Implement prediction_options.

Implement prediction_options.  This provides two functions in the chrome_browser_net namespace that can be used on either UI or OI thread to determine whether predictive network actions (which will include DNS, preconnect, prefetching, rendering) are allowed given user preference and current network state.

BUG=334602
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=280525

[jkarlin, 2014-06-26 16:28:53]

Drive by comments

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
File chrome/browser/net/prediction_helper.cc (right):

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
chrome/browser/net/prediction_helper.cc:31: return false;
DCHECK(false) with:
NOTREACHED() << "Unknown kAllowNetworkPrediction value.";

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
File chrome/browser/net/prediction_helper.h (right):

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
chrome/browser/net/prediction_helper.h:17: }
  // namespace user_prefs

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
chrome/browser/net/prediction_helper.h:21: }
  // namespace content

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
File chrome/browser/net/predictor.h (right):

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
chrome/browser/net/predictor.h:86: // NETWORK_PREDICTION_UNSET means that the
old setting(s) should be observed.
Unclear what "the old settings" refers to here.

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/profiles/...
File chrome/browser/profiles/profile_io_data.cc (right):

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/profiles/...
chrome/browser/profiles/profile_io_data.cc:441: pref_service);
formatting.  Please run 'git cl format'

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/profiles/...
File chrome/browser/profiles/profile_io_data.h (right):

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/profiles/...
chrome/browser/profiles/profile_io_data.h:8: #include <map>
why add this?

[Bence, 2014-06-26 20:02:06]

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
File chrome/browser/net/prediction_helper.cc (right):

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
chrome/browser/net/prediction_helper.cc:31: return false;
On 2014/06/26 16:28:52, jkarlin wrote:
> DCHECK(false) with:
> NOTREACHED() << "Unknown kAllowNetworkPrediction value.";

Done.

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
File chrome/browser/net/prediction_helper.h (right):

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
chrome/browser/net/prediction_helper.h:17: }
On 2014/06/26 16:28:52, jkarlin wrote:
>   // namespace user_prefs

Done.

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
chrome/browser/net/prediction_helper.h:21: }
On 2014/06/26 16:28:52, jkarlin wrote:
>   // namespace content

Done.

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
File chrome/browser/net/predictor.h (right):

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/net/predi...
chrome/browser/net/predictor.h:86: // NETWORK_PREDICTION_UNSET means that the
old setting(s) should be observed.
On 2014/06/26 16:28:52, jkarlin wrote:
> Unclear what "the old settings" refers to here.

Done.

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/profiles/...
File chrome/browser/profiles/profile_io_data.cc (right):

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/profiles/...
chrome/browser/profiles/profile_io_data.cc:441: pref_service);
On 2014/06/26 16:28:52, jkarlin wrote:
> formatting.  Please run 'git cl format'

Done.

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/profiles/...
File chrome/browser/profiles/profile_io_data.h (right):

https://codereview.chromium.org/357723003/diff/60001/chrome/browser/profiles/...
chrome/browser/profiles/profile_io_data.h:8: #include <map>
On 2014/06/26 16:28:52, jkarlin wrote:
> why add this?

git cl lint was complaining.  Should I ignore it?

[Bence, 2014-06-26 20:05:30]

Hi,

mmenke: could you please review changes in browser/net and browser/profiles?

battre: could you please review browser/prefs?

Thank you,

bnc

[mmenke, 2014-06-26 20:41:02]

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
File chrome/browser/net/prediction_helper.cc (right):

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:23:
net::NetworkChangeNotifier::GetConnectionType());
nit:  Suggest just using a 4-space indent here.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:29: NOTREACHED() << "Unknown
kAllowNetworkPrediction value.";
Include base/logging.h for NOTREACHED and DCHECK.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:34: }  // anonymous namespace
nit:  --anonymous

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:46: // To be executed on the IO thread
only.
Function level comments should generally be with their declaration, rather than
their definition.  In this case, the comment is redundant with what you have in
the header file, so should just get rid of it.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:49: DCHECK(resource_context);
nit:  Generally, there's a blank line between DCHECKs and other code, if the
function is long enough.  Makes it easier to review.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:51: if (io_data == NULL)
Can this ever be NULL?  Looks like other consumers of this function don't do
this check.  In general, the Chrome philosophy is not to handle things that can
never happen, unless they could result in a security bug.  In this case, we'd
get crash dumps.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:58: // To be executed on the UI thread
only.
nit:  Remove comment.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
File chrome/browser/net/prediction_helper.h (right):

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.h:8: // this is implemented on a polling
basis (as opposed to observers).
nit:  Maybe "... this polls the current network state as needed, instead of
caching it."

