Rearrange Flash component registration.

Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1

Committed: https://crrev.com/ae2a704ab712680c8b53a86e77eba78d0e591106
Committed: https://crrev.com/66a5763d27e17e175c46fcb1036f73c2c91e5f7f
Cr-Original-Commit-Position: refs/heads/master@{#405148}
Cr-Commit-Position: refs/heads/master@{#405597}

[waffles, 2016-07-07 18:44:16]

waffles@chromium.org changed reviewers:
+ kerrnel@chromium.org, wfh@chromium.org

[waffles, 2016-07-07 18:44:17]

Will, Greg: PTAL. Tested on Mac.

The main thing I want to accomplish here is to put RegisterWithChrome back into
the synchronous component updater flow, so that I can later hook the component
updater's update-complete notification to reload missing plugin placeholders (in
a race-free way).

[Will Harris, 2016-07-07 18:52:16]

On 2016/07/07 18:44:17, waffles wrote:
> Will, Greg: PTAL. Tested on Mac.
> 
> The main thing I want to accomplish here is to put RegisterWithChrome back
into
> the synchronous component updater flow, so that I can later hook the component
> updater's update-complete notification to reload missing plugin placeholders
(in
> a race-free way).

I'm a bit behind on the latest state of this code.

Am I right that we will no longer be bundling flash as a special "bundled"
distro with special code to handle this (and magic #defines like
FLAPPER_VERSION) but instead will always be just bundling the component?

[waffles, 2016-07-07 19:39:10]

On 2016/07/07 18:52:16, Will Harris wrote:
> On 2016/07/07 18:44:17, waffles wrote:
> > Will, Greg: PTAL. Tested on Mac.
> > 
> > The main thing I want to accomplish here is to put RegisterWithChrome back
> into
> > the synchronous component updater flow, so that I can later hook the
component
> > updater's update-complete notification to reload missing plugin placeholders
> (in
> > a race-free way).
> 
> I'm a bit behind on the latest state of this code.
> 
> Am I right that we will no longer be bundling flash as a special "bundled"
> distro with special code to handle this (and magic #defines like
> FLAPPER_VERSION) but instead will always be just bundling the component?

Exactly! That is the goal. This changelist does not accomplish that quite yet,
thanks to the Linux sandbox issues, but I think the main usage of
FLAPPER_VERSION is in GetBundledPepperFlash, which is Linux only (after this
change). Hopefully if we get the Linux sandbox issues resolved, we can fully
disentangle any Flash compilation from Chrome compilation, and pave the way for
physically separating the delivery of Chrome and Flash.

[Will Harris, 2016-07-07 19:47:05]

I think we'd need some kind of manual testing plan to verify the existing
behavior still works especially things like Debug flash taking precedence over
bundled component (people get upset if we break this)... can you modify/copy/use
https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul...
as a template here?

[waffles, 2016-07-08 16:59:31]

On 2016/07/07 19:47:05, Will Harris wrote:
> I think we'd need some kind of manual testing plan to verify the existing
> behavior still works especially things like Debug flash taking precedence over
> bundled component (people get upset if we break this)... can you
modify/copy/use
>
https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul...
> as a template here?

Yes, that's very helpful. We definitely need a manual testing plan like once the
end-to-end changes are complete to verify that all the behavior is WAI. To be
clear: are you suggesting also that we should run the manual testing plan after
this CL (and each CL on the road to the goal)?

[Will Harris, 2016-07-08 17:07:01]

On 2016/07/08 16:59:31, waffles wrote:
> On 2016/07/07 19:47:05, Will Harris wrote:
> > I think we'd need some kind of manual testing plan to verify the existing
> > behavior still works especially things like Debug flash taking precedence
over
> > bundled component (people get upset if we break this)... can you
> modify/copy/use
> >
>
https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul...
> > as a template here?
> 
> Yes, that's very helpful. We definitely need a manual testing plan like once
the
> end-to-end changes are complete to verify that all the behavior is WAI. To be
> clear: are you suggesting also that we should run the manual testing plan
after
> this CL (and each CL on the road to the goal)?

yes I think some sort of testing per CL is worth it since this code is pretty
hairy. Have you checked that system-installed Flash still works, and that a
flash content debugger will override the system component correctly?

