Adding utility method to test the version of the OS.

Adding utility method to test the version of the OS.

 This had method to check which version of the OS the application is
running.
 It also changes the ID generation mechanism to use identifierForVendor
on iOS6 so that the ID are persisted across application reinstallation.

BUG=None
TEST=None
R=stuartmorgan@chromium.org,mark@chromium.org

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=160414

[stuartmorgan, 2012-10-05 12:19:43]

Per discussion, we should use base/ios/ios_util to match the similar methods in
base/mac/mac_util

https://chromiumcodereview.appspot.com/11031066/diff/7001/base/ios/OWNERS
File base/ios/OWNERS (right):

https://chromiumcodereview.appspot.com/11031066/diff/7001/base/ios/OWNERS#new...
base/ios/OWNERS:1: set noparent
This is discouraged; we shouldn't do it here.

https://chromiumcodereview.appspot.com/11031066/diff/7001/base/ios/device_uti...
File base/ios/device_util.mm (right):

https://chromiumcodereview.appspot.com/11031066/diff/7001/base/ios/device_uti...
base/ios/device_util.mm:8: #import <Foundation/Foundation.h>
I believe UIKit includes Foundation, so this can probably go.

[Mark Mentovai, 2012-10-05 12:49:04]

Plus Stuart’s comments.

https://chromiumcodereview.appspot.com/11031066/diff/7002/base/ios/OWNERS
File base/ios/OWNERS (right):

https://chromiumcodereview.appspot.com/11031066/diff/7002/base/ios/OWNERS#new...
base/ios/OWNERS:2: qsr@chromium.org
This isn’t a file copy, can you make git not act like it is? (I think Rohit
knows how to do this, I’ve seen this before with him.)

https://chromiumcodereview.appspot.com/11031066/diff/7002/base/sys_info.h
File base/sys_info.h (right):

https://chromiumcodereview.appspot.com/11031066/diff/7002/base/sys_info.h#new...
base/sys_info.h:112: static bool IsRunningOnOrLater(int major, int minor, int
bugFix);
bug_fix

https://chromiumcodereview.appspot.com/11031066/diff/7002/base/sys_info_ios.mm
File base/sys_info_ios.mm (right):

https://chromiumcodereview.appspot.com/11031066/diff/7002/base/sys_info_ios.m...
base/sys_info_ios.mm:91: static int32 digits[3] = { 0, 0, 0 };
You know how I feel about these. I’d prefer to just initialize the static with a
call to a simple adapter function, like:

  static int32* digits =
OperatingSystemVersionNumbersPointerToArray_PickABetterName();

https://chromiumcodereview.appspot.com/11031066/diff/7002/base/sys_info_ios.m...
base/sys_info_ios.mm:93: OperatingSystemVersionNumbers(&digits[0], &digits[1],
&digits[2]);
mac_util uses uname for its runtime OS version checks like IsOSLionOrLater.
There’s a comment in base/mac/mac_util.mm DarwinMajorVersionInternal that
explains what the story is, but uname is faster and more side-effect-free than
other methods of getting the version. It’s fine if that’s not a concern on iOS,
but I thought I’d point it out.

https://chromiumcodereview.appspot.com/11031066/diff/7002/base/sys_info_ios.m...
base/sys_info_ios.mm:95: if (digits[0] < major)
Seems like this could be written as a loop. Or would a Version (base/version.h)
be appropriate?

[qsr, 2012-10-05 14:53:55]

http://codereview.chromium.org/11031066/diff/7001/base/ios/OWNERS
File base/ios/OWNERS (right):

http://codereview.chromium.org/11031066/diff/7001/base/ios/OWNERS#newcode1
base/ios/OWNERS:1: set noparent
On 2012/10/05 12:19:43, stuartmorgan wrote:
> This is discouraged; we shouldn't do it here.

Done.

http://codereview.chromium.org/11031066/diff/7001/base/ios/device_util.mm
File base/ios/device_util.mm (right):

http://codereview.chromium.org/11031066/diff/7001/base/ios/device_util.mm#new...
base/ios/device_util.mm:8: #import <Foundation/Foundation.h>
On 2012/10/05 12:19:43, stuartmorgan wrote:
> I believe UIKit includes Foundation, so this can probably go.

Done.

http://codereview.chromium.org/11031066/diff/7002/base/ios/OWNERS
File base/ios/OWNERS (right):

http://codereview.chromium.org/11031066/diff/7002/base/ios/OWNERS#newcode2
base/ios/OWNERS:2: qsr@chromium.org
On 2012/10/05 12:49:05, Mark Mentovai wrote:
> This isn’t a file copy, can you make git not act like it is? (I think Rohit
> knows how to do this, I’ve seen this before with him.)

