Subject: Re: [boost] [git] open an "admin" link doesn't works
From: Rene Rivera (grafikrobot_at_[hidden])
Date: 2013-11-01 16:47:07
On Fri, Nov 1, 2013 at 12:50 PM, Beman Dawes <bdawes_at_[hidden]> wrote:
> On Thu, Oct 31, 2013 at 2:03 PM, Eric Niebler <eniebler_at_[hidden]> wrote:
> > On 10/30/2013 10:58 PM, Daniel Pfeifer wrote:
> > > Beman, Dave, Eric,
> > >
> > > Should we attend admin issues now or wait until after the final
> > > switchover to git?
> > >
> > > Advantage of starting now: Authors can experiment with their libraries
> > > inside boostorg.
> > > Disclaimer: They even can push their changes, but need to keep in mind
> > > that Boost2Git continuously recreates the whole repository structure
> > > based on the SVN contents.
> > >
> > > Advantage of doing it when Boost2Git job is switched off: Avoid
> > > surprises that commits get deleted/overridden.
> > My $0.02: Couldn't hurt to grant folks access when they ask for it. When
> > we close the admin issues, we should make it clear that any changes made
> > to the repos before the switch-over will be overwritten.
> I don't see how we can test Boost2Git without letting authors experiment
> with their libraries, including pushing to the public repo.
> An attempt to push was how the https vs git protocol issue came up, and I
> expect more testing by more developers to raise more issues.
> Remember, some developers have very little experience with git, and others
> have little or no experience with sub-modules. Testing will will help bring
> people up to speed.
> We can limit granting write access to those who acknowledge that they
> understand that changes will be overwritten until Boost2Git job is switched
I would agree with that. And I have a suggestion.. There are a number of
developers, myself included, who already have their libraries in github
(and at least in my case already in the appropriate structure). It would
seem that giving the authors with those libraries permission to directly
push their existing repos to the Boost canonical ones would be the most
beneficial way to test. Especially since it would remove the git->svn->git
translation that is already going on.
-- -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk