Traceback with git-am

[Yann E. MORIN]

Bryan, All,

On 2015-01-13 22:40 -0800, Bryan Hundven spake thusly:
> On Thu, Jan 8, 2015 at 10:45 PM, Bryce Harrington <bryce at osg.samsung.com> wrote:
> > On Thu, Jan 08, 2015 at 07:54:41PM -0800, Bryan Hundven wrote:
> >> Hello list,
> >>
> >> While applying patches to my project with pwclient, I got this traceback.
> >> I'm not really that good with python, so I figured I'd post the
> >> traceback here and see if you can figure it out.
> >
> > In this case, it's just saying there was an error in the XML data sent
> > between the client and server, but doesn't indicate what that error
> > was... just that it was the second character on on line 171.
> >
> > Does the failure happen with just one specific patch, and does it occur
> > reliably with it?  If so, then it would probably help if you also posted
> > the patch.
> 
> It happens reliably with: https://patchwork.ozlabs.org/patch/424224/
> full log: http://pastebin.com/vPrQzEMR
> 
> I've never seen this with any patch I've applied.
> 
> I see that Yann (CC'd) has also found this issue. Maybe he can provide
> his example?
> https://lists.ozlabs.org/pipermail/patchwork/2012-November/000779.html

Yeah, I remember this issue.

The problem I encountered comes from a bug in libexpat, in which the
ASCII character 0x0c (line-feed) is considered invalid.

See the initial report to the Python bug tracker:
    http://bugs.python.org/issue11804
    http://bugs.python.org/issue11804#msg161701

Comment #161701 hints at a feature request in expat to support XML-1.1:
    http://sourceforge.net/tracker/?func=detail&atid=110127&aid=891265&group_id=10127

In patch 424224, we can see there are 6 line-feed chars, on lines: 127,
756, 766, 1409, 2038 and 2048. It does not seem to match the error that
is reported against line 171, though... But I noticed that the line
numbers are not always to be trusted, and that there can be an offset.

I am not really an XML pro, to say the least, so I am not able to judge
where the responsibility really lies: python-expat, expat or xmlrpclib?

So, maybe such character have to be somehow encoded server-side, and
then decoded client-side. Again, not sure where the responsibility lies...

Sorry, I can't help you further on this topic... :-(

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'