Rich's suggestion sounds clear to me. --Mark
On Thu, 31 Jul 2008, Richard Lowe wrote: > Date: Thu, 31 Jul 2008 18:28:41 -0400 > From: Richard Lowe <richlowe at richlowe.net> > To: Mark J. Nelson <Mark.J.Nelson at Sun.COM> > Cc: Valerie Bubb Fenwick <Valerie.Fenwick at sun.com>, > John Beck <John.Beck at sun.com>, scm-migration-dev at opensolaris.org > Subject: Re: Sponsor task SCM updates > > "Mark J. Nelson" <Mark.J.Nelson at Sun.COM> writes: > >>>> The full context here, from the current site: >>>> "As opposed to the SCCS delta comment, the putback comment should have an >>>> additional line, of one of the following two formats, either: >>>> >>>> Contributed by Firstname Lastname. >>>> >>>> or >>>> >>>> Contributed by Firstname Lastname <something at example dot com>. >>>> >>>> Which format to use is up to the contributor." >>>> >>>> So.... I assume s/SCCS delta/Mercurial changeset/ but don't know >>>> what to do s/putback comment/ with... >>> >>> Crap, sorry, Val, I didn't read carefully enough. >>> >>> There is NOT a "putback comment" analog with Mercurial. You push a >>> changegroup, consisting of one or more changesets, each of which has >>> changeset comments. >>> >>> The way you do this in Mercurial SHOULD be by having the contributor as >>> the user in the changeset that the sponsor is pushing. >>> >>> Sometime tomorrow, can you pop over to #onnv-scm on freenode, and discuss >>> this with the team? >> >> So Val did this, and here's what we came up with, plus some additions >> that writing it out made me think of: >> >> The existing SCCS version: >> >>> Since the contributor's name and e-mail address are listed in the bug >>> report, the SCCS delta comment should be the usual bug-ID and synopsis >>> plus any ARC cases: the contributor's name should not be listed. >> >> The Mercurial version: >> >>> The contributed changeset(s) should be committed by the Contributor, not >>> the Sponsor. This should be done using either "login <e-mail at address>" >>> or "full name <e-mail at address>" format for hg user. If the Contributor >>> does a commit and supplies a Mercurial bundle of the changeset(s), this >>> will be done correctly. If the Contributor supplies a patch, and the >>> Sponsor then applies the patch and commits the changes, they will need >>> to use "hg commit -u" or "hg recommit -u" to set this correctly. >>> >>> When the Sponsor pushes the changes to the gate, the From: field of the >>> notification e-mail will identify the Sponsor, not the Contributor. >>> Both Sponsor and Contributor can be identified through either the bug >>> report(s) or the RTI. > > If we're doing it that way, we need to mention maintaining author: > over the reci more clearly. > > ... If you have cause to merge and thus recommit, you should use the > -u option to recommit to preserve the author: field. > > something like that? > > -- Rich >