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.h:10: // See https://crbug.com/334602 for
more details.
Move comments to before function definitions (Maybe with a blank line below it,
to indicate it applies to multiple code blocks).

Comments just below the license header often get overlooked.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.h:10: // See https://crbug.com/334602 for
more details.
I'm not sure the bug is interesting enough to link to here, since it talks a lot
about how things are before you're done.

Could mention when uses the setting, once things start using it, though that may
have a tendency to get outdated, and it's easy enough to grep for callsites,
though I'm not opposed to adding it, if you think it adds value.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.h:13: #define
CHROME_BROWSER_NET_PREDICTION_HELPER_H_
Unless you have plans to add other stuff to this file, suggest calling it
prediction_config.h.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.h:25: namespace chrome_browser_net {
Using the chrome_browser_net namespace has fallen out of favor.  Suggest not
using it in new files.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
File chrome/browser/net/predictor.h (right):

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/predictor.h:84: // The same enum must be used by the
platform-dependent components.
Should mention that order matters here, since they're stored in the prefs file.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/predictor.h:93: };
Suggest moving these into the new file, since they're going to be used outside
the predictor as well.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/profiles...
File chrome/browser/profiles/profile_io_data.h (right):

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/profiles...
chrome/browser/profiles/profile_io_data.h:194: IntegerPrefMember*
allow_network_prediction() const {
Using an integer for something named "allow_network_prediction", which sounds
like a bool, seems wrong to me.  You called the enum "NetworkPredictionOptions"
in the other CL, which is better.  Not perfect, but I have no better ideas.

[battre, 2014-06-27 06:34:57]

chrome/browser/prefs/browser_prefs.cc LGTM.

[Bence, 2014-06-27 15:11:51]

mmenke: We should brainstorm about four things:

1. What is the best filename for browser/net/prediction_helper.h?
2. What is the best namespace for its two functions?
3. What is the best place for enum NetworkPredictionOptions?  (See second
comment at https://codereview.chromium.org/336543003#msg2).
4. What is the best name for this enum and the corresponding preference?

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
File chrome/browser/net/prediction_helper.cc (right):

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:23:
net::NetworkChangeNotifier::GetConnectionType());
On 2014/06/26 20:41:01, mmenke wrote:
> nit:  Suggest just using a 4-space indent here.

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:29: NOTREACHED() << "Unknown
kAllowNetworkPrediction value.";
On 2014/06/26 20:41:01, mmenke wrote:
> Include base/logging.h for NOTREACHED and DCHECK.

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:34: }  // anonymous namespace
On 2014/06/26 20:41:01, mmenke wrote:
> nit:  --anonymous

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:46: // To be executed on the IO thread
only.
On 2014/06/26 20:41:01, mmenke wrote:
> Function level comments should generally be with their declaration, rather
than
> their definition.  In this case, the comment is redundant with what you have
in
> the header file, so should just get rid of it.

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:49: DCHECK(resource_context);
On 2014/06/26 20:41:01, mmenke wrote:
> nit:  Generally, there's a blank line between DCHECKs and other code, if the
> function is long enough.  Makes it easier to review.

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:51: if (io_data == NULL)
On 2014/06/26 20:41:01, mmenke wrote:
> Can this ever be NULL?  Looks like other consumers of this function don't do
> this check.  In general, the Chrome philosophy is not to handle things that
can
> never happen, unless they could result in a security bug.  In this case, we'd
> get crash dumps.

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.cc:58: // To be executed on the UI thread
only.
On 2014/06/26 20:41:01, mmenke wrote:
> nit:  Remove comment.

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
File chrome/browser/net/prediction_helper.h (right):

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.h:8: // this is implemented on a polling
basis (as opposed to observers).
On 2014/06/26 20:41:01, mmenke wrote:
> nit:  Maybe "... this polls the current network state as needed, instead of
> caching it."

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.h:10: // See https://crbug.com/334602 for
more details.
On 2014/06/26 20:41:01, mmenke wrote:
> I'm not sure the bug is interesting enough to link to here, since it talks a
lot
> about how things are before you're done.
> 
> Could mention when uses the setting, once things start using it, though that
may
> have a tendency to get outdated, and it's easy enough to grep for callsites,
> though I'm not opposed to adding it, if you think it adds value.

