Re: Using Github for HTTP2

One very minor annoyance I've encountered is working on multiple edits in
parallel (in separate feature branches) and keeping the HTML in sync
(across my branches and merges from upstream). I'm not sure if anyone else
has this problem. My two cents: it'd be nice to ditch the HTML from the
main repo/branch and use https://help.github.com/articles/post-receive-hooks to
process the commits to XML and generate the other formats (TXT+HTML) and
post them somewhere (another repo/branch/website/etc).

That said, it's probably not important since I bet very few folks outside
the editors will make changes here, so as long as it works for them, then
great. Just noting this in case anyone felt it was useful to address.


On Thu, Feb 7, 2013 at 6:15 PM, Mark Nottingham <mnot@mnot.net> wrote:

> For the HTTP/1.1 work, we've used the IETF's infrastructure on
> tools.ietf.org -- mostly, the SVN repository and the Trac instance for
> issue tracking.
>
> After briefly discussing it in Tokyo and conferring with the editors and
> our AD, we've moved the source for the HTTP/2 draft and related issues to
> Github.
>
> There are a couple of reasons for this. First of all, we would like to
> encourage participation from a broader audience, and more implementers have
> github accounts and are familiar with that workflow, vs. IETF tools
> accounts. Furthermore, we've had a few requests from people that want to
> fork the specification for experimentation, to be used as potential input
> to our work (as pull requests). Finally, our editors have expressed some
> preference for working with Git over SVN.
>
> As such, this isn't a problem with the IETF tools themselves (which are
> well-supported and have been very useful to date); it's more driven by the
> specific situation we're in.
>
> So, see:
>   https://github.com/http2/http2-spec
>
> I'm now in the process of closing the relevant Trac issues and adjusting
> our wiki to point to the relevant places.
>
> Note that this is itself an experiment; e.g., we may need to establish
> some process around issue handling, especially, if the input there gets out
> of hand.
>
> Also, note that this does NOT affect any of our work on HTTP/1.1; that
> will remain at the IETF tools site.
>
> Cheers,
>
>
> --
> Mark Nottingham   http://www.mnot.net/
>
>
>
>
>