IRC or
> users@ with somebody saying, "hey, my script suddenly behaves
> differently." So I'm not too worried.

The very first time I tried to do merges with a 1.1.x client for the
1.1.1 client I ran into this issue. I doubt very many people are doing
operations that we changed some of the meaning of the lack of peg revs
from scripts. But I'm willing to bet people have had perplexing
behavior changes that they just didn't know why and never bothered to
ask about. Further, the cases where we changed the meanings of the
commands aren't all that common.

I'd have to go dig in my IRC logs to even remember exactly what I was
doing to run into it. But it required a specific set of repository
conditions to run into.

-- 
Ben Reser <ben@reser.org>
http://ben.reser.org
"Conscience is the inner voice which warns us somebody may be looking."
- H.L. Mencken
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org