Re: [PATCH] to fix a compile error of merge-tracking on Windows

On Mon, 2006-07-24 at 20:55 +0200, Branko Èibej wrote:
> > It works perfectly fine (I've very heavily load tested it) on servers
> > that support nfs locking properly. This includes windows nt, linux,
> > solaris, etc.

I'm curious what level of locking support is necessary, since AFS
supports file locking but not byte ranges. Just a point of curiosity,
though.

I'm also curious what index we're using sqlite for. The patches being
circulated look like they're against the trunk, but the sqlite index I
remember being discussed was for the merge-tracking branch.

No one's required to satisfy my curiosity, of course; these are just
places I've fallen behind on following Subversion development.

> All right ... but, doesn't more or less the same argument hold for BDB
> over NFS?

No, BDB memory-maps the environment and does pthreads-type locking in
the memory-mapped space. That's nigh impossible to make work over a
network filesystem.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Jul 24 21:00:47 2006