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.h:10: // See https://crbug.com/334602 for
more details.
On 2014/06/26 20:41:01, mmenke wrote:
> I'm not sure the bug is interesting enough to link to here, since it talks a
lot
> about how things are before you're done.
> 
> Could mention when uses the setting, once things start using it, though that
may
> have a tendency to get outdated, and it's easy enough to grep for callsites,
> though I'm not opposed to adding it, if you think it adds value.

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.h:13: #define
CHROME_BROWSER_NET_PREDICTION_HELPER_H_
On 2014/06/26 20:41:01, mmenke wrote:
> Unless you have plans to add other stuff to this file, suggest calling it
> prediction_config.h.

I do not have such plans.  Let's discuss it.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.h:25: namespace chrome_browser_net {
On 2014/06/26 20:41:01, mmenke wrote:
> Using the chrome_browser_net namespace has fallen out of favor.  Suggest not
> using it in new files.

Sure.  What namespace do you recommend for these two particular functions?

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
File chrome/browser/net/predictor.h (right):

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/predictor.h:84: // The same enum must be used by the
platform-dependent components.
On 2014/06/26 20:41:02, mmenke wrote:
> Should mention that order matters here, since they're stored in the prefs
file.

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/predictor.h:93: };
On 2014/06/26 20:41:02, mmenke wrote:
> Suggest moving these into the new file, since they're going to be used outside
> the predictor as well.

Cf. https://codereview.chromium.org/336543003#msg2.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/profiles...
File chrome/browser/profiles/profile_io_data.h (right):

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/profiles...
chrome/browser/profiles/profile_io_data.h:194: IntegerPrefMember*
allow_network_prediction() const {
On 2014/06/26 20:41:02, mmenke wrote:
> Using an integer for something named "allow_network_prediction", which sounds
> like a bool, seems wrong to me.  You called the enum
"NetworkPredictionOptions"
> in the other CL, which is better.  Not perfect, but I have no better ideas.

Great catch.  Let's do some brainstorming before we rename this preference, it
would already have to be changed at a lot of places, including
https://codereview.chromium.org/336543003.

[Bence, 2014-06-27 16:11:52]

Results of discussion with mmenke:

1. What is the best filename for browser/net/prediction_helper.h?

browser/net/network_prediction_options.h

2. What is the best namespace for its two functions?

In default of a clear guideline, they will stay in chrome_browser_net.

3. What is the best place for enum NetworkPredictionOptions?  (See second
comment at https://codereview.chromium.org/336543003#msg2).

browser/net/network_prediction_options.h.

4. What is the best name for this enum and the corresponding preference?

enum NetworkPredictionOptions
const char kNetworkPredictionOptions[] = "net.network_prediction_options";

Calling every aspect of a single functionality by the same name, including the
files, yields in the least cognitive overhead.

[Bence, 2014-06-27 16:20:50]

Since browser/net/network_prediction_options.{cc,h} is redundant, I will use
browser/net/prediction_options.{cc,h} instead.

On 2014/06/27 16:11:52, bnc wrote:
> Results of discussion with mmenke:
> 
> 1. What is the best filename for browser/net/prediction_helper.h?
> 
> browser/net/network_prediction_options.h
> 
> 2. What is the best namespace for its two functions?
> 
> In default of a clear guideline, they will stay in chrome_browser_net.
> 
> 3. What is the best place for enum NetworkPredictionOptions?  (See second
> comment at https://codereview.chromium.org/336543003#msg2).
> 
> browser/net/network_prediction_options.h.
> 
> 4. What is the best name for this enum and the corresponding preference?
> 
> enum NetworkPredictionOptions
> const char kNetworkPredictionOptions[] = "net.network_prediction_options";
> 
> Calling every aspect of a single functionality by the same name, including the
> files, yields in the least cognitive overhead.

[mmenke, 2014-06-27 16:26:29]

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
File chrome/browser/net/prediction_helper.h (right):

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.h:27: void
RegisterPredictionHelperProfilePrefs(
OH, this should be renamed, too, using our new favorite string.  Other names
look fine.

[Bence, 2014-06-27 18:26:45]

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
File chrome/browser/net/prediction_helper.h (right):

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/prediction_helper.h:27: void
RegisterPredictionHelperProfilePrefs(
On 2014/06/27 16:26:29, mmenke wrote:
> OH, this should be renamed, too, using our new favorite string.  Other names
> look fine.

Well, this is where a unique namespace would be nice, because every other
function called by chrome/browser/prefs/browser_prefs.cc is named
RegisterProfilePrefs.

