Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 25 Sep 2020 19:23:07 +0200
From:      =?UTF-8?Q?Ulrich_Sp=C3=B6rlein?= <uqs@freebsd.org>
To:        Brooks Davis <brooks@freebsd.org>
Cc:        git@freebsd.org
Subject:   Re: Question on (my) workflow migration svn -> git
Message-ID:  <CAJ9axoR=0zPxUSF-wH-ZT%2BeeHN3-ZB%2B=M6NcbWR1zKncU2m%2BKA@mail.gmail.com>
In-Reply-To: <20200924205326.GB64154@spindle.one-eyed-alien.net>
References:  <20200924201715.GR1390@albert.catwhisker.org> <20200924205326.GB64154@spindle.one-eyed-alien.net>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Sep 24, 2020 at 10:53 PM Brooks Davis <brooks@freebsd.org> wrote:
>
> On Thu, Sep 24, 2020 at 01:17:15PM -0700, David Wolfskill wrote:
> > The migration (of my workflow) from cvs to svn was more "cosmetic"
> > than anything else -- the data store changed and the commands
> > changed, but the underlying thought processes behind what I was
> > doing stayed pretty much the same.
> >
> > I have no idea if that will be the case for svn -> git, so I'm
> > putting this out for comments and, perhaps, some useful advice from
> > folks who are more familiar with git than I am.
> >
> > Please note: I am not a FreeBSD committer.  I have no desire (let
> > alone requirement) to cause changes to the FreeBSD code base to
> > show up with my fingerprints on them.
> >
> > Details on the below-described workflow may be found at
> > http://www.catwhisker.org/~david/FreeBSD/upgrade.html; the current
> > state of things may be seen at
> > http://www.catwhisker.org/~david/FreeBSD/history/ (in case either
> > of those is of interest).
> >
> > TL; DR:  I sync private local mirrors, update working copies, then make
> > the running systems match the working copies.
> >
> > Details:
> > I have a couple of machines that I consider "development" machines;
> > one is the laptop that I use for day-to-day activities; the other
> > is a dedicated "build machine" (named "freebeast").  On ecah of
> > these machines, I perform a daily update of:
> >
> > * FreeBSD stable/${head - 1}  [currently: 12]
> > * Locally-installed ports (using the ports "head" branch)
> > * FreeBSD head
> >
> > Perhaps more significantly, I keep the two machines in lock-step,
> > as my activities on my laptop constitute the bulk of my testing of
> > what will be installed on my small handful of "production" machines
> > each Sunday (as they get the most recent snapshot of stable/${head
> > - 1} that freebeast built Sunday morning).
> >
> > To keep them in lock-step, the build machine performs a cron-initiated
> > "svnsync" to update a local private mirror of the svn repositories
> > (src, ports, and doc).  For the laptop, it depends:
> >
> > * If I'm away from home, it also does the same sort of svnsync (at
> >   the same time).
> > * If I'm home, it waits a couple of minutes, then does an rsync
> >   from the build machine's private local mirror.
> >
> > In any case, each of them has a private local mirror that is only
> > updated by the "mirroring" operation -- I don't commit to it.
> >
> > For each of stable/${head - 1}, head, and ports, the appopriate svn
> > working copy is updated, using the svn repo as a read-only reference.
> >
> > Thus (for example), when FreeBSD stable/12 was branched, it was
> > straightforward for me to "clone" the slice used for head (to a
> > slice that I was not actively using until then), then use "svn
> > switch" to change the src working copy to point to stable/12 (vs.
> > head) -- there would be little substantitiative change, so the
> > behavior should be quite similar to what I had already experienced
> > with head.  And once I was comfortable with stable/12, I could
> > "clone" that to overlay the old stable/11 slice.
> >
> > (I've been doing this -- well, substitute a few references to cvs
> > inplace of the svn ones for early days -- since around FreeBSD 3.x
> > (I'm pretty sure I skipped 5.x, though).  And it's been working
> > pretty well for me so far.  I've evn been able to point out an
> > occasional breakage now and then.)
> >
> > So: Is doing something ... similar ... to the above feasible using
> > git (vs. svn)?  And if so, may I have a pointer or two to get me
> > started?  Because it's not at all obvious to me how to accomplish
> > it.  (E.g., in my use of git so far, the notions of "repository"
> > and "working copy" seem to be conflated, which does not lead to
> > warm and fuzzy feelings on my part.)
>
> To continue having the local mirror of the git repo you could use `git
> clone --bare <freebsd-repo>` to download a copy and not checkout out
> a work tree.  I believe your daily updates updates would just be `git
> fetch`, but I haven't worked with bare repos much.  You could then do
> a local clone of that (at the cost of two complete copies of FreeBSD
> history).
>
> Your laptop situation is more complicated, but I think I see how it
> could work.  You'd initially clone from your local buildbox's mirror
> with something like `git clone ssh://freebeast/path/to/freebsd/repo`.
> You'd then checkout out whatever branch you wanted and update with `git
> pull`.  To handle the away scenario you have two choices, you can change
> your url of your remote branch to point to upstream freebsd and update
> as usual or your could add a second remote (e.g. freebsd vs the default
> origin) and perform a "fast-foward" merge from the corresponding branch
> on that remote to your working branch.  This would look something like
> the following (totally untested with made up branch names) example:
>
> git clone ssh://freebeast/path/to/freebsd/repo freebsd
> cd freebsd
> git remote add freebsd <freebsd-repo-url>
> git checkout stable/12
>
> So long as you don't commit any local changes your update process at
> home would look like:
>
> cd freebsd
> git pull
>
> To update while away, you'd so something like:
>
> cd freebsd
> git fetch freebsd
> git merge --ff-only freebsd/stable/12
>
> I don't see a good way to keep a bare repo on the laptop.
>
> I hope this give you an idea where to start.

Why not a bare repo same as with the svnsync copies before?

On each machine (doesn't matter if connected or not) he could:

git clone --bare https://.../{src,ports}

and "git fetch --all --prune" (or some combination) whenever internet
connectivity exists. Then it's easy enough to clone from that local
copy for the various branches, but it's obviously suboptimal. I
recommend to read `man git-worktree`. From those bare repos, he could
check out both stable/12 and head to some arbitrary location and keep
them updated with ... um git pull I believe.

so for svn switch, look at a bare clone + multiple worktrees (if you
need just a single worktree, then don't do the bare repo, obviously).

I also assume that your combination of svnsync + rsync was only due to
speed, yes? You can ignore all that and always pull directly from the
cgit.freebsd.org host or whatever the final URL will be. The savings
in bandwidth are not worth it to do something else, but you can of
course `git remote add ...` on your laptop and fetch first from the
local clone and then from the internet.

In short, the svnsync crutch is no longer needed, git can do that all
natively and will be so much faster doing it, to boot. But, you'll
need to read up on clones, fetches, bare repos, worktrees and other
stuff.

Happy to answer more questions, once you've played around a bit to get
a feel for what's possible.

Cheers
Uli



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ9axoR=0zPxUSF-wH-ZT%2BeeHN3-ZB%2B=M6NcbWR1zKncU2m%2BKA>