Le 31/07/2019 à 20:13, Ladislav Lhotka a écrit :
On Wed, 2019-07-31 at 11:04 -0500, Nico Williams wrote:
On Wed, Jul 31, 2019 at 03:15:26PM +0200, Ladislav Lhotka wrote:
On Wed, 2019-07-31 at 08:29 -0400, Theodore Y. Ts'o wrote:
Git is a distributed system. As such, the same git repository can
be made available on multiple git servers. For example:
[...]
Are all mirrors of each other. So even if/when github.com were to cut
off access to some country, people still have access from the other
two git repos.
For the git part, yes. What's more problematic are extra things that are not
replicated, e.g. issues.
It's easy enough to host issues within a normal Git repo. In fact,
there are tools out there for doing just that[0]. Typically an issue's
Of course, but in my experience it is not how IETF WGs use GitHub. As a matter
of fact, draft-ietf-git-using-github-00 describes the use of GitHub's issues.
"number" then becomes a commit hash, an abbreviated commit hash, since
there's no easy way to generate sequential issue numbers in Git, but I
think that's quite fine. Really, each issue can be its own ref, and you
can have one periodically-rebuilt ref to list all known issues so they
are easy to find.
One can even run such a Git-based issues system even while using GH/GL/
whatever, and one could even have a bi-di issues sync with the GH issues
so as to still be able to use the GH BUI while being able to access
issues via Git alone and w/o reference to the GH APIs or BUIs.
Wikis too can be hosted inside Git, as can web pages. Indeed, GitHub
does that for both. Oddly they chose to use the same repo for the web
pages (on a branch named gh-pages) but a different repo (with related
URI) for the wiki. Odd choices, but the point remains.
All of a WG's work can be hosted in a Git repo in a way that is
independent of the enhanced-Git GitHubs/GitLabs/Bitbuckets/... of the
world. All we need is Git. And since some of these tools exist
already, we may not even need to invest much in IETF-specific tooling.
Personally, I would be absolutely fine with such a simple solution. But then, I
am also happy with editing xml2rfc XML directly in emacs. :-)
I also use emacs to edit the xml and process on xml2rfc.ietf.org, and
some times I store the file on github.
In the process I loose accents from peoples' names: ä from Harri, ø from
Troan and the entire 神明 達哉. I cant figure out where's the error:
xml2rfc.ietf.org, github or emacs encode-coding-region utf-8.
I have tried a few different values for encode-coding-region (utf-16,
jp-*, iso-8859-*) but to no avail.
Alex
Lada
Nico
[0] E.g., https://github.com/dspinellis/git-issue -- note the "related
work" links at the bottom.
