Re: [Qemu-devel] [PULL for-2.9 0/9] virtio, vhost, pc: fixes

[Michael S. Tsirkin]

On Fri, Jan 06, 2017 at 02:38:32PM -0600, Eric Blake wrote:
> On 01/06/2017 02:04 PM, Michael S. Tsirkin wrote:
> > On Sat, Dec 17, 2016 at 11:18:03AM +0000, Peter Maydell wrote:
> >> On 16 December 2016 at 21:32,  <address@hidden> wrote:
> >>> Your series seems to have some coding style problems. See output below for
> >>> more information:
> >>>
> >>> Type: series
> >>> Subject: [Qemu-devel] [PULL for-2.9 0/9] virtio, vhost, pc: fixes
> >>> Message-id: address@hidden
> >>
> >>> Checking PATCH 5/9: contrib: add libvhost-user...
> >>> ERROR: Invalid UTF-8, patch and commit message should be encoded in UTF-8
> >>> #75: FILE: contrib/libvhost-user/libvhost-user.c:9:
> >>> + *  Marc-Andr Lureau <address@hidden>
> >>>               ^
> >>>
> >>> ERROR: Invalid UTF-8, patch and commit message should be encoded in UTF-8
> >>> #1579: FILE: contrib/libvhost-user/libvhost-user.h:8:
> >>> + *  Marc-Andr Lureau <address@hidden>
> >>>               ^
> >>>
> >>> total: 2 errors, 0 warnings, 1950 lines checked
> >>
> >> Michael, could you check the encoding in that file and respin
> >> the patchset if necessary, please?
> >>
> >> thanks
> >> -- PMM
> > 
> > I have no idea what this means.
> 
> Look at the headers to patch 5/9:
> 
> Subject: [PULL for-2.9 5/9] contrib: add libvhost-user
> Message-ID: <address@hidden>
> References: <address@hidden>
> MIME-Version: 1.0
> Content-Type: text/plain; charset=iso-8859-1
> 
> Although that mail body says
> From: Marc-André Lureau <address@hidden>
> it was doing so with the single-byte character from the iso-8859-1
> character set, instead of the multibyte UTF-8 character set.
> 
> Compare that to another recent pull request from Gerd:
> 
> Date: Wed,  4 Jan 2017 12:21:20 +0100
> Message-Id: <address@hidden>
> In-Reply-To: <address@hidden>
> References: <address@hidden>
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: quoted-printable
> 
> where the mail body says:
> From: Marc-Andr=C3=A9 Lureau <address@hidden>
> 
> So somehow, your git settings differ from Gerd, in that your 'git
> send-email' (or however you are sending your pull request) is not
> picking the correct encoding.  The buildbot is NOT seeing UTF-8, but
> ISO-8859-1 characters from your pull request, and is apparently unable
> to run iconv on the result before testing the patches out, resulting in
> the dropped bytes that are not already UTF-8 encoded.  My mail reader,
> on the other hand, is smart enough to recognize both encodings, and my
> mail window sees the name correctly formatted in both mails, with no
> visual indication that the sources for the two mails were in different
> encodings.  Which makes it a rather insidious problem to track down, if
> you don't spot the difference until dealing with some tool that isn't
> smart enough to do automatic conversion.
> 
> But how to force git to send mail in UTF-8 is not something I've had to
> figure out (I'm relying on my distro's defaults) - you'll have to poke
> around your .git/config (or ~/.gitconfig) and see if you have some
> really old settings from an earlier time, or ask around on freenode's
> #git channel for better ideas.
> 
> -- 
> Eric Blake   eblake redhat com    +1-919-301-3266
> Libvirt virtualization library http://libvirt.org
> 

So the email might be malformed, but git commits are ok I think.
Peter, could you pls merge? I'll investigate my email settings later.

-- 
MST