On Fri, Nov 26, 2010 at 11:18:36AM -0800, Matthew Jacob wrote: > Something like that. 209704 got itself recorded as merged from head. But > nothing changed. I'm undoing that so I can do it right. This is the > reason why the geom multipath stuff ended up totally broken in 8.1 even > though I though I had merged it. > > Or have I missed something? Might be I am confused as well, but r215873 did not reverted recording of the merge of r209704. Instead, the r215873 is revision that seem to recorded it as merged.
> > On 11/26/2010 8:38 AM, Kostik Belousov wrote: > >On Fri, Nov 26, 2010 at 04:21:44PM +0000, Matt Jacob wrote: > >>Author: mjacob > >>Date: Fri Nov 26 16:21:44 2010 > >>New Revision: 215873 > >>URL: http://svn.freebsd.org/changeset/base/215873 > >> > >>Log: > >> Revert a completely fouled up merge. > >> > >>Modified: > >>Directory Properties: > >> stable/8/sbin/geom/ (props changed) > >> stable/8/sbin/geom/class/part/ (props changed) > >> stable/8/sbin/geom/class/sched/gsched.8 (props changed) > >> stable/8/sbin/geom/class/stripe/ (props changed) > >Is this a revert ? The commit looks as a record of the merge of > >r209704, that in fact did not happen ?
pgplTPNVcJ6Wm.pgp
Description: PGP signature
