On Fri, Mar 11, 2011 at 03:40:50PM -0800, Sarah Sharp wrote:
> On Fri, Mar 11, 2011 at 02:21:29PM -0800, Dmitry Torokhov wrote:
> > On Fri, Mar 11, 2011 at 01:36:25PM -0800, Sarah Sharp wrote:
> > > Ok, no worries. It's just that we're very careful to make sure that the
> > > Signed-off-by email addresses match the actual email of the sender.
> > > It's a contract according to Documentation/SubmittingPatches (see #12),
> > > and we want to make sure the email address is correct. I don't think a
> > > Reviewed-by tag is as big deal as a Signed-off-by tag, but I just
> > > thought I'd check which email to use. :)
> >
> > Hm, is it Intel's internal policy? Because as far as I know DCO requires
> > using one's real name and simply a reachable e-mail address; at least
> > most of my patches are signed off using dtor@xxxxxxx but sent from (or
> > committed as) gmail.com address.
>
> No, it's not an Intel policy, it's something you'll run into from most
> Linux kernel maintainers. I know Greg KH harassed me the first time I
> sent a patch with a Signed-off-by line that didn't match the From
> address. And if you've sent me patches before that didn't have matching
> addresses, I've probably only queued them without hassle because I
> didn't notice the mismatch.
I am not sure why Greg did that to you, as I said DCO requires usable
address, that's it.
So I got curious how many people besides myself are sending patches
from addresses other than their "master" address and here we are (for
commits between 2.6.18..2.6.37):
Name E-mail Signoff email Count
-----------------------------------------------------------------------------------------------------
Russell King rmk@xxxxxxxxxxxxxxxxxxxxxxx rmk+kernel@xxxxxxxxxxxxxxxx : 647
David S. Miller davem@xxxxxxxxxxxxxxxxxxxx davem@xxxxxxxxxxxxx : 537
Steven Rostedt srostedt@xxxxxxxxxx rostedt@xxxxxxxxxxx : 286
Dmitry Torokhov dmitry.torokhov@xxxxxxxxx dtor@xxxxxxx : 262
Linus Torvalds torvalds@xxxxxxxxxxxxxxxxxxxxxxxxxx torvalds@xxxxxxxxxxxxxxxxxxxx : 142
Dmitry Torokhov dtor@xxxxxxxxxxxxx dtor@xxxxxxx : 116
Jeff Garzik jeff@xxxxxxxxxx jgarzik@xxxxxxxxxx : 109
Lennert Buytenhek buytenh@xxxxxxxxxxxxxx buytenh@xxxxxxxxxxx : 69
Ben Dooks ben@xxxxxxxxxxxx ben-linux@xxxxxxxxx : 65
Dave Airlie airlied@xxxxxxxx airlied@xxxxxxxxxx : 64
Nicolas Pitre nico@xxxxxxx nico@xxxxxxxxxxx : 51
Jeremy Fitzhardinge jeremy@xxxxxxxx jeremy.fitzhardinge@xxxxxxxxxx: 46
Matthew Wilcox matthew@xxxxxx willy@xxxxxxxxxxxxxxx : 43
Mauro Carvalho Chehab mchehab@xxxxxxxxxxxxx mchehab@xxxxxxxxxx : 32
Andi Kleen andi@xxxxxxxxxxxxxx ak@xxxxxxxxxxxxxxx : 31
Linus Torvalds torvalds@xxxxxxxxxxxxxx torvalds@xxxxxxxx : 31
...
As you can see it is quite common practice so please do not invent new
rules.
Attached is a quick-n-dirty script I used to generate the list.
Thanks.
--
Dmitry
#!/usr/bin/python
from git import *
import email.utils
import re
import operator
repo = Repo("~dtor/kernel/work")
authors = dict()
for commit in Commit.iter_items(repo, "v2.6.18..v2.6.37"):
signoff = re.search("Signed-off-by: (" + re.escape(commit.author.name) + ".*)$", commit.message)
if signoff:
s_name, s_email = email.utils.parseaddr(signoff.group(1))
if commit.author.email.lower() != s_email.lower():
a = (commit.author.name, commit.author.email, s_email)
authors[a] = authors.get(a, 0) + 1
fmt = '{0:24} {1:30} {2:30}: {3:>6}'
print "{0:24} {1:30} {2:30} {3:6}".format('Name', 'E-mail', 'Signoff email', 'Count')
print "-----------------------------------------------------------------------------------------------------"
for a in sorted(authors.iteritems(), key=operator.itemgetter(1), reverse=True):
print fmt.format(a[0][0], a[0][1], a[0][2], a[1])
