android envsetup: Remove --host-os flag.

android envsetup: Remove --host-os flag.

Clients should set the host_os gyp variable instead. (As far as I can tell,
this flag was added for the webview build, and this CL updates gyp_webview
to use the gyp define instead. I haven't found other clients.)

Since several gyp targets check the value of host_os in a
target_conditions block and most of the build/common.gypi variables
aren't available during target_conditions processing time, use the same
trick that chromium_code uses to make host_os available at target_conditions
time.

BUG=330631
NOTRY=true

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=251138

[Torne, 2014-02-11 10:37:09]

lgtm

[Nico, 2014-02-13 06:27:07]

The CQ bit was checked by thakis@chromium.org

[commit-bot: I haz the power, 2014-02-13 06:28:04]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/thakis@chromium.org/153623011/330001

[commit-bot: I haz the power, 2014-02-13 07:52:55]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-02-13 07:52:56]

Retried try job too often on win_rel for step(s) content_unittests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[Nico, 2014-02-13 18:27:46]

The CQ bit was checked by thakis@chromium.org

[commit-bot: I haz the power, 2014-02-13 18:28:44]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/thakis@chromium.org/153623011/330001

[commit-bot: I haz the power, 2014-02-13 19:57:18]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/thakis@chromium.org/153623011/330001

[commit-bot: I haz the power, 2014-02-13 21:48:29]

Change committed as 251138

[Torne, 2014-02-21 11:32:49]

Sorry about this, I missed an issue with this cl.

https://codereview.chromium.org/153623011/diff/330001/build/android/envsetup_...
File build/android/envsetup_functions.sh (right):

https://codereview.chromium.org/153623011/diff/330001/build/android/envsetup_...
build/android/envsetup_functions.sh:172: '${ANDROID_SDK_ROOT}/../tools/' +
sys.platform.rstrip('23'), \
Unfortunately this broke downstream mac builds and I didn't notice.

We generate all the makefiles in one go, on linux, even for mac hosts, so
sys.platform here is always "linux[23]", never "darwin". So, mac is now trying
to run the linux version of the sdk tools and failing. We need to respect
host_os and not detect anything based on the current platform :/

To make matters more irritating, this CL made the M34 branch, but a fix on trunk
will not, so we probably need to fix this separately on M34. I'll see if I have
time to deal with that later today..

[Nico, 2014-02-21 15:07:43]

https://codereview.chromium.org/153623011/diff/330001/build/android/envsetup_...
File build/android/envsetup_functions.sh (right):

https://codereview.chromium.org/153623011/diff/330001/build/android/envsetup_...
build/android/envsetup_functions.sh:172: '${ANDROID_SDK_ROOT}/../tools/' +
sys.platform.rstrip('23'), \
On 2014/02/21 11:32:49, Torne wrote:
> Unfortunately this broke downstream mac builds and I didn't notice.
> 
> We generate all the makefiles in one go, on linux, even for mac hosts, so
> sys.platform here is always "linux[23]", never "darwin". So, mac is now trying
> to run the linux version of the sdk tools and failing. We need to respect
> host_os and not detect anything based on the current platform :/

D'oh, sorry!

> To make matters more irritating, this CL made the M34 branch, but a fix on
trunk
> will not, so we probably need to fix this separately on M34. I'll see if I
have
> time to deal with that later today..

Can't you do the regular merge process? (File a bug, fix on trunk, set
merge-requested on bug, drover over to branch)?

[Torne, 2014-02-21 15:10:01]

On 2014/02/21 15:07:43, Nico wrote:
>
https://codereview.chromium.org/153623011/diff/330001/build/android/envsetup_...
> File build/android/envsetup_functions.sh (right):
> 
>
https://codereview.chromium.org/153623011/diff/330001/build/android/envsetup_...
> build/android/envsetup_functions.sh:172: '${ANDROID_SDK_ROOT}/../tools/' +
> sys.platform.rstrip('23'), \
> On 2014/02/21 11:32:49, Torne wrote:
> > Unfortunately this broke downstream mac builds and I didn't notice.
> > 
> > We generate all the makefiles in one go, on linux, even for mac hosts, so
> > sys.platform here is always "linux[23]", never "darwin". So, mac is now
trying
> > to run the linux version of the sdk tools and failing. We need to respect
> > host_os and not detect anything based on the current platform :/
> 
> D'oh, sorry!
> 
> > To make matters more irritating, this CL made the M34 branch, but a fix on
> trunk
> > will not, so we probably need to fix this separately on M34. I'll see if I
> have
> > time to deal with that later today..
> 
> Can't you do the regular merge process? (File a bug, fix on trunk, set
> merge-requested on bug, drover over to branch)?

Yes, but the same CL probably won't apply on both as I don't think *all* of your
envsetup refactoring made the M34 branch point. If I'm mistaken then a direct
merge may work :)