Done.

http://codereview.chromium.org/11031066/diff/7002/base/sys_info.h
File base/sys_info.h (right):

http://codereview.chromium.org/11031066/diff/7002/base/sys_info.h#newcode112
base/sys_info.h:112: static bool IsRunningOnOrLater(int major, int minor, int
bugFix);
On 2012/10/05 12:49:05, Mark Mentovai wrote:
> bug_fix

Done.

http://codereview.chromium.org/11031066/diff/7002/base/sys_info_ios.mm
File base/sys_info_ios.mm (right):

http://codereview.chromium.org/11031066/diff/7002/base/sys_info_ios.mm#newcode91
base/sys_info_ios.mm:91: static int32 digits[3] = { 0, 0, 0 };
On 2012/10/05 12:49:05, Mark Mentovai wrote:
> You know how I feel about these. I’d prefer to just initialize the static with
a
> call to a simple adapter function, like:
> 
>   static int32* digits =
> OperatingSystemVersionNumbersPointerToArray_PickABetterName();

Done.

http://codereview.chromium.org/11031066/diff/7002/base/sys_info_ios.mm#newcode93
base/sys_info_ios.mm:93: OperatingSystemVersionNumbers(&digits[0], &digits[1],
&digits[2]);
On 2012/10/05 12:49:05, Mark Mentovai wrote:
> mac_util uses uname for its runtime OS version checks like IsOSLionOrLater.
> There’s a comment in base/mac/mac_util.mm DarwinMajorVersionInternal that
> explains what the story is, but uname is faster and more side-effect-free than
> other methods of getting the version. It’s fine if that’s not a concern on
iOS,
> but I thought I’d point it out.

 On iOS, we are calling: [[UIDevice currentDevice] systemVersion]

 1) I don't think this will call Getstalt, as it returns a fully constructed
version, not individual element and as this value is cached here, I'm not that
worried.

http://codereview.chromium.org/11031066/diff/7002/base/sys_info_ios.mm#newcode95
base/sys_info_ios.mm:95: if (digits[0] < major)
 I'll do a loop. Version is working on string. I do not want to transform that
on string before sending that to Version that will cut it back.

[Mark Mentovai, 2012-10-05 15:03:51]

If it LGT Stuart, it LGTM with these small changes.

http://codereview.chromium.org/11031066/diff/9017/base/ios/ios_util.mm
File base/ios/ios_util.mm (right):

http://codereview.chromium.org/11031066/diff/9017/base/ios/ios_util.mm#newcode13
base/ios/ios_util.mm:13: static int32 digits[3];
For a single-use function whose return value is stuffed into a static itself,
I’d use “new” instead of “static” here.

http://codereview.chromium.org/11031066/diff/9017/base/ios/ios_util.mm#newcode18
base/ios/ios_util.mm:18: }
}  // namespace

http://codereview.chromium.org/11031066/diff/9017/base/ios/ios_util.mm#newcode33
base/ios/ios_util.mm:33: int32 version[3] = { major, minor, bug_fix };
[3] not necessary.

http://codereview.chromium.org/11031066/diff/9017/base/ios/ios_util.mm#newcode34
base/ios/ios_util.mm:34: for (int i = 0; i < 3; i++) {
Use arraysize instead of 3.

[stuartmorgan, 2012-10-05 15:17:51]

LGTM with nits

http://codereview.chromium.org/11031066/diff/6014/base/base.gypi
File base/base.gypi (right):

http://codereview.chromium.org/11031066/diff/6014/base/base.gypi#newcode166
base/base.gypi:166: 'ios/scoped_critical_action.mm',
Why are you moving this out of alphabetical order?

http://codereview.chromium.org/11031066/diff/6014/base/ios/ios_util.h
File base/ios/ios_util.h (right):

http://codereview.chromium.org/11031066/diff/6014/base/ios/ios_util.h#newcode14
base/ios/ios_util.h:14: // Running on iOS 5 or later?
Please follow Google style for these comments.

[qsr, 2012-10-05 15:33:16]

http://codereview.chromium.org/11031066/diff/9017/base/ios/ios_util.mm
File base/ios/ios_util.mm (right):

http://codereview.chromium.org/11031066/diff/9017/base/ios/ios_util.mm#newcode13
base/ios/ios_util.mm:13: static int32 digits[3];
On 2012/10/05 15:03:51, Mark Mentovai wrote:
> For a single-use function whose return value is stuffed into a static itself,
> I’d use “new” instead of “static” here.

Done.

http://codereview.chromium.org/11031066/diff/9017/base/ios/ios_util.mm#newcode18
base/ios/ios_util.mm:18: }
On 2012/10/05 15:03:51, Mark Mentovai wrote:
> }  // namespace

Done.

http://codereview.chromium.org/11031066/diff/9017/base/ios/ios_util.mm#newcode33
base/ios/ios_util.mm:33: int32 version[3] = { major, minor, bug_fix };
On 2012/10/05 15:03:51, Mark Mentovai wrote:
> [3] not necessary.

Done.

http://codereview.chromium.org/11031066/diff/9017/base/ios/ios_util.mm#newcode34
base/ios/ios_util.mm:34: for (int i = 0; i < 3; i++) {
On 2012/10/05 15:03:51, Mark Mentovai wrote:
> Use arraysize instead of 3.

Done.

http://codereview.chromium.org/11031066/diff/6014/base/base.gypi
File base/base.gypi (right):

http://codereview.chromium.org/11031066/diff/6014/base/base.gypi#newcode166
base/base.gypi:166: 'ios/scoped_critical_action.mm',
 I was putting that in order :).

 Sorry, don't know why, I was seeing that as beeing critical_xxx

http://codereview.chromium.org/11031066/diff/6014/base/ios/ios_util.h
File base/ios/ios_util.h (right):

http://codereview.chromium.org/11031066/diff/6014/base/ios/ios_util.h#newcode14
base/ios/ios_util.h:14: // Running on iOS 5 or later?
On 2012/10/05 15:17:52, stuartmorgan wrote:
> Please follow Google style for these comments.

Done.

[commit-bot: I haz the power, 2012-10-05 15:33:26]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/qsr@chromium.org/11031066/1018

[commit-bot: I haz the power, 2012-10-05 17:49:03]

Change committed as 160414