Serve mojo WebUI resources from the same origin as the WebUI itself.

Serve mojo WebUI resources from the same origin as the WebUI itself.

Currently, mojo resources are served from chrome://mojo and each mojo
WebUI controller replaces the WebUIDataSource for that origin with one
containing the resource it needs on construction. If a WebUI page's
controller (or an equivalent one) was not the last one to be
constructed, any requests for mojo bindings for its mojo interfaces
will fail. This change fixes this problem by serving the mojo resources
from the same origin as the WebUI itself.

BUG=557540
Committed: https://crrev.com/7e10653acb8254226f9b3ca5e0233c298c2ffbc3
Cr-Commit-Position: refs/heads/master@{#361275}

Committed: https://crrev.com/afd813ed4b6ce7fbaef2ed9dc39802bcd5587c79
Cr-Commit-Position: refs/heads/master@{#362078}

[Sam McNally, 2015-11-19 05:41:48]

sammc@chromium.org changed reviewers:
+ sky@chromium.org

[sky, 2015-11-19 17:43:37]

sky@chromium.org changed reviewers:
+ jam@chromium.org

[sky, 2015-11-19 17:43:38]

+jam for content (I'm not an owner there)

You should also get a security review too. Maybe tsepez.

https://codereview.chromium.org/1457623004/diff/1/content/renderer/web_ui_moj...
File content/renderer/web_ui_mojo_context_state.cc (right):

https://codereview.chromium.org/1457623004/diff/1/content/renderer/web_ui_moj...
content/renderer/web_ui_mojo_context_state.cc:138: if (module_prefix_.empty())
Can module_prefix_ be set correctly in the constructor?

[jam, 2015-11-20 01:22:13]

I'm not familiar with the mojo-webui code. Scott: can you review since I believe
you wrote this, and I can rs?

[sky, 2015-11-20 16:36:47]

On 2015/11/20 01:22:13, jam wrote:
> I'm not familiar with the mojo-webui code. Scott: can you review since I
believe
> you wrote this, and I can rs?

I did. I'm not ready to approve yet based on comment 2, but it's a trivial
change.

[Sam McNally, 2015-11-22 23:43:30]

sammc@chromium.org changed reviewers:
+ tsepez@chromium.org

[Sam McNally, 2015-11-22 23:43:30]

+tsepez for security

https://codereview.chromium.org/1457623004/diff/1/content/renderer/web_ui_moj...
File content/renderer/web_ui_mojo_context_state.cc (right):

https://codereview.chromium.org/1457623004/diff/1/content/renderer/web_ui_moj...
content/renderer/web_ui_mojo_context_state.cc:138: if (module_prefix_.empty())
On 2015/11/19 17:43:38, sky wrote:
> Can module_prefix_ be set correctly in the constructor?

Done.

[sky, 2015-11-23 16:11:56]

LGTM

[Tom Sepez, 2015-11-23 17:07:10]

LGTM.  I'm a little surprised that you registered the resource under a .mojom
extension, rather than .mojom.js or something, since it's generated JS, and not
the mojom IDL you're serving, right?

[Sam McNally, 2015-11-23 23:47:14]

On 2015/11/23 17:07:10, Tom Sepez wrote:
> LGTM.  I'm a little surprised that you registered the resource under a .mojom
> extension, rather than .mojom.js or something, since it's generated JS, and
not
> the mojom IDL you're serving, right?

The mojo JS modules are named without the .js extension. The module loader for
WebUI just sends a request for that path.

[jam, 2015-11-24 00:05:44]

rubberstamp lgtm

[Sam McNally, 2015-11-24 06:24:15]

The CQ bit was checked by sammc@chromium.org

[commit-bot: I haz the power, 2015-11-24 06:25:36]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1457623004/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1457623004/20001

[commit-bot: I haz the power, 2015-11-24 07:49:24]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-11-24 07:50:04]

Patchset 2 (id:??) landed as
https://crrev.com/7e10653acb8254226f9b3ca5e0233c298c2ffbc3
Cr-Commit-Position: refs/heads/master@{#361275}

[magjed_chromium, 2015-11-24 08:55:50]

magjed@chromium.org changed reviewers:
+ magjed@chromium.org

[magjed_chromium, 2015-11-24 08:55:50]

https://codereview.chromium.org/1457623004/diff/20001/content/browser/webui/w...
File content/browser/webui/web_ui_mojo_browsertest.cc (right):

https://codereview.chromium.org/1457623004/diff/20001/content/browser/webui/w...
content/browser/webui/web_ui_mojo_browsertest.cc:71:
CHECK(base::ReadFileToString(path, &contents, std::string::npos)) << id;
After your CL, we are hitting this CHECK on the Cast Linux bot,
content_browsertests, with the following failing tests:
WebUIMojoTest.EndToEndPing and WebUIMojoTest.ConnectToApplication.
https://build.chromium.org/p/chromium.linux/builders/Cast%20Linux/builds/1315...

