[darcs-users] Re: patch numbers and annotate and diff etc.

[Ralph Corderoy]

Hi Mark,

> > I'm in favour of a simple, repo-local, number to identify a point in
> > the repo's state.  Could your suggestion above be considered as
> > similar to an implicit `tag' at each point?  Also, would a repo's
> > definition of the numbers be visible to others so my local 137 was
> > visible to you as http://...137.
> 
> I think we are talking about a shorter unique patch ID than the
> current hash.

Yeah, I want the incrementing numbers from 1 with negatives working
backwards from the end.

> Your suggestion is interesting but different.  If you publish the
> number, wouldn't it cease to be "repo-local"?

Publishing it doesn't make it non-local.  If I say to you right now that
"It's 137 in my repository over here" I've published it, but it's still
repo-specific, and not accessible to you.  I'm trying to point out that
they seem similar to tags in some ways in what they represent, and that
by combining a repo's location with 137 the repo-specific 137 is then
accessible by you.

> Two identical repos can have different state numbers, if one records
> and unpulls and a patch that the other one never had.

I don't think that's relevant, hopefully my above explanation will
clarify.

Cheers,


Ralph.