Deal correctly with BrowsingData API calls during startup.

Deal correctly with BrowsingData API calls during startup.

We're apparently running into situations where an extension is attempting to
clear browsing data before either the browser or the profile is correctly set
up. This CL takes the simplest approach: if we can't grab the browser or the
profile, exit early (hopefully) without crashing.

BUG=139075
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=150583

[Mike West, 2012-08-08 11:31:29]

Hello Bernhard and Marja! Would you mind taking a look at this CL, which I hope
will mitigate the #3 crasher on M22? :)

Thanks!

-mike

[Bernhard Bauer, 2012-08-08 11:38:27]

We shouldn't use GetCurrentBrowser() to get the profile.
UIThreadExtensionFunction has a profile() method that returns the profile
without having to go through the browser (which requires a window).

[Mike West, 2012-08-08 11:44:12]

On 2012/08/08 11:38:27, Bernhard Bauer wrote:
> We shouldn't use GetCurrentBrowser() to get the profile.
> UIThreadExtensionFunction has a profile() method that returns the profile
> without having to go through the browser (which requires a window).

Ok. That's much simpler. :)

[Bernhard Bauer, 2012-08-08 11:46:51]

http://codereview.chromium.org/10855049/diff/1003/chrome/browser/extensions/a...
File chrome/browser/extensions/api/browsing_data/browsing_data_api.cc (right):

http://codereview.chromium.org/10855049/diff/1003/chrome/browser/extensions/a...
chrome/browser/extensions/api/browsing_data/browsing_data_api.cc:147: if
(removal_mask_ & BrowsingDataRemover::REMOVE_PLUGIN_DATA && profile()) {
I think we do have a profile at this point. I would prefer not to silently
ignore it if we don't.

[Mike West, 2012-08-08 12:17:31]

On 2012/08/08 11:46:51, Bernhard Bauer wrote:
>
http://codereview.chromium.org/10855049/diff/1003/chrome/browser/extensions/a...
> File chrome/browser/extensions/api/browsing_data/browsing_data_api.cc (right):
> 
>
http://codereview.chromium.org/10855049/diff/1003/chrome/browser/extensions/a...
> chrome/browser/extensions/api/browsing_data/browsing_data_api.cc:147: if
> (removal_mask_ & BrowsingDataRemover::REMOVE_PLUGIN_DATA && profile()) {
> I think we do have a profile at this point. I would prefer not to silently
> ignore it if we don't.

How would you prefer I deal with it? DCHECK?

[Bernhard Bauer, 2012-08-08 12:20:53]

On 2012/08/08 12:17:31, Mike West (chromium) wrote:
> On 2012/08/08 11:46:51, Bernhard Bauer wrote:
> >
>
http://codereview.chromium.org/10855049/diff/1003/chrome/browser/extensions/a...
> > File chrome/browser/extensions/api/browsing_data/browsing_data_api.cc
(right):
> > 
> >
>
http://codereview.chromium.org/10855049/diff/1003/chrome/browser/extensions/a...
> > chrome/browser/extensions/api/browsing_data/browsing_data_api.cc:147: if
> > (removal_mask_ & BrowsingDataRemover::REMOVE_PLUGIN_DATA && profile()) {
> > I think we do have a profile at this point. I would prefer not to silently
> > ignore it if we don't.
> 
> How would you prefer I deal with it? DCHECK?

Don't deal with it at all. If there is no profile, we will eventually crash
anyway, so we gain nothing from a DCHECK.

[Mike West, 2012-08-08 12:28:29]

On 2012/08/08 12:20:53, Bernhard Bauer wrote:
> On 2012/08/08 12:17:31, Mike West (chromium) wrote:
> > On 2012/08/08 11:46:51, Bernhard Bauer wrote:
> > >
> >
>
http://codereview.chromium.org/10855049/diff/1003/chrome/browser/extensions/a...
> > > File chrome/browser/extensions/api/browsing_data/browsing_data_api.cc
> (right):
> > > 
> > >
> >
>
http://codereview.chromium.org/10855049/diff/1003/chrome/browser/extensions/a...
> > > chrome/browser/extensions/api/browsing_data/browsing_data_api.cc:147: if
> > > (removal_mask_ & BrowsingDataRemover::REMOVE_PLUGIN_DATA && profile()) {
> > > I think we do have a profile at this point. I would prefer not to silently
> > > ignore it if we don't.
> > 
> > How would you prefer I deal with it? DCHECK?
> 
> Don't deal with it at all. If there is no profile, we will eventually crash
> anyway, so we gain nothing from a DCHECK.

As discussed briefly over Talk, we gain explicitness. And, if the browser bit
wasn't the issue (though I think you're right: that was it), we gain a very
clear crash, as opposed to what we have at the moment. :)

[Bernhard Bauer, 2012-08-08 12:54:11]

LGTM

[commit-bot: I haz the power, 2012-08-08 12:54:53]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/mkwst@chromium.org/10855049/5002

[commit-bot: I haz the power, 2012-08-08 14:34:55]

Try job failure for 10855049-5002 (retry) on mac_rel for step "browser_tests".
It's a second try, previously, step "browser_tests" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=mac_rel&nu...

[commit-bot: I haz the power, 2012-08-08 14:37:19]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/mkwst@chromium.org/10855049/5002

[commit-bot: I haz the power, 2012-08-08 17:46:06]

Change committed as 150583