On Thu, Jun 18, 2015 at 01:45:24PM -0400, Ron W wrote: > On Wed, Jun 17, 2015 at 3:17 PM, Joerg Sonnenberger <[email protected] > > wrote: > > > On Wed, Jun 17, 2015 at 12:10:36PM -0400, Richard Hipp wrote: > > > On 6/17/15, Jan Danielsson <[email protected]> wrote: > > > > Out of curiosity; why aren't pulls 100% read-only on the server? > > > > > > > > > > The server might decide to create a new cluster artifact > > > (https://www.fossil-scm.org/fossil/doc/trunk/www/fileformat.wiki#cluster > > ) > > > to help it with the sync. > > > > But it should not have to do that on a pull. > > > See section 5.1, step 5 in > https://www.fossil-scm.org/index.html/doc/trunk/www/sync.wiki > > Of course, if the read-only repo is a "snap shot clone" of a repo with > fewer than 101 unclustered artifacts, then new clusters won't be created.
I'm aware of the sync protocol. My point is that all clusters should already exist on the public master repository as part of other sync operations and I would strongly argue that this is the case that should be optimised for. It is a pity that the nature of the sync protocol doesn't make it possible to enforce this more strongly. Joerg _______________________________________________ fossil-users mailing list [email protected] http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
