Make it so the windows store protocol is properly handled

Make it so the windows store protocol is properly handled

The core problem is that the key that controls the execution of a protocol
HKEY_CLASSES_ROOT\<protocol>\Shell\Open\Command

Has changed, we expect the Default value to be a valid command like foo.exe %1
but this is no longer the case. This value can be empty and other values will
control what happens, for instance DelegateExecute.

This CL also whitelists the protocol, alternatively we could hardcode a meaniful
string to present to the user but that presents localization issues. We take the
same approach as IE and just launch the windows store if we see this protocol.

BUG=159096
TEST=see bug
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=175678

[cpu_(ooo_6.6-7.5), 2013-01-08 01:47:05]

In this version of the change we whitelist the protocol, like we do for mailto,
news and snews. This is consistent with IE10.

An alternative is to present the security dialog, the problem there is that the
methods I know to find the name to present return the non-nonsensical name
"twinui.dll" as the target app for the scheme. This happens to firefox as well.

I could also invest time into finding the right name, but I feel comfortable
with the whitelist. Let me know how do you want to proceed.

[jschuh, 2013-01-08 04:06:00]

lgtm with one nit.

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
File chrome/browser/platform_util_win.cc (right):

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
chrome/browser/platform_util_win.cc:168: if
(!ValidateShellCommandForScheme(url.scheme()))
I feel like Nicolas should sign off on this one given the original bug.

[darin (slow to review), 2013-01-08 06:00:14]

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
File chrome/browser/platform_util_win.cc (right):

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
chrome/browser/platform_util_win.cc:129: return false;
nit: 2 space indent

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
chrome/browser/platform_util_win.cc:132: }  // namespace
nit: new line before close of namespace

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
chrome/browser/platform_util_win.cc:167: if (base::win::GetVersion() <
base::win::VERSION_WIN7) {
Why less than Win7 and not less than Win8?  Win7 also supports the
DelegateExecute method?

[cpu_(ooo_6.6-7.5), 2013-01-08 19:01:16]

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
File chrome/browser/platform_util_win.cc (right):

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
chrome/browser/platform_util_win.cc:129: return false;
On 2013/01/08 06:00:14, darin wrote:
> nit: 2 space indent

Done.

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
chrome/browser/platform_util_win.cc:132: }  // namespace
On 2013/01/08 06:00:14, darin wrote:
> nit: new line before close of namespace

Done.

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
chrome/browser/platform_util_win.cc:167: if (base::win::GetVersion() <
base::win::VERSION_WIN7) {
On 2013/01/08 06:00:14, darin wrote:
> Why less than Win7 and not less than Win8?  Win7 also supports the
> DelegateExecute method?

Yes.

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
chrome/browser/platform_util_win.cc:168: if
(!ValidateShellCommandForScheme(url.scheme()))
On 2013/01/08 04:06:00, Justin Schuh wrote:
> I feel like Nicolas should sign off on this one given the original bug.

Ok, I'll ping him.

[darin (slow to review), 2013-01-08 19:12:33]

LGTM

[jschuh, 2013-01-08 19:22:37]

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
File chrome/browser/platform_util_win.cc (right):

https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
chrome/browser/platform_util_win.cc:168: if
(!ValidateShellCommandForScheme(url.scheme()))
On 2013/01/08 19:01:16, cpu wrote:
> On 2013/01/08 04:06:00, Justin Schuh wrote:
> > I feel like Nicolas should sign off on this one given the original bug.
> 
> Ok, I'll ping him.

I'm pretty sure I typed that after you handed me a tequila. Please disregard.

[nsylvain, 2013-01-08 19:58:16]

On 2013/01/08 19:22:37, Justin Schuh wrote:
>
https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
> File chrome/browser/platform_util_win.cc (right):
> 
>
https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
> chrome/browser/platform_util_win.cc:168: if
> (!ValidateShellCommandForScheme(url.scheme()))
> On 2013/01/08 19:01:16, cpu wrote:
> > On 2013/01/08 04:06:00, Justin Schuh wrote:
> > > I feel like Nicolas should sign off on this one given the original bug.
> > 
> > Ok, I'll ping him.
> 
> I'm pretty sure I typed that after you handed me a tequila. Please disregard.

Nicolas thinks the code looks fine...  I assume you tried an empty string (with
not alternate ways of describing the command) on Win7 and it did not crash, just
failed silently?

[cpu_(ooo_6.6-7.5), 2013-01-08 20:16:53]

On 2013/01/08 19:58:16, nsylvain wrote:
> On 2013/01/08 19:22:37, Justin Schuh wrote:
> >
>
https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
> > File chrome/browser/platform_util_win.cc (right):
> > 
> >
>
https://codereview.chromium.org/11316038/diff/1/chrome/browser/platform_util_...
> > chrome/browser/platform_util_win.cc:168: if
> > (!ValidateShellCommandForScheme(url.scheme()))
> > On 2013/01/08 19:01:16, cpu wrote:
> > > On 2013/01/08 04:06:00, Justin Schuh wrote:
> > > > I feel like Nicolas should sign off on this one given the original bug.
> > > 
> > > Ok, I'll ping him.
> > 
> > I'm pretty sure I typed that after you handed me a tequila. Please
disregard.
> 
> Nicolas thinks the code looks fine...  I assume you tried an empty string
(with
> not alternate ways of describing the command) on Win7 and it did not crash,
just
> failed silently?

Yep, no crash.