[TUHS] A repository with 44 years of Unix evolution gets the MSR '15 Best Data Showcase Award

Steffen Nurpmeso sdaoden at yandex.com
Tue May 19 23:48:05 AEST 2015


Jacob Goense <dugo at xs4all.nl> wrote:
 |On 2015-05-19 12:04, Erik E. Fair wrote:
 |> The NetBSD source repository has not changed source control systems
 |> since project inception - we're still using CVS, and the whole thing
 |> can be browsed at anoncvs.netbsd.org. A bunch of stuff is now "in the
 |> attic" (nominally "deleted" from the repository, but so far as I know
 |> that just means it's not fetched by default with an unadorned "get"
 |> command - it's still in the repository archive).
 |> 
 |> The history you're looking to collect is available for copy any time.
 |
 |There is a pile of "revision #.# intentionally removed" in there. See
 |http://www.onlamp.com/pub/a/bsd/2006/09/14/netbsd_future.html?page=2
 |for Hannum's story behind it.
 |
 |Deadly quote "and nobody cares about that early code history any more
 |--so this is all water under the bridge."
 |
 |They are available in a manilla folder through a coughing man in a
 |raincoat if you know in which parking garage you have to look. Not
 |sure what the consequences are if they appear as a pull request for
 |dspinellis's truly awesome unix-history-repo.

I also thought that was bad advise – Jörg Sonnenberger of NetBSD
converted the repository to Fossil.  I have lost that address; he
automatically converts _that_ to git(1) on github [1], which
i track myself, however.

Note that he is the one who is developing the used converter, and
that in turn had some errors in the past, resulting in history to
become invalid.  That didn't happen for quite a while now (afaik!)
but since it was very weird about two years ago and occasionally
happened still last year one possibly should ask him before
integrating all those millions of objects into such a large
repository.

  [1] https://github.com/jsonn/src

--steffen



More information about the TUHS mailing list