> From: Julian Foad [mailto:julianfoad@btopenworld.com]
> Sent: Sunday, September 21, 2003 12:26 AM
> I am concerned that the implementation of special revision ids may go beyond what is useful and what is documented. I am
> not concerned about the amount of code or execution time, but about the maintenance cost if people start to depend on
> these things unnecessarily.
>
> subversion/libsvn_subr/opt.c (revision_from_word) implements the following familiar words:
>
> HEAD
> BASE
> COMMITTED
> PREV
>
> but also these:
>
> CHANGED: a synonym for COMMITTED
>
> Do we really need that?
I don't really have an opinion on this one. I tend to believe less is good,
but if some people like, or, more importantly, use, the alias, let's keep it in.
> FIRST: a synonym for 0 (number zero)
>
> Do we really need that? I can imagine a meaning of "FIRST" that would be more useful: "The first revision in which the
> item existed."
Hmmm, what does usage of FIRST do with the various commands nowadays?
What is the expected output? Is that intuitive?
> It allows them to be specified in any mixture of upper and lower case:
>
> HEAD head Head hEaD
>
> Do we really need that? I checked in the source code of CVS; it only allows upper case, and Subversion's documentation
> always specifies upper case.
+1 on only allowing uppercase.
> My opinion, if you haven't guessed, is that none of these are useful. Sure, case insensitivity save you from having to
> press the SHIFT key, and CHANGED saves two key strokes over COMMITTED, but I don't think those are good enough reasons.
> But maybe there are other reasons, like compatibility with some other system.
Sander
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Sep 21 12:03:15 2003