Date: Thu, 18 Nov 2021 09:46:50 -0700 From: Warner Losh <imp@bsdimp.com> To: Brooks Davis <brooks@freebsd.org> Cc: Philip Paeps <philip@freebsd.org>, Mark Millard <marklmi@yahoo.com>, "Ulrich Sp??rlein" <uqs@freebsd.org>, Graham Perrin <grahamperrin@gmail.com>, freebsd-git <freebsd-git@freebsd.org> Subject: Re: cgit, ages and chronological order Message-ID: <CANCZdfrWT1OxgK3T2b-5R8wD2rBJ41wdYwx8iDdh=VSWvzcEqQ@mail.gmail.com> In-Reply-To: <20211118162927.GG81740@spindle.one-eyed-alien.net> References: <9766b3e1-fb5d-1993-46e2-057e2567315a@gmail.com> <CAJ9axoT6kEwC%2Bt5zHSKPSFgFmaOt8-CXPAG5jsanWobT4LZhpA@mail.gmail.com> <36020FD7-32A4-4869-B6A2-2622F50F6478@yahoo.com> <FFB55106-BA73-4639-9772-0B3608B3E8DC@freebsd.org> <20211118162927.GG81740@spindle.one-eyed-alien.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Nov 18, 2021 at 9:29 AM Brooks Davis <brooks@freebsd.org> wrote: > On Thu, Nov 18, 2021 at 12:46:49PM +0800, Philip Paeps wrote: > > On 2021-11-17 06:32:52 (+0800), Mark Millard via freebsd-git wrote: > > > information being based on local git commit timing (and clocks) > > > vs. when the commits are pushed to FreeBSD servers: The display > > > order is from the timing on the FreeBSD servers but the Age is > > > based on the original commit (before the push). The longer the > > > delay between commit and push, the more noticeable the > > > distinction is. > > > > Some projects require a "git rebase --ignore-date" (or "git rebase > > --reset-author-date", which I consider the more obvious spelling) before > > pushing. A hook could potentially reject commits with timestamps that > > are too far off to the server's liking. > > > > I can't comment on whether we need or want either the policy or the hook > > or both. I don't really have a problem with the default Git behaviour > > here. > > I always use --ignore-date when curating pre-commit. I'd like to at > an absolute minimum enforce that CommitDate be newer than the previous > commit and older than the push time. There is no good argument for > allowing non-linear CommitDates since the only requirement is that the > committer have their clock set more or less correctly. > At the very least, we should document this suggestion. I also like the idea of enforcing this as a pre-commit hook, but before that we should have it the docs... wouldn't hurt to have a section in the docs about each of the major points we enforce to tell people have to fix the issue should they encounter it. Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfrWT1OxgK3T2b-5R8wD2rBJ41wdYwx8iDdh=VSWvzcEqQ>