Add support for MIDL include directories

Add support for MIDL include directories

This adds support for 'midl_include_dirs', which is handled like 'include_dirs'
and 'resource_include_dirs', except this is for the MIDL tool.

One notable difference is that, unlike 'resource_include_dirs', the
'midl_include_dirs' property doesn't default to 'include_dirs'.  This is to
prevent any change in behavior for any existing gyp files that don't set
'midl_include_dirs'.  Also, MIDL include directories are quite a different
concept to C/C++ include directories.

BUG=
R=scottmg@chromium.org

Committed: https://code.google.com/p/gyp/source/detail?r=1986

[Shezan Baig (Bloomberg), 2014-09-25 16:28:00]

sbaig1@bloomberg.net changed reviewers:
+ scottmg@chromium.org

[scottmg, 2014-09-25 16:53:10]

https://codereview.chromium.org/602073003/diff/1/pylib/gyp/generator/msvs.py
File pylib/gyp/generator/msvs.py (right):

https://codereview.chromium.org/602073003/diff/1/pylib/gyp/generator/msvs.py#...
pylib/gyp/generator/msvs.py:1195: midl_include_dirs =
config.get('midl_include_dirs', [])
should msvs_system_include_dirs be included here for the common ones like
ObjIdl.idl, etc?

https://codereview.chromium.org/602073003/diff/1/test/win/idl-includedirs/sub...
File test/win/idl-includedirs/subdir/bar.idl (right):

https://codereview.chromium.org/602073003/diff/1/test/win/idl-includedirs/sub...
test/win/idl-includedirs/subdir/bar.idl:5: import "oaidl.idl";
(or i guess this one)

https://codereview.chromium.org/602073003/diff/1/test/win/idl-includedirs/sub...
File test/win/idl-includedirs/subdir/foo.idl (right):

https://codereview.chromium.org/602073003/diff/1/test/win/idl-includedirs/sub...
test/win/idl-includedirs/subdir/foo.idl:6: import "bar.idl";
would this have been found based on "." probably being in the default search
locations anyway?

[Shezan Baig (Bloomberg), 2014-09-25 17:00:21]

https://codereview.chromium.org/602073003/diff/1/pylib/gyp/generator/msvs.py
File pylib/gyp/generator/msvs.py (right):

https://codereview.chromium.org/602073003/diff/1/pylib/gyp/generator/msvs.py#...
pylib/gyp/generator/msvs.py:1195: midl_include_dirs =
config.get('midl_include_dirs', [])
On 2014/09/25 16:53:10, scottmg wrote:
> should msvs_system_include_dirs be included here for the common ones like
> ObjIdl.idl, etc?

Hmm, I didn't seem to need it in my test.  I can add it though, if you think it
is required.

https://codereview.chromium.org/602073003/diff/1/test/win/idl-includedirs/sub...
File test/win/idl-includedirs/subdir/foo.idl (right):

https://codereview.chromium.org/602073003/diff/1/test/win/idl-includedirs/sub...
test/win/idl-includedirs/subdir/foo.idl:6: import "bar.idl";
On 2014/09/25 16:53:10, scottmg wrote:
> would this have been found based on "." probably being in the default search
> locations anyway?

Did you mean '.' relative to the 'subdir'?  I don't think that works, because
the test would fail if I remove 'subdir' from 'midl_include_dirs'.

[scottmg, 2014-09-25 17:05:31]

https://codereview.chromium.org/602073003/diff/1/pylib/gyp/generator/msvs.py
File pylib/gyp/generator/msvs.py (right):

https://codereview.chromium.org/602073003/diff/1/pylib/gyp/generator/msvs.py#...
pylib/gyp/generator/msvs.py:1195: midl_include_dirs =
config.get('midl_include_dirs', [])
On 2014/09/25 17:00:21, Shezan Baig (Bloomberg) wrote:
> On 2014/09/25 16:53:10, scottmg wrote:
> > should msvs_system_include_dirs be included here for the common ones like
> > ObjIdl.idl, etc?
> 
> Hmm, I didn't seem to need it in my test.  I can add it though, if you think
it
> is required.

I think by default midl uses cl /E, which means it uses %INCLUDE% which is
probably set to include the system dirs on the bots
http://msdn.microsoft.com/en-us/library/windows/desktop/aa367299(v=vs.85).aspx

But, in general, it wouldn't have to be (and I know Nico had to do ... something
with the clang-win build because of Chrome's non-system toolchain install)

So, I think it would be sensible to include it in that list, if it doesn't cause
any problems.

https://codereview.chromium.org/602073003/diff/1/test/win/idl-includedirs/sub...
File test/win/idl-includedirs/subdir/foo.idl (right):

https://codereview.chromium.org/602073003/diff/1/test/win/idl-includedirs/sub...
test/win/idl-includedirs/subdir/foo.idl:6: import "bar.idl";
On 2014/09/25 17:00:21, Shezan Baig (Bloomberg) wrote:
> On 2014/09/25 16:53:10, scottmg wrote:
> > would this have been found based on "." probably being in the default search
> > locations anyway?
> 
> Did you mean '.' relative to the 'subdir'?  I don't think that works, because
> the test would fail if I remove 'subdir' from 'midl_include_dirs'.

Yes, that's what I was obtusely getting at. :) Thanks for confirming that it
doesn't work without the midl_include_dirs addition.

[Shezan Baig (Bloomberg), 2014-09-25 17:16:00]

https://codereview.chromium.org/602073003/diff/1/pylib/gyp/generator/msvs.py
File pylib/gyp/generator/msvs.py (right):

https://codereview.chromium.org/602073003/diff/1/pylib/gyp/generator/msvs.py#...
pylib/gyp/generator/msvs.py:1195: midl_include_dirs =
config.get('midl_include_dirs', [])
On 2014/09/25 17:05:31, scottmg wrote:
> On 2014/09/25 17:00:21, Shezan Baig (Bloomberg) wrote:
> > On 2014/09/25 16:53:10, scottmg wrote:
> > > should msvs_system_include_dirs be included here for the common ones like
> > > ObjIdl.idl, etc?
> > 
> > Hmm, I didn't seem to need it in my test.  I can add it though, if you think
> it
> > is required.
> 
> I think by default midl uses cl /E, which means it uses %INCLUDE% which is
> probably set to include the system dirs on the bots
> http://msdn.microsoft.com/en-us/library/windows/desktop/aa367299(v=vs.85).aspx
> 
> But, in general, it wouldn't have to be (and I know Nico had to do ...
something
> with the clang-win build because of Chrome's non-system toolchain install)
> 
> So, I think it would be sensible to include it in that list, if it doesn't
cause
> any problems.


Sounds good -- I've uploaded a new patch.  I copied exactly how 'include_dirs'
is initialized, replacing 'include_dirs' with 'midl_include_dirs'

[scottmg, 2014-09-25 17:17:50]

and then lgtm. Thanks!

https://codereview.chromium.org/602073003/diff/40001/pylib/gyp/msvs_emulation.py
File pylib/gyp/msvs_emulation.py (right):

https://codereview.chromium.org/602073003/diff/40001/pylib/gyp/msvs_emulation...
pylib/gyp/msvs_emulation.py:325: includes = midl_include_dirs + self._Setting(
similiarly here the same as above

[Shezan Baig (Bloomberg), 2014-09-25 19:00:10]

Committed patchset #4 (id:60001) manually as 1986 (presubmit successful).