On Fri, 2012-04-20 at 16:00 -0700, Nishanth Aravamudan wrote:
> On 20.04.2012 [22:46:54 +0000], Spear, Raymond (Mission Critical Linux) wrote:
> > If this is the wrong place to ask this, I apologize.  But, it's the
> > only place I could find to ask.
> > 
> > Is there an approximate count of how many people are contributing
> > patches worldwide to autotest?  If there is, the count broken down by
> > core vs the tests themselves?
> > 
> > I've done git short-log -s but that seems to be too few people to be
> > all patch contributors.
> 
> -s just tells you the Authors. You need all the contributors:
> 
> git log | grep 'Signed-off-by' | sort | uniq | wc -l
> 220
> 
> That is probably closer, but may still miss things like Tested-by,
> Reported-by, etc.

Well, I was asked about something similar the previous week.

So, bottom line, this is *mostly* the result of keeping the code on an
svn repo for 6 years and doing a git conversion through a simple git
mirror.

In all respects, we're doing a different thing than the Linux kernel,
specially because of the code being kept on svn for so long, so in
autotest, the defacto authorship of a patch is the signed-off-by: line.

Now, I suppose we could write a script of some sort to rewrite the
entire project history to ensure all patches have a from: field
correspondent to the Signed-off-by: line.

Downsides are that 3rd parties having their own autotest trees that are
rebased from time to time might have some hard time rebasing.

Once rebased, then we have to adopt a stricter policy about preserving
the authorship (the from: field) on patches coming to the mailing list.
I can say I am guilty of hand editing/rebasing patches so they end up
having myself as the 'author'. 

Well, I can tell that this was not made on bad faith, only that I didn't
really understand that (from:) is the author field, and that was the
convention used on the Linux kernel.

Thoughts?

_______________________________________________
Autotest mailing list
[email protected]
http://test.kernel.org/cgi-bin/mailman/listinfo/autotest

Reply via email to