New security-related pyauto tests for Chrome on ChromeOS that verify extension permissions.

New security-related pyauto tests for Chrome on ChromeOS that verify extension permissions.

BUG=chromium-os:15994
TEST=None

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=88351

[dennis_jeffrey, 2011-06-03 00:04:29]

I still need to verify that this works on an official Chrome-on-ChromeOS build
too (once Chrome revision 87673 gets pulled into an official ChromeOS build). 
But wanted to send this out in the meantime to start the review process.

[Nirnimesh, 2011-06-03 18:30:28]

http://codereview.chromium.org/6969105/diff/1/chrome/test/functional/chromeos...
File chrome/test/functional/chromeos_security.py (right):

http://codereview.chromium.org/6969105/diff/1/chrome/test/functional/chromeos...
chrome/test/functional/chromeos_security.py:17: COMPONENT_EXTENSION_BASELINE = [
Are these the same on marios, alex and zgb?

http://codereview.chromium.org/6969105/diff/1/chrome/test/functional/chromeos...
chrome/test/functional/chromeos_security.py:146:
self._VerifyExtensionPermissions(self.COMPONENT_EXTENSION_BASELINE)
Should there be a count match somewhere?

http://codereview.chromium.org/6969105/diff/1/chrome/test/functional/chromeos...
chrome/test/functional/chromeos_security.py:152: os.path.join('/', 'opt',
'google', 'chrome', 'extensions'))
This is one of the places where you don't want to use join()
Use the string instead: '/opt/goog/e/chrome/extensions'

http://codereview.chromium.org/6969105/diff/1/chrome/test/functional/chromeos...
chrome/test/functional/chromeos_security.py:162: msg='Extension install
failed.')
add: |crx_file| in the mesg

[dennis_jeffrey, 2011-06-03 22:04:19]

http://codereview.chromium.org/6969105/diff/1/chrome/test/functional/chromeos...
File chrome/test/functional/chromeos_security.py (right):

http://codereview.chromium.org/6969105/diff/1/chrome/test/functional/chromeos...
chrome/test/functional/chromeos_security.py:17: COMPONENT_EXTENSION_BASELINE = [
On 2011/06/03 18:30:28, Nirnimesh wrote:
> Are these the same on marios, alex and zgb?

I think so.  I'm getting this information from
ProfileImpl::RegisterComponentExtensions, which contains the logic for whether
different component extensions are registered or not (it checks whether we're on
ChromeOS, whether it's an official build, etc).  I didn't see anything that
would suggest differences based on ChromeOS board type.

http://codereview.chromium.org/6969105/diff/1/chrome/test/functional/chromeos...
chrome/test/functional/chromeos_security.py:146:
self._VerifyExtensionPermissions(self.COMPONENT_EXTENSION_BASELINE)
On 2011/06/03 18:30:28, Nirnimesh wrote:
> Should there be a count match somewhere?

Done.  This requires that I know which installed extensions are component
extensions.  I had to slightly modify the GetExtensionsInfo automation hook to
do it.

http://codereview.chromium.org/6969105/diff/1/chrome/test/functional/chromeos...
chrome/test/functional/chromeos_security.py:152: os.path.join('/', 'opt',
'google', 'chrome', 'extensions'))
On 2011/06/03 18:30:28, Nirnimesh wrote:
> This is one of the places where you don't want to use join()
> Use the string instead: '/opt/goog/e/chrome/extensions'

Done.  What is the reason for this?

http://codereview.chromium.org/6969105/diff/1/chrome/test/functional/chromeos...
chrome/test/functional/chromeos_security.py:162: msg='Extension install
failed.')
On 2011/06/03 18:30:28, Nirnimesh wrote:
> add: |crx_file| in the mesg

Done.

[jimhebert, 2011-06-03 23:34:43]

http://codereview.chromium.org/6969105/diff/5001/chrome/test/functional/chrom...
File chrome/test/functional/chromeos_security.py (right):

http://codereview.chromium.org/6969105/diff/5001/chrome/test/functional/chrom...
chrome/test/functional/chromeos_security.py:19: COMPONENT_EXTENSION_BASELINE = [
Is it typical pyauto style to just inline these expectations? If so OK as long
as it matches local pyauto style.

(In the autotests I've been keeping these serialized as seperate files...)

http://codereview.chromium.org/6969105/diff/5001/chrome/test/functional/chrom...
chrome/test/functional/chromeos_security.py:46: BUNDLED_CRX_BASELINE = [
The autotest version of this test is kept in autotest-private because it will on
occasion end up containing name/details of unannounced Chrome OS features which
themselves are built out of private/internal bits. I don't know what the
equivalent 'private' mechanism is for pyauto but I think the concern carries
over. How do we address that in pyauto?