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
File chrome/browser/net/predictor.h (right):

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/net/pred...
chrome/browser/net/predictor.h:93: };
On 2014/06/26 20:41:02, mmenke wrote:
> Suggest moving these into the new file, since they're going to be used outside
> the predictor as well.

Done.

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/profiles...
File chrome/browser/profiles/profile_io_data.h (right):

https://codereview.chromium.org/357723003/diff/120001/chrome/browser/profiles...
chrome/browser/profiles/profile_io_data.h:194: IntegerPrefMember*
allow_network_prediction() const {
On 2014/06/27 15:11:51, bnc wrote:
> On 2014/06/26 20:41:02, mmenke wrote:
> > Using an integer for something named "allow_network_prediction", which
sounds
> > like a bool, seems wrong to me.  You called the enum
> "NetworkPredictionOptions"
> > in the other CL, which is better.  Not perfect, but I have no better ideas.
> 
> Great catch.  Let's do some brainstorming before we rename this preference, it
> would already have to be changed at a lot of places, including
> https://codereview.chromium.org/336543003.

Done.

[mmenke, 2014-06-27 18:52:06]

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/net/pred...
File chrome/browser/net/prediction_options.cc (right):

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/net/pred...
chrome/browser/net/prediction_options.cc:48:
DCHECK(content::BrowserThread::CurrentlyOn(content::BrowserThread::IO));
include browser_thread.h

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/net/pred...
File chrome/browser/net/prediction_options.h (right):

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/net/pred...
chrome/browser/net/prediction_options.h:10: }  // user_prefs
nit:  Close namespace comments usually aren't used for short forward declaration
sections.

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/net/pred...
chrome/browser/net/prediction_options.h:14: }  // namespace content
nit:  Close namespace comments usually aren't used for short forward declaration
sections.

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/net/pred...
chrome/browser/net/prediction_options.h:25: }  // namespace
Delete this block of code, it's not needed.

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/profiles...
File chrome/browser/profiles/profile_io_data.cc (right):

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/profiles...
chrome/browser/profiles/profile_io_data.cc:441: pref_service);
nit:  +2 indent

[Bence, 2014-06-27 19:10:24]

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/net/pred...
File chrome/browser/net/prediction_options.cc (right):

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/net/pred...
chrome/browser/net/prediction_options.cc:48:
DCHECK(content::BrowserThread::CurrentlyOn(content::BrowserThread::IO));
On 2014/06/27 18:52:05, mmenke wrote:
> include browser_thread.h

Done.

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/net/pred...
File chrome/browser/net/prediction_options.h (right):

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/net/pred...
chrome/browser/net/prediction_options.h:10: }  // user_prefs
On 2014/06/27 18:52:05, mmenke wrote:
> nit:  Close namespace comments usually aren't used for short forward
declaration
> sections.

Done.

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/net/pred...
chrome/browser/net/prediction_options.h:14: }  // namespace content
On 2014/06/27 18:52:05, mmenke wrote:
> nit:  Close namespace comments usually aren't used for short forward
declaration
> sections.

Done.

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/net/pred...
chrome/browser/net/prediction_options.h:25: }  // namespace
On 2014/06/27 18:52:05, mmenke wrote:
> Delete this block of code, it's not needed.

Comment moved to prediction_options.cc.

Done.

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/profiles...
File chrome/browser/profiles/profile_io_data.cc (right):

https://codereview.chromium.org/357723003/diff/160001/chrome/browser/profiles...
chrome/browser/profiles/profile_io_data.cc:441: pref_service);
On 2014/06/27 18:52:05, mmenke wrote:
> nit:  +2 indent

Done.

[mmenke, 2014-06-27 19:14:17]

LGTM!

[Bence, 2014-06-27 19:20:06]

The CQ bit was checked by bnc@chromium.org

[commit-bot: I haz the power, 2014-06-27 19:21:26]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/bnc@chromium.org/357723003/180001

[Bence, 2014-06-27 20:30:32]

The CQ bit was checked by bnc@chromium.org

[commit-bot: I haz the power, 2014-06-27 20:31:58]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/bnc@chromium.org/357723003/200001

[commit-bot: I haz the power, 2014-06-28 01:29:39]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  win_chromium_compile_dbg on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/win_chromium_compile_...)
  win_chromium_rel on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/win_chromium_rel/buil...)

[commit-bot: I haz the power, 2014-06-28 20:37:17]

Change committed as 280525