Re: [Monotone-devel] Re: [HACKERS] SCMS question

[Joel Crisp]

Actually, IMHO the $id$ thing is made redundant by the use of hashing.

In most cases it is used to determine the original version a file came
from. In monotone, this devolves to generating an SHA1 hash of the
unknown file, followed by a DB query which  will give you as much (or
as little) information about that file as possible.

The only point at which this doesn't help is if the unknown file has
been modified, in which case I'm sure there are easy to implement
similarity heuristics which would help.

Joel

On 2/22/07, Alvaro Herrera <address@hidden> wrote:
> A.M. wrote:
> >
> > On Feb 22, 2007, at 13:00 , Alvaro Herrera wrote:
> >
> >
> > >
> > >One thing that Monotone doesn't do to which we are used to, is
> > >$Id$-style keyword expansion.  The only other problem I see
> > >currently is
> > >the long time to get an initial pull.
> >
> > Monotone will never support such templating because it relies on SHA
> > hashing to track the repo history, so changing the repo before the
> > hashing would mean that monotone would have to track pre- and post-
> > templated files.
>
> I know :-(   My question is, what is the mechanism used to replace it?
>
> One thing of note is that we have some files derived from NetBSD, Tcl,
> zic, and others, and we have $NetBSD$ tags in there (etc) to track the
> exact file we copied.  It would be nice to have a simple mechanism to do
> that as well.
>
> How does Monotone solve this problem?
>
> --
> Alvaro Herrera                         http://www.flickr.com/photos/alvherre/
> "Cuando mañana llegue pelearemos segun lo que mañana exija" (Mowgli)
>
>
> _______________________________________________
> Monotone-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/monotone-devel