[Openais] Misc Corrections

Mark Haverkamp markh at osdl.org
Wed Oct 27 15:13:51 PDT 2004


On Thu, 2004-10-28 at 00:02 +0200, Daniel Stodden wrote:
> On Wed, 2004-10-27 at 14:17 -0700, Steven Dake wrote:
> > Mark
> > 
> > What I did was create a new openais tree then
> > cat patchfile | bk receive
> > bk resolve
> 
> correct. the actual command is bk takepatch. 'bk receive' is the
> complement to 'bk send', which means it can filter RFC 822 mail
> messages. usage of bk send seems to be preferred by bitmover for some
> reason.
> 
> a few comments (what i've learned so far):
> 
> 1. bk send
> 
> if you do a 'bk send openais at lists.osdl.org', the mailing is generated
> automatically. the last patch range already sent for each individual
> destination is remembered by bk which means repeatedly calling send will
> send incremental changesets as expected. careful: this implies that you
> use -r1.xxx.. the _first_ time you send to a new recipient, where 1.xxx
> is the last changeset you know the receiver already has in its tree.
> otherwise this sends _any_ changeset available, basically the entire
> project history.
> 
> this may be the preferred way to send patches to the list. though, send
> has the problem that it sends a patch and nothing else, which looks a
> little bit pushy if the receipients didn't expect you :).
> 
> steven: you told be to create a gnu diff for the list, i know. the
> problem with this approach that you're losing all my checkin comments
> this way, let alone me having to reextract them to explain what the
> patches were good for.
> 
> the alternative is 'bk makepatch'. this creates a bitkeeper patch to
> stdout for a desired revision range. this can be attached to a cleartext
> message with additional comments. that's how the message above was
> created.
> 
> mark: is it okay if i proceed like this for a while? if desired, i can
> attach unified diffs as well. this would make it easier for people to
> send comments from watching the mail message alone.
> again, i'm not a bk guru. still looking for the optimal process
> myself...

Sure,  I just hadn't seen that format before and didn't know what it
was.  The unified diffs do make it easier to review without having to
apply to a bk tree.

Mark.

-- 
Mark Haverkamp <markh at osdl.org>




More information about the Openais mailing list