[magjed_chromium, 2015-11-24 09:01:03]

A revert of this CL (patchset #2 id:20001) has been created in
https://codereview.chromium.org/1467133004/ by magjed@chromium.org.

The reason for reverting is: Causing test failures in the Cast Linux bot,
content_browsertests, tests WebUIMojoTest.EndToEndPing and
WebUIMojoTest.ConnectToApplication.
https://build.chromium.org/p/chromium.linux/builders/Cast%20Linux/builds/1315...
Stack trace:
[ RUN      ] WebUIMojoTest.ConnectToApplication
[8517:8517:1124/000407:287044551:FATAL:web_ui_mojo_browsertest.cc(71)] Check
failed: base::ReadFileToString(path, &contents, std::string::npos).
web_ui_mojo_shell_test.html
#0 0x000000b4fa0e base::debug::StackTrace::StackTrace()
#1 0x000000b5bbeb logging::LogMessage::~LogMessage()
#2 0x0000006d96a7 content::(anonymous namespace)::GetResource()
#3 0x000002f21749 content::WebUIDataSourceImpl::StartDataRequest()
#4 0x000002f1fb9f content::URLDataManagerBackend::CallStartRequest()
#5 0x000002f20588
_ZN4base8internal12InvokeHelperILb0EvNS0_15RunnableAdapterIPFv13scoped_refptrIN7content17URLDataSourceImplEERKSsiiiEEENS0_8TypeListIJPS5_S8_RKiSF_SF_EEEE8MakeItSoESB_SD_S8_SF_SF_SF_
#6 0x000000ba8a3d base::debug::TaskAnnotator::RunTask()
#7 0x000000b6091b base::MessageLoop::RunTask()
#8 0x000000b60c29 base::MessageLoop::DeferOrRunPendingTask()
#9 0x000000b60dfd base::MessageLoop::DoWork()
#10 0x000000b48e89 base::MessagePumpLibevent::Run()
#11 0x000000b6e600 base::RunLoop::Run()
#12 0x0000009b669c content::RunThisRunLoop()
#13 0x00000095d853 content::NavigateToURL()
#14 0x0000006d9f68 content::(anonymous
namespace)::WebUIMojoTest_ConnectToApplication_Test::RunTestOnMainThread()
#15 0x00000095d2c5 content::ContentBrowserTest::RunTestOnMainThreadLoop()
#16 0x0000009a7a9a content::BrowserTestBase::ProxyRunTestOnMainThreadLoop()
#17 0x00000099f86e content::ShellBrowserMainParts::PreMainMessageLoopRun()
#18 0x000002d9b015 content::BrowserMainLoop::PreMainMessageLoopRun()
#19 0x000002ee8397 content::StartupTaskRunner::RunAllTasksNow()
#20 0x000002d991ea content::BrowserMainLoop::CreateStartupTasks()
#21 0x000002d9dc96 content::BrowserMainRunnerImpl::Initialize()
#22 0x00000097e635 ShellBrowserMain()
#23 0x000000978af2 content::ShellMainDelegate::RunProcess()
#24 0x000002d73ea2 content::RunNamedProcessTypeMain()
#25 0x000002d74731 content::ContentMainRunnerImpl::Run()
#26 0x000002d73610 content::ContentMain()
#27 0x0000009a75ea content::BrowserTestBase::SetUp()
#28 0x00000095d1b3 content::ContentBrowserTest::SetUp()
#29 0x000000ac41b8 testing::Test::Run()
#30 0x000000ac517a testing::TestInfo::Run()
#31 0x000000ac5613 testing::TestCase::Run()
#32 0x000000acc9d9 testing::internal::UnitTestImpl::RunAllTests()
#33 0x000000acc67e testing::UnitTest::Run()
#34 0x0000009c9dc1 base::TestSuite::Run()
#35 0x0000009603d6 content::ContentTestLauncherDelegate::RunTestSuite()
#36 0x0000009b2806 content::LaunchTests()
#37 0x000000960383 main
#38 0x7f964b45cec5 __libc_start_main
#39 0x000000459519 <unknown>.

[Sam McNally, 2015-11-25 00:27:20]

Description was changed from

==========
Serve mojo WebUI resources from the same origin as the WebUI itself.

Currently, mojo resources are served from chrome://mojo and each mojo
WebUI controller replaces the WebUIDataSource for that origin with one
containing the resource it needs on construction. If a WebUI page's
controller (or an equivalent one) was not the last one to be
constructed, any requests for mojo bindings for its mojo interfaces
will fail. This change fixes this problem by serving the mojo resources
from the same origin as the WebUI itself.

