[darcs-users] Re: Possibly a very simplistic solution

Martin Pool mbp at sourcefrog.net
Fri May 21 08:21:59 UTC 2004

On Wed, 19 May 2004 07:20:34 -0400, David Roundy wrote:

> I think the best direction to move would be towards supporting
> (optionally or configurably) a berkeleydb database for _darcs/current. 
> This "shouldn't" be too hard, and would give us all sorts of nice
> features like transactions, etc.  It would also probably be faster than
> the current implementation, since we could tune the database to our
> access times, and we'd be able to avoid system calls to a large extent. 
> On big repos, the bottleneck is often stat(2) calls to find modification
> times and file sizes.

Please think twice (or more) before doing this.  I think many people would
agree that berkeleydb is one of the prime sources of annoyance in svn.

I think stat'ing the current and working copy of every file an operation
reads is a reasonable thing to do.  In my little experiments with large
trees it seemed like memory usage was more of a problem than stat calls.

If you must use a db then tdb is worth a look.


More information about the darcs-users mailing list