hack hack hack on android scripts

hack hack hack

Edit down the android scripts a bit.
  - set -e
  - detect ccache by default
  - -t for BUILDTYPE

I figure the -t changes are going to break some bots.

BUG=skia:

Committed: http://code.google.com/p/skia/source/detail?r=14061

[epoger, 2014-04-03 17:13:57]

If/when you commit this, please give it a more descriptive title to show up in
the list of commits...

[djsollen, 2014-04-03 17:24:23]

https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
File platform_tools/android/bin/android_make (right):

https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
platform_tools/android/bin/android_make:3: set -e
add the comment back to why this is useful.

# Fail-fast if anything in the script fails.

https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
File platform_tools/android/bin/android_setup.sh (right):

https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
platform_tools/android/bin/android_setup.sh:5: set -e
why set it here and in android_make?

[mtklein, 2014-04-03 17:30:17]

https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
File platform_tools/android/bin/android_make (right):

https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
platform_tools/android/bin/android_make:3: set -e
On 2014/04/03 17:24:23, djsollen wrote:
> add the comment back to why this is useful.
> 
> # Fail-fast if anything in the script fails.

Done.  Though this is sort of like

// Increment by 1.
++x;

https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
File platform_tools/android/bin/android_setup.sh (right):

https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
platform_tools/android/bin/android_setup.sh:5: set -e
On 2014/04/03 17:24:23, djsollen wrote:
> why set it here and in android_make?

android_make's set -e covers this when sourced from android_make, but we're
relying on it specifically in here too.  This keeps it safe when sourced from
places like android_run_skia that don't set -e themselves.

[djsollen, 2014-04-03 17:54:32]

lg2m pending trybot success and description update.

https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
File platform_tools/android/bin/android_setup.sh (right):

https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
platform_tools/android/bin/android_setup.sh:5: set -e
So I see why we would want this here so that everything that includes it gets it
set, but why do we still need it in android_make at all.

[mtklein, 2014-04-03 17:59:46]

https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
File platform_tools/android/bin/android_setup.sh (right):

https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
platform_tools/android/bin/android_setup.sh:5: set -e
On 2014/04/03 17:54:33, djsollen wrote:
> So I see why we would want this here so that everything that includes it gets
it
> set, but why do we still need it in android_make at all.

For the build.  There's no point running ninja if gyp fails, that sort of thing.

set -e should just be the first line of all scripts.  There's no benefit trying
to make guarantees about who will set it... everyone who cares can just set it.

[djsollen, 2014-04-03 18:02:20]

On 2014/04/03 17:59:46, mtklein wrote:
>
https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
> File platform_tools/android/bin/android_setup.sh (right):
> 
>
https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
> platform_tools/android/bin/android_setup.sh:5: set -e
> On 2014/04/03 17:54:33, djsollen wrote:
> > So I see why we would want this here so that everything that includes it
gets
> it
> > set, but why do we still need it in android_make at all.
> 
> For the build.  There's no point running ninja if gyp fails, that sort of
thing.
> 
> set -e should just be the first line of all scripts.  There's no benefit
trying
> to make guarantees about who will set it... everyone who cares can just set
it.

I agree that it should be set in android_make, but since every android script
sources android_setup.sh (perhaps we should rename it common_includes.sh) I
didn't know if we needed it explicitly listed in android.mk.  Though it isn't a
big deal that it is.

[mtklein, 2014-04-03 18:04:06]

On 2014/04/03 18:02:20, djsollen wrote:
> On 2014/04/03 17:59:46, mtklein wrote:
> >
>
https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
> > File platform_tools/android/bin/android_setup.sh (right):
> > 
> >
>
https://codereview.chromium.org/223613004/diff/20001/platform_tools/android/b...
> > platform_tools/android/bin/android_setup.sh:5: set -e
> > On 2014/04/03 17:54:33, djsollen wrote:
> > > So I see why we would want this here so that everything that includes it
> gets
> > it
> > > set, but why do we still need it in android_make at all.
> > 
> > For the build.  There's no point running ninja if gyp fails, that sort of
> thing.
> > 
> > set -e should just be the first line of all scripts.  There's no benefit
> trying
> > to make guarantees about who will set it... everyone who cares can just set
> it.
> 
> I agree that it should be set in android_make, but since every android script
> sources android_setup.sh (perhaps we should rename it common_includes.sh) I
> didn't know if we needed it explicitly listed in android.mk.  Though it isn't
a
> big deal that it is.

It's an implementation detail of android_setup.sh that it sets set -e.

[djsollen, 2014-04-04 16:06:17]

lgtm

[mtklein, 2014-04-04 16:06:27]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-04-04 16:06:33]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/mtklein@chromium.org/223613004/80001

[commit-bot: I haz the power, 2014-04-04 16:26:44]

Change committed as 14061