[Nirnimesh, 2011-06-04 22:53:08]

http://codereview.chromium.org/6969105/diff/5001/chrome/test/functional/chrom...
File chrome/test/functional/chromeos_security.py (right):

http://codereview.chromium.org/6969105/diff/5001/chrome/test/functional/chrom...
chrome/test/functional/chromeos_security.py:46: BUNDLED_CRX_BASELINE = [
On 2011/06/03 23:34:44, jimhebert wrote:
> The autotest version of this test is kept in autotest-private because it will
on
> occasion end up containing name/details of unannounced Chrome OS features
which
> themselves are built out of private/internal bits. I don't know what the
> equivalent 'private' mechanism is for pyauto but I think the concern carries
> over. How do we address that in pyauto?

In this case, the baseline needs to be read out from a file in
chrome/test/data/pyauto_private/chromeos (this is private)

I'd prefer to have all data together in a single data file, in the private repo.

[dennis_jeffrey, 2011-06-07 19:02:20]

Thanks for the comments.  I'll create a separate CL to check in the baseline
information into the pyauto private data directory (will send out the review
soon).

I will not commit the current CL until the other CL gets committed.

http://codereview.chromium.org/6969105/diff/5001/chrome/test/functional/chrom...
File chrome/test/functional/chromeos_security.py (right):

http://codereview.chromium.org/6969105/diff/5001/chrome/test/functional/chrom...
chrome/test/functional/chromeos_security.py:19: COMPONENT_EXTENSION_BASELINE = [
On 2011/06/03 23:34:44, jimhebert wrote:
> Is it typical pyauto style to just inline these expectations? If so OK as long
> as it matches local pyauto style.
> 
> (In the autotests I've been keeping these serialized as seperate files...)

The baseline information has been moved to a separate private data file.

http://codereview.chromium.org/6969105/diff/5001/chrome/test/functional/chrom...
chrome/test/functional/chromeos_security.py:46: BUNDLED_CRX_BASELINE = [
On 2011/06/03 23:34:44, jimhebert wrote:
> The autotest version of this test is kept in autotest-private because it will
on
> occasion end up containing name/details of unannounced Chrome OS features
which
> themselves are built out of private/internal bits. I don't know what the
> equivalent 'private' mechanism is for pyauto but I think the concern carries
> over. How do we address that in pyauto?

Done.  The baseline info has been moved to a separate private data file.

http://codereview.chromium.org/6969105/diff/5001/chrome/test/functional/chrom...
chrome/test/functional/chromeos_security.py:46: BUNDLED_CRX_BASELINE = [
On 2011/06/04 22:53:08, Nirnimesh wrote:
> On 2011/06/03 23:34:44, jimhebert wrote:
> > The autotest version of this test is kept in autotest-private because it
will
> on
> > occasion end up containing name/details of unannounced Chrome OS features
> which
> > themselves are built out of private/internal bits. I don't know what the
> > equivalent 'private' mechanism is for pyauto but I think the concern carries
> > over. How do we address that in pyauto?
> 
> In this case, the baseline needs to be read out from a file in
> chrome/test/data/pyauto_private/chromeos (this is private)
> 
> I'd prefer to have all data together in a single data file, in the private
repo.

Done.  I created a file in a "security" subfolder of the pyauto private data
directory.

[Nirnimesh, 2011-06-07 19:23:51]

http://codereview.chromium.org/6969105/diff/10001/chrome/test/functional/chro...
File chrome/test/functional/chromeos_security.py (right):

http://codereview.chromium.org/6969105/diff/10001/chrome/test/functional/chro...
chrome/test/functional/chromeos_security.py:25: self.BUNDLED_CRX_DIRECTORY =
baseline_info['BUNDLED_CRX_DIRECTORY']
this is not a constant anymore. Why CAPS?

[dennis_jeffrey, 2011-06-07 20:45:45]

http://codereview.chromium.org/6969105/diff/10001/chrome/test/functional/chro...
File chrome/test/functional/chromeos_security.py (right):

http://codereview.chromium.org/6969105/diff/10001/chrome/test/functional/chro...
chrome/test/functional/chromeos_security.py:25: self.BUNDLED_CRX_DIRECTORY =
baseline_info['BUNDLED_CRX_DIRECTORY']
On 2011/06/07 19:23:51, Nirnimesh wrote:
> this is not a constant anymore. Why CAPS?

Done.

[Nirnimesh, 2011-06-07 20:49:14]

LGTM