Reimplement RequestNetworkServiceConnect, RequestRequirePin, RequestEnterPin, RequestEnterPin, Requ…

Reimplement RequestNetworkServiceConnect, RequestRequirePin, RequestEnterPin, RequestEnterPin, RequestChangePin, RequestCellularRegister without Libcros

BUG=chromium-os:16557
TEST=unit_tests --gtest_filter="CrosNetworkFunctions*"

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=133654

[hashimoto, 2012-04-23 06:59:38]

This patch depends on http://codereview.chromium.org/10170020/ ("Rework
chromeos::FlimflamDevice/ServiceClient with CallMethodWithErrorCallback ")

[satorux1, 2012-04-23 17:15:20]

http://codereview.chromium.org/10134027/diff/2001/chrome/browser/chromeos/cro...
File chrome/browser/chromeos/cros/cros_network_functions_unittest.cc (right):

http://codereview.chromium.org/10134027/diff/2001/chrome/browser/chromeos/cro...
chrome/browser/chromeos/cros/cros_network_functions_unittest.cc:360:
reinterpret_cast<MonitorDataPlanCallback>(42);  // Dummy value.
casting 42 to a callback object looks scary...

http://codereview.chromium.org/10134027/diff/2001/chrome/browser/chromeos/cro...
chrome/browser/chromeos/cros/cros_network_functions_unittest.cc:395: void*
object = reinterpret_cast<void*>(42);  // dummy
I'm guessing that you are using 42 because you cannot use NULL for this.

However, casting 42 to void* looks rather scary, as dereferencing the point
would likely cause a SEGV. It might be nicer to create a dummy object?

scoped_ptr<int> dummy_object(42);
void* object = dummy_object.get();

[hashimoto, 2012-04-24 04:29:46]

http://codereview.chromium.org/10134027/diff/2001/chrome/browser/chromeos/cro...
File chrome/browser/chromeos/cros/cros_network_functions_unittest.cc (right):

http://codereview.chromium.org/10134027/diff/2001/chrome/browser/chromeos/cro...
chrome/browser/chromeos/cros/cros_network_functions_unittest.cc:360:
reinterpret_cast<MonitorDataPlanCallback>(42);  // Dummy value.
On 2012/04/23 17:15:20, satorux1 wrote:
> casting 42 to a callback object looks scary...

Done.

http://codereview.chromium.org/10134027/diff/2001/chrome/browser/chromeos/cro...
chrome/browser/chromeos/cros/cros_network_functions_unittest.cc:395: void*
object = reinterpret_cast<void*>(42);  // dummy
On 2012/04/23 17:15:20, satorux1 wrote:
> I'm guessing that you are using 42 because you cannot use NULL for this.
> 
I can pass NULL, but I just want to make sure that the tested function does not
overwrite the argument with NULL.
> However, casting 42 to void* looks rather scary, as dereferencing the point
> would likely cause a SEGV. It might be nicer to create a dummy object?
> 
> scoped_ptr<int> dummy_object(42);
> void* object = dummy_object.get();
I found 'this' can be used as a void* argument and it does not cause SEGV.

[satorux1, 2012-04-24 06:05:54]

LGTM

http://codereview.chromium.org/10134027/diff/8004/chrome/browser/chromeos/cro...
File chrome/browser/chromeos/cros/cros_network_functions_unittest.cc (right):

http://codereview.chromium.org/10134027/diff/8004/chrome/browser/chromeos/cro...
chrome/browser/chromeos/cros/cros_network_functions_unittest.cc:370:
MonitorDataPlanCallback callback = &OnDataPlansUpdate;
maybe safer to add the class name ?

&CrosNetworkFunctionsLibcrosTest::OnDataPlansUpdate

http://codereview.chromium.org/10134027/diff/8004/chrome/browser/chromeos/cro...
chrome/browser/chromeos/cros/cros_network_functions_unittest.cc:404: void*
object = this;
I like this idea. cool

[hashimoto, 2012-04-24 06:57:06]

http://codereview.chromium.org/10134027/diff/8004/chrome/browser/chromeos/cro...
File chrome/browser/chromeos/cros/cros_network_functions_unittest.cc (right):

http://codereview.chromium.org/10134027/diff/8004/chrome/browser/chromeos/cro...
chrome/browser/chromeos/cros/cros_network_functions_unittest.cc:370:
MonitorDataPlanCallback callback = &OnDataPlansUpdate;
On 2012/04/24 06:05:55, satorux1 wrote:
> maybe safer to add the class name ?
> 
> &CrosNetworkFunctionsLibcrosTest::OnDataPlansUpdate

Done.