BUG=557540

Committed: https://crrev.com/7e10653acb8254226f9b3ca5e0233c298c2ffbc3
Cr-Commit-Position: refs/heads/master@{#361275}
==========

to

==========
Serve mojo WebUI resources from the same origin as the WebUI itself.

Currently, mojo resources are served from chrome://mojo and each mojo
WebUI controller replaces the WebUIDataSource for that origin with one
containing the resource it needs on construction. If a WebUI page's
controller (or an equivalent one) was not the last one to be
constructed, any requests for mojo bindings for its mojo interfaces
will fail. This change fixes this problem by serving the mojo resources
from the same origin as the WebUI itself.

BUG=557540

Committed: https://crrev.com/7e10653acb8254226f9b3ca5e0233c298c2ffbc3
Cr-Commit-Position: refs/heads/master@{#361275}
==========

[Sam McNally, 2015-11-26 03:34:00]

Patchset #4 (id:60001) has been deleted

[Sam McNally, 2015-11-30 00:17:06]

Patchset #9 (id:180001) has been deleted

[Sam McNally, 2015-11-30 01:42:13]

The CQ bit was checked by sammc@chromium.org

[Sam McNally, 2015-11-30 01:42:13]

The patchset sent to the CQ was uploaded after l-g-t-m from tsepez@chromium.org,
jam@chromium.org, sky@chromium.org
Link to the patchset: https://codereview.chromium.org/1457623004/#ps200001
(title: "rebase")

[commit-bot: I haz the power, 2015-11-30 01:42:47]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1457623004/200001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1457623004/200001

[commit-bot: I haz the power, 2015-11-30 01:46:57]

Description was changed from

==========
Serve mojo WebUI resources from the same origin as the WebUI itself.

Currently, mojo resources are served from chrome://mojo and each mojo
WebUI controller replaces the WebUIDataSource for that origin with one
containing the resource it needs on construction. If a WebUI page's
controller (or an equivalent one) was not the last one to be
constructed, any requests for mojo bindings for its mojo interfaces
will fail. This change fixes this problem by serving the mojo resources
from the same origin as the WebUI itself.

BUG=557540

Committed: https://crrev.com/7e10653acb8254226f9b3ca5e0233c298c2ffbc3
Cr-Commit-Position: refs/heads/master@{#361275}
==========

to

==========
Serve mojo WebUI resources from the same origin as the WebUI itself.

Currently, mojo resources are served from chrome://mojo and each mojo
WebUI controller replaces the WebUIDataSource for that origin with one
containing the resource it needs on construction. If a WebUI page's
controller (or an equivalent one) was not the last one to be
constructed, any requests for mojo bindings for its mojo interfaces
will fail. This change fixes this problem by serving the mojo resources
from the same origin as the WebUI itself.

BUG=557540

Committed: https://crrev.com/7e10653acb8254226f9b3ca5e0233c298c2ffbc3
Cr-Commit-Position: refs/heads/master@{#361275}
==========

[commit-bot: I haz the power, 2015-11-30 01:46:58]

Committed patchset #9 (id:200001)

[commit-bot: I haz the power, 2015-11-30 01:47:47]

Description was changed from

==========
Serve mojo WebUI resources from the same origin as the WebUI itself.

Currently, mojo resources are served from chrome://mojo and each mojo
WebUI controller replaces the WebUIDataSource for that origin with one
containing the resource it needs on construction. If a WebUI page's
controller (or an equivalent one) was not the last one to be
constructed, any requests for mojo bindings for its mojo interfaces
will fail. This change fixes this problem by serving the mojo resources
from the same origin as the WebUI itself.

BUG=557540

Committed: https://crrev.com/7e10653acb8254226f9b3ca5e0233c298c2ffbc3
Cr-Commit-Position: refs/heads/master@{#361275}
==========

to

==========
Serve mojo WebUI resources from the same origin as the WebUI itself.

Currently, mojo resources are served from chrome://mojo and each mojo
WebUI controller replaces the WebUIDataSource for that origin with one
containing the resource it needs on construction. If a WebUI page's
controller (or an equivalent one) was not the last one to be
constructed, any requests for mojo bindings for its mojo interfaces
will fail. This change fixes this problem by serving the mojo resources
from the same origin as the WebUI itself.

BUG=557540

Committed: https://crrev.com/7e10653acb8254226f9b3ca5e0233c298c2ffbc3
Cr-Commit-Position: refs/heads/master@{#361275}

Committed: https://crrev.com/afd813ed4b6ce7fbaef2ed9dc39802bcd5587c79
Cr-Commit-Position: refs/heads/master@{#362078}
==========

[commit-bot: I haz the power, 2015-11-30 01:47:48]

Patchset 9 (id:??) landed as
https://crrev.com/afd813ed4b6ce7fbaef2ed9dc39802bcd5587c79
Cr-Commit-Position: refs/heads/master@{#362078}