Factoring GUID generation from metrics to base

Factoring GUID generation from metrics to base

Factors GUID generation into base/rand_util.  The Autofill feature is in need of this utility so am factoring GUID generation out of metrics and moving to the commons.

BUG=58813
TEST=RandUtilTest.GUIDGeneratesAllZeroes, RandUtilTest.GUIDGeneratesCorrectly, RandUtilTest.GUIDCorrectlyFormatted, MetricsServiceTest.ClientIdCorrectlyFormatted

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=62480

[Mark Mentovai, 2010-10-13 20:18:32]

LGTM

http://codereview.chromium.org/3800001/diff/1007/5003
File base/rand_util_unittest.cc (right):

http://codereview.chromium.org/3800001/diff/1007/5003#newcode70
base/rand_util_unittest.cc:70: }
I’d also add a test that calls GenerateGUID twice and ensures that the result is
different each time.

Optional: do the GUIDCorrectlyFormatted and uniqueness tests a few times in a
loop.

[dhollowa, 2010-10-13 20:52:12]

http://codereview.chromium.org/3800001/diff/1007/5003
File base/rand_util_unittest.cc (right):

http://codereview.chromium.org/3800001/diff/1007/5003#newcode70
base/rand_util_unittest.cc:70: }
On 2010/10/13 20:18:33, Mark Mentovai wrote:
> I’d also add a test that calls GenerateGUID twice and ensures that the result
is
> different each time.
> 
> Optional: do the GUIDCorrectlyFormatted and uniqueness tests a few times in a
> loop.

Done.

[Mark Mentovai, 2010-10-13 20:55:41]

LGTM

[brettw, 2010-10-14 00:16:55]

Why did this need to get added to base? It's not used by anything but the Chrome
browser, so I'd prefer to keep it as narrowly scoped as possible. In this case,
it seems like it should go in browser or some subdirectory thereof.

[dhollowa, 2010-10-14 00:50:11]

On 2010/10/14 00:16:55, brettw wrote:
> Why did this need to get added to base? It's not used by anything but the
Chrome
> browser, so I'd prefer to keep it as narrowly scoped as possible. In this
case,
> it seems like it should go in browser or some subdirectory thereof.

It seemed an appropriately general facility.  But I'm happy to reduce scope.

I propose:

  chrome/browser/guid.h
  chrome/browser/guid_posix.cc
  chrome/browser/guid_unittest.cc
  chrome/browser/guid_win.cc

[brettw, 2010-10-14 00:51:29]

On Wed, Oct 13, 2010 at 5:50 PM,  <dhollowa@chromium.org> wrote:
> On 2010/10/14 00:16:55, brettw wrote:
>>
>> Why did this need to get added to base? It's not used by anything but the
>
> Chrome
>>
>> browser, so I'd prefer to keep it as narrowly scoped as possible. In this
>
> case,
>>
>> it seems like it should go in browser or some subdirectory thereof.
>
> It seemed an appropriately general facility.  But I'm happy to reduce scope.
>
> I propose:
>
>  chrome/browser/guid.h
>  chrome/browser/guid_posix.cc
>  chrome/browser/guid_unittest.cc
>  chrome/browser/guid_win.cc

SGTM. I'm trying to make base smaller since it's becoming
overwhelmingly large (and therefore less useful to non-Chrome
targets).

Brett

[dhollowa, 2010-10-14 00:54:57]

On 2010/10/14 00:51:29, brettw wrote:
> On Wed, Oct 13, 2010 at 5:50 PM,  <mailto:dhollowa@chromium.org> wrote:
> > On 2010/10/14 00:16:55, brettw wrote:
> >>
> >> Why did this need to get added to base? It's not used by anything but the
> >
> > Chrome
> >>
> >> browser, so I'd prefer to keep it as narrowly scoped as possible. In this
> >
> > case,
> >>
> >> it seems like it should go in browser or some subdirectory thereof.
> >
> > It seemed an appropriately general facility. &nbsp;But I'm happy to reduce
scope.
> >
> > I propose:
> >
> > &nbsp;chrome/browser/guid.h
> > &nbsp;chrome/browser/guid_posix.cc
> > &nbsp;chrome/browser/guid_unittest.cc
> > &nbsp;chrome/browser/guid_win.cc
> 
> SGTM. I'm trying to make base smaller since it's becoming
> overwhelmingly large (and therefore less useful to non-Chrome
> targets).
> 
> Brett
> 

NP.  Thanks for the drive-by.

[dhollowa, 2010-10-14 04:32:28]

Follow up here: http://codereview.chromium.org/3800003