Subject: Re: [boost] [partly OT] Re: [review queue] What to do about the library review queue?
From: Tom Kent (lists_at_[hidden])
Date: 2017-03-17 00:12:11
On Thu, Mar 16, 2017 at 2:08 PM, Vinnie Falco via Boost <
> On Thu, Mar 16, 2017 at 2:54 PM, Robert Ramey via Boost
> <boost_at_[hidden]> wrote:
> > If we "require" git, note that we'll be strapping ourselves to git for a
> > very long times. Good or Bad it should be considered. If that's good,
> > might just deemphasize distribution via zip file and just encourage
> > to clone the git repository. For me this is a convenient way to work.
> > But what about when boost outlives git - as it has a host of other
> > services (like sourceforge)
> I think when you say git you mean GitHub (since issues are a GitHub
> feature not a git feature).
> GitHub has an API allowing you to extract all the issues, so if Boost
> does outlive GitHub then there is a path for migrating all of the data
> by extracting it first:
I love GitHub's interface and am happy to use it for now, but we should
already be planning the next transition even though we don't know where we
will be going. Ideally we'd have some server running this API daily, so
that if one night GitHub decides that you have to start paying to use
issues and the API, we already have a backup.
Same with the source code, today everyone is using GitHub for source code
and collaboration, but do we have a backup (our own server at git.boost.org?)
for the day when GitHub decides not to play nice anymore?
We've already seen the pain on the distribution side with sourceforge,
where we wanted to leave, but didn't have the infrastructure that would
allow us to.
I don't have any answers to these questions, but I'd rather see boost hire
someone to manage these things and pay for our own services/hardware than
to pay for reviewers.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk