GN: Get Mac build working for IPC target at least.

GN: Get Mac build working for IPC target at least.

- Move sysroot.gni up a directory, as it applies to multiple
platforms now.
- Use existing find_sdk.py script to locate Mac SDK.
- Fix some gtest includes for Mac files.

BUG=
R=brettw@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=243654

[bbudge, 2014-01-07 22:31:22]

https://codereview.chromium.org/116763005/diff/1/build/config/sysroot.gni
File build/config/sysroot.gni (right):

https://codereview.chromium.org/116763005/diff/1/build/config/sysroot.gni#new...
build/config/sysroot.gni:24: find_sdk_args = ["--verify", mac_sdk_min
"--sdk_path=" + mac_sdk_path]
I'm probably missing some escaping on the last argument. Not sure if this will
work anyway.

[brettw, 2014-01-08 00:24:20]

https://codereview.chromium.org/116763005/diff/50001/build/config/sysroot.gni
File build/config/sysroot.gni (right):

https://codereview.chromium.org/116763005/diff/50001/build/config/sysroot.gni...
build/config/sysroot.gni:22: declare_args() {
I think this should go in a separate mac SDK .gni file that this one imports.
The reason is that the SDK location doesn't really have anything to do with the
sysroot. If another BUILD.gn file wants the Mac SDK location, it makes more
sense for them to import the SDK .gni file thatn the sysroot one.

You can see my equivalent Windows one:
https://codereview.chromium.org/126073005/

So maybe call it build/config/mac/mac_sdk.gni (or something), move this code
there, and replace it with an import.

[bbudge, 2014-01-08 01:13:15]

https://codereview.chromium.org/116763005/diff/50001/build/config/sysroot.gni
File build/config/sysroot.gni (right):

https://codereview.chromium.org/116763005/diff/50001/build/config/sysroot.gni...
build/config/sysroot.gni:22: declare_args() {
On 2014/01/08 00:24:20, brettw wrote:
> I think this should go in a separate mac SDK .gni file that this one imports.
> The reason is that the SDK location doesn't really have anything to do with
the
> sysroot. If another BUILD.gn file wants the Mac SDK location, it makes more
> sense for them to import the SDK .gni file thatn the sysroot one.
> 
> You can see my equivalent Windows one:
> https://codereview.chromium.org/126073005/
> 
> So maybe call it build/config/mac/mac_sdk.gni (or something), move this code
> there, and replace it with an import.

Done.

[brettw, 2014-01-08 06:52:02]

https://codereview.chromium.org/116763005/diff/150001/build/config/sysroot.gni
File build/config/sysroot.gni (right):

https://codereview.chromium.org/116763005/diff/150001/build/config/sysroot.gn...
build/config/sysroot.gni:30: mac_sdk_version =
I was thinking everything up to here inclusive would also go into the .gni so
callers could get the SDK version and not just the path.

[bbudge, 2014-01-08 13:41:55]

https://codereview.chromium.org/116763005/diff/150001/build/config/sysroot.gni
File build/config/sysroot.gni (right):

https://codereview.chromium.org/116763005/diff/150001/build/config/sysroot.gn...
build/config/sysroot.gni:30: mac_sdk_version =
On 2014/01/08 06:52:02, brettw wrote:
> I was thinking everything up to here inclusive would also go into the .gni so
> callers could get the SDK version and not just the path.

Done.

[brettw, 2014-01-08 18:43:43]

LGTM, thanks!

https://codereview.chromium.org/116763005/diff/250001/build/config/linux/pkg_...
File build/config/linux/pkg_config.gni (right):

https://codereview.chromium.org/116763005/diff/250001/build/config/linux/pkg_...
build/config/linux/pkg_config.gni:5: import("../sysroot.gni")
Can you do a full path here? I've recently started doing full paths for imports
to match the C++ style (so //build/config/sysroot.gni)

https://codereview.chromium.org/116763005/diff/250001/build/config/mac/mac_sd...
File build/config/mac/mac_sdk.gni (right):

https://codereview.chromium.org/116763005/diff/250001/build/config/mac/mac_sd...
build/config/mac/mac_sdk.gni:9: # Path to a specific version of the Mac SDK.
Can you document the behavior if it's left empty? I think it would also be good
to say if this needs to have a slash at the end (looks like not).

https://codereview.chromium.org/116763005/diff/250001/build/config/mac/mac_sd...
build/config/mac/mac_sdk.gni:14: find_sdk_args = ["--verify", mac_sdk_min
"--sdk_path=" + mac_sdk_path]
Can you put spaces inside the [] here and below (but not for the "[0]")? I've
been doing that.

https://codereview.chromium.org/116763005/diff/250001/build/config/sysroot.gni
File build/config/sysroot.gni (right):

https://codereview.chromium.org/116763005/diff/250001/build/config/sysroot.gn...
build/config/sysroot.gni:24: sysroot = mac_sdk_path
Probably don't need so much vertical whitespace.

[bbudge, 2014-01-08 20:22:55]

https://codereview.chromium.org/116763005/diff/250001/build/config/linux/pkg_...
File build/config/linux/pkg_config.gni (right):

https://codereview.chromium.org/116763005/diff/250001/build/config/linux/pkg_...
build/config/linux/pkg_config.gni:5: import("../sysroot.gni")
On 2014/01/08 18:43:44, brettw wrote:
> Can you do a full path here? I've recently started doing full paths for
imports
> to match the C++ style (so //build/config/sysroot.gni)

Done.

https://codereview.chromium.org/116763005/diff/250001/build/config/mac/mac_sd...
File build/config/mac/mac_sdk.gni (right):

https://codereview.chromium.org/116763005/diff/250001/build/config/mac/mac_sd...
build/config/mac/mac_sdk.gni:9: # Path to a specific version of the Mac SDK.
On 2014/01/08 18:43:44, brettw wrote:
> Can you document the behavior if it's left empty? I think it would also be
good
> to say if this needs to have a slash at the end (looks like not).

Done.

https://codereview.chromium.org/116763005/diff/250001/build/config/sysroot.gni
File build/config/sysroot.gni (right):

https://codereview.chromium.org/116763005/diff/250001/build/config/sysroot.gn...
build/config/sysroot.gni:24: sysroot = mac_sdk_path
On 2014/01/08 18:43:44, brettw wrote:
> Probably don't need so much vertical whitespace.

I think the deleted line's wrap causes this to look like an extra line. The
actual code only has a newline after the 'import' statement, which I think helps
readability.

[commit-bot: I haz the power, 2014-01-08 20:26:51]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/bbudge@chromium.org/116763005/370001

[commit-bot: I haz the power, 2014-01-08 20:44:27]

Retried try job too often on linux_rel for step(s) base_unittests,
browser_tests, cacheinvalidation_unittests, cc_unittests, check_deps,
check_deps2git, chromedriver_unittests, components_unittests,
content_browsertests, content_unittests, crypto_unittests,
google_apis_unittests, gpu_unittests, interactive_ui_tests, ipc_tests,
jingle_unittests, media_unittests, nacl_integration, net_unittests,
ppapi_unittests, printing_unittests, remoting_unittests,
sandbox_linux_unittests, sql_unittests, sync_integration_tests, sync_unit_tests,
unit_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_rel&...

[bbudge, 2014-01-08 21:42:23]

Committed patchset #5 manually as r243654 (presubmit successful).