Fix some scan_sources.py dependencies

Fix some scan_sources.py dependencies

Search for include files relative to the
directory of the file that included them.

BUG=169539
TEST=manual, run gyp from chrome and verify includes are found

Committed: https://src.chromium.org/viewvc/native_client?view=rev&revision=10614

[Roland McGrath, 2013-01-11 21:03:25]

Technically, the directory containing the source file will only be searched for
#include "..." syntax, not #include <...> syntax.  So to make the script really
match the behavior of the compiler, it ought to distinguish the two.  Probably
it doesn't matter in practice, but it at least deserves a comment about the
known mismatch in behavior.

[elijahtaylor1, 2013-01-11 21:19:52]

I can add a comment if desired, but seeing as this script has always erred on
the side of over-including, I would consider this par for the course.

For example, it doesn't consider any defines when scanning source files so we
see things like mac/windows specific headers attempting to be resolved by this
native_client tool.

On 2013/01/11 21:03:25, Roland McGrath wrote:
> Technically, the directory containing the source file will only be searched
for
> #include "..." syntax, not #include <...> syntax.  So to make the script
really
> match the behavior of the compiler, it ought to distinguish the two.  Probably
> it doesn't matter in practice, but it at least deserves a comment about the
> known mismatch in behavior.

[elijahtaylor1, 2013-01-14 19:02:37]

ping

On 2013/01/11 21:19:52, elijahtaylor1 wrote:
> I can add a comment if desired, but seeing as this script has always erred on
> the side of over-including, I would consider this par for the course.
> 
> For example, it doesn't consider any defines when scanning source files so we
> see things like mac/windows specific headers attempting to be resolved by this
> native_client tool.
> 
> On 2013/01/11 21:03:25, Roland McGrath wrote:
> > Technically, the directory containing the source file will only be searched
> for
> > #include "..." syntax, not #include <...> syntax.  So to make the script
> really
> > match the behavior of the compiler, it ought to distinguish the two. 
Probably
> > it doesn't matter in practice, but it at least deserves a comment about the
> > known mismatch in behavior.

[binji, 2013-01-14 19:12:05]

lgt.m but you should probably get Noel to review.

[noelallen1, 2013-01-16 01:08:58]

lgtm

[commit-bot: I haz the power, 2013-01-16 19:46:38]

CQ is trying da patch. Follow status at
https://nativeclient-status.appspot.com/cq/elijahtaylor@chromium.org/11882002/1

[commit-bot: I haz the power, 2013-01-16 19:46:40]

Presubmit check for 11882002-1 failed and returned exit status 1.

Running presubmit commit checks ...
Error processing /b/commit-queue/workdir/nacl/build/scan_sources.py:
  Cannot execute command ['pychecker', '-stdlib']. failed reason: [Errno 2] No
such file or directory 

** Presubmit Warnings **
License must match:
.*? Copyright (\(c\) )?(2013|2012|2011|2010|2009|2008|2007|2006|2006-2008) The
Native Client Authors\. All rights reserved\.\n.*? Use of this source code is
governed by a BSD-style license that can be\n.*? found in the LICENSE file\.(?:
\*/)?\n
Found a bad license header in these files:
  build/scan_sources.py