Support linking skia statically on Android

Support linking skia statically on Android

Support linking skia statically on Android. Skia symbols should be
exported by the app library, to the extent that skia_launcher needs
them.

Syncs libskia_android.so to the device only if it exists.

Committed: http://code.google.com/p/skia/source/detail?r=12399

[djsollen, 2013-11-07 17:03:16]

not fundamentally opposed to this change, but I'm not sure what the real benefit
is.  Can you convince me this is better than what we do currently?

[Kimmo Kinnunen, 2013-11-08 06:18:26]

On 2013/11/07 17:03:16, djsollen wrote:
> not fundamentally opposed to this change, but I'm not sure what the real
benefit
> is.  Can you convince me this is better than what we do currently?

Only soft rationales:

a) Desktop build of Skia provides two modes of linking -> this makes the Android
build Skia consistent with that (less surprises)

b) Existing Android Skia build system (gyp files) contains conditionals for
static and dynamic build, implying that both should work -> this makes the
static build actually work  (less surprises)

c) On Android, every GUI process has already platform Skia linked in when custom
Skia is being dynamically added. This is fairly undefined, as it is not
immediately (if ever?) clear whether or not the functions called by the custom
library / app are always bound correctly to the custom library entry points.
Supporting static build would provide a easy way for a self-doubting developer
to rule out one possible point of bugs, since custom app would have all its
skia-related calls bound statically.

[Kimmo Kinnunen, 2013-11-20 10:55:00]

On 2013/11/08 06:18:26, kkinnunen wrote:
> On 2013/11/07 17:03:16, djsollen wrote:
> > not fundamentally opposed to this change, but I'm not sure what the real
> benefit
> > is.  Can you convince me this is better than what we do currently?

And adding: it's not really trying to replace anything done currently, rather
fixing an existing mode.

Though, position can be taken that it adds few extra lines of code, and by that
it creates maintenance burden..

[djsollen, 2013-11-20 15:37:00]

I can concede that this will allow us to support a mode that we provide on other
platforms, but I am still worried about it bit rotting since we won't normally
exercise this path.

However, I don't have that strong of a conviction so I'll go ahead and accept
this for now with the expectation that if it does turn out to be a maintenance
burden we will remove it and document it as such.

lgtm

[commit-bot: I haz the power, 2013-11-21 06:00:49]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/kkinnunen@nvidia.com/61893008/1

[commit-bot: I haz the power, 2013-11-21 06:06:58]

Retried try job too often on Build-Ubuntu12-GCC-x86_64-Release-Trybot for
step(s) BuildBench, BuildEverything, BuildGm, BuildSkiaLib, BuildTests,
BuildTools
http://skiabot-master.pogerlabs.com:10117/buildstatus?builder=Build-Ubuntu12-...

[commit-bot: I haz the power, 2013-11-22 07:59:40]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/kkinnunen@nvidia.com/61893008/1

[commit-bot: I haz the power, 2013-11-22 08:04:50]

Retried try job too often on Build-Ubuntu12-GCC-x86_64-Release-Trybot for
step(s) BuildBench, BuildEverything, BuildGm, BuildSkiaLib, BuildTests,
BuildTools
http://skiabot-master.pogerlabs.com:10117/buildstatus?builder=Build-Ubuntu12-...

[commit-bot: I haz the power, 2013-11-22 08:10:48]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/kkinnunen@nvidia.com/61893008/190001

[commit-bot: I haz the power, 2013-11-22 08:15:00]

Retried try job too often on Build-Ubuntu12-GCC-x86_64-Release-Trybot for
step(s) BuildBench, BuildEverything, BuildGm, BuildSkiaLib, BuildTests,
BuildTools
http://skiabot-master.pogerlabs.com:10117/buildstatus?builder=Build-Ubuntu12-...

[Kimmo Kinnunen, 2013-11-22 08:19:13]

On 2013/11/22 08:15:00, I haz the power (commit-bot) wrote:
> Retried try job too often on Build-Ubuntu12-GCC-x86_64-Release-Trybot for
> step(s) BuildBench, BuildEverything, BuildGm, BuildSkiaLib, BuildTests,
> BuildTools
>
http://skiabot-master.pogerlabs.com:10117/buildstatus?builder=Build-Ubuntu12-...


I wonder who would know why does applying the patch fails for this patch?

[08:14:12.370796] error:
a/platform_tools/android/app/src/com/skia/SkiaSampleActivity.java: No such file
or directory

[djsollen, 2013-11-25 21:37:29]

lgtm

trying again as our infrastructure team thinks this may be fixed.

[commit-bot: I haz the power, 2013-11-25 21:37:35]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/kkinnunen@nvidia.com/61893008/190001

[commit-bot: I haz the power, 2013-11-25 21:43:30]

Retried try job too often on Build-Mac10.7-Clang-x86-Release-Trybot for step(s)
BuildBench, BuildEverything, BuildGm, BuildSkiaLib, BuildTests, BuildTools
http://skiabot-master.pogerlabs.com:10117/buildstatus?builder=Build-Mac10.7-C...

[djsollen, 2013-11-25 21:47:15]

looks like you need to rebase the change locally and upload it again.

[commit-bot: I haz the power, 2013-11-26 13:15:45]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/kkinnunen@nvidia.com/61893008/290001

[commit-bot: I haz the power, 2013-11-26 20:39:16]

Change committed as 12399