svn commit: r336448 - stable/10

[Rodney W. Grimes]

> On 7/18/18 12:05 PM, Peter Jeremy wrote:
> > On 2018-Jul-18 07:41:23 -0700, "Rodney W. Grimes" <freebsd at pdx.rh.CN85.dnsmgr.net> wrote:
> >>> Author: peterj
> >>> Date: Wed Jul 18 09:32:43 2018
> >>> New Revision: 336448
> >>> URL: https://svnweb.freebsd.org/changeset/base/336448
> >>>
> >>> Log:
> >>>   Retrospectively document SVN branch point for stable-10 and its releases.
> >>>   
> >>>   This is a direct commit to stable/10 because the releases are taken
> >>>   from the stable/10 branch.
> >>>   
> >>>   Approved by:	jhb (mentor)
> >>>   Differential Revision:	D16263
> >>
> >> Actually I see no reason not to document these in the mainline
> >> UPDATING file and making these MFC's.  As is now when looking
> >> at UPDATING from head I can not easily find the branch point
> >> for any of these releases and that is probably the most useful
> >> time for this information.  If I already have a branch I probably
> >> already know what its anchor point is.
> > 
> > I only put the releng/x.y branch points into the relevant stable/x/UPDATING
> > because releng/x.y is branched off stable/x and I don't think it makes much
> > sense to document those in head/UPDATING.  The stable/x branchpoints are in
> > both head/UPDATING and stable/x/UPDATING.  Note that the stable/10 branch-
> > point was already in head/UPDATING.
> 
> I agree with this.  We should document them in the source branch, but not
> in grandparents like head where there is no single head commit that becomes
> releng/X.Y.

My only counter to this is that often what I am investigating is
the change between ^head/ and some release, say 11.2 and I really
do not want to go grovel in stable/11 to find the rXXXXXX for 11.2,
that is just a PITA.  Though the branch point for stable/11 is interesting,
it is rarely usefull for any thing very meaningful.

It would actually be useful to have a file that listed all of these
for all branch points maintained in ^head/.

> > Do you have a quick way to find branch points?  The best I've found is
> > "svn log -r 1:HEAD --limit 1 --stop-on-copy" within a branch and that
> > is quite resource intensive on the SVN server.
> 
> Finding a file that doesn't change often like MAINTAINERS and only doing the
> log against that shouldn't be as bad.

Thats a good tip.

> John Baldwin

-- 
Rod Grimes                                                 rgrimes at freebsd.org