[waffles, 2016-07-08 18:27:00]

On 2016/07/08 17:07:01, Will Harris wrote:
> On 2016/07/08 16:59:31, waffles wrote:
> > On 2016/07/07 19:47:05, Will Harris wrote:
> > > I think we'd need some kind of manual testing plan to verify the existing
> > > behavior still works especially things like Debug flash taking precedence
> over
> > > bundled component (people get upset if we break this)... can you
> > modify/copy/use
> > >
> >
>
https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul...
> > > as a template here?
> > 
> > Yes, that's very helpful. We definitely need a manual testing plan like once
> the
> > end-to-end changes are complete to verify that all the behavior is WAI. To
be
> > clear: are you suggesting also that we should run the manual testing plan
> after
> > this CL (and each CL on the road to the goal)?
> 
> yes I think some sort of testing per CL is worth it since this code is pretty
> hairy. Have you checked that system-installed Flash still works, and that a
> flash content debugger will override the system component correctly?

That seems reasonable, although I feel that this CL is low-risk w.r.t
system-installed flash. OK, everything in the test plan is Windows-specific, so
I guess I'll have to update my checkout in my VM. As I said, I have tested the
actual behavioral differences on Mac; further testing will be purely for the
sake of detecting regressions.

Please review the code anyways. Maenwhile, I'll work on getting the manual tests
done.

[Will Harris, 2016-07-08 21:00:07]

yup I stared at it long enough to think this is lgtm. the system/debug flash is
still registered earlier than this call in ChromeContentClient::AddPepperPlugins
so I think the logic still works. I'd really like to have some assurance via
some testing that it does work though, even if it's just through putting a TEST=
line in the CL and asking Chrome TE to do the test manually (by installing flash
content debugger, making sure it shows up).

[Greg K, 2016-07-11 17:56:23]

On 2016/07/08 21:00:07, Will Harris wrote:
> yup I stared at it long enough to think this is lgtm. the system/debug flash
is
> still registered earlier than this call in
ChromeContentClient::AddPepperPlugins
> so I think the logic still works. I'd really like to have some assurance via
> some testing that it does work though, even if it's just through putting a
TEST=
> line in the CL and asking Chrome TE to do the test manually (by installing
flash
> content debugger, making sure it shows up).

I stared at it as well and LGTM.

Please try running the flash debugger in OS X at least as Will said.

Regards,

Greg

[waffles, 2016-07-11 19:08:02]

Description was changed from

==========
Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
==========

to

==========
Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1
==========

[waffles, 2016-07-11 19:10:11]

On 2016/07/11 17:56:23, Greg Kerr wrote:
> On 2016/07/08 21:00:07, Will Harris wrote:
> > yup I stared at it long enough to think this is lgtm. the system/debug flash
> is
> > still registered earlier than this call in
> ChromeContentClient::AddPepperPlugins
> > so I think the logic still works. I'd really like to have some assurance via
> > some testing that it does work though, even if it's just through putting a
> TEST=
> > line in the CL and asking Chrome TE to do the test manually (by installing
> flash
> > content debugger, making sure it shows up).
> 
> I stared at it as well and LGTM.
> 
> Please try running the flash debugger in OS X at least as Will said.
> 
> Regards,
> 
> Greg

Added the TEST= line. Also manually did cases 1,4,6,7,9 on OS X.

[waffles, 2016-07-11 19:10:23]

The CQ bit was checked by waffles@chromium.org

[commit-bot: I haz the power, 2016-07-11 19:11:01]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-11 19:19:06]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-07-11 19:19:07]

Try jobs failed on following builders:
  chromium_presubmit on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[waffles, 2016-07-11 19:21:08]

waffles@chromium.org changed reviewers:
+ thakis@chromium.org

[waffles, 2016-07-11 19:21:09]

Nico, PTAL.

[Nico, 2016-07-11 19:22:25]

rs-lgtm, sounds like folks looked at this in detail already

[waffles, 2016-07-11 21:04:26]

The CQ bit was checked by waffles@chromium.org

[commit-bot: I haz the power, 2016-07-11 21:04:46]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-12 01:07:06]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-07-12 01:07:07]

Try jobs failed on following builders:
  linux_chromium_asan_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[waffles, 2016-07-12 16:01:36]

The CQ bit was checked by waffles@chromium.org

[commit-bot: I haz the power, 2016-07-12 16:03:23]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-12 20:16:17]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-07-12 20:16:18]

Try jobs failed on following builders:
  linux_chromium_asan_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[waffles, 2016-07-13 14:00:52]

The CQ bit was checked by waffles@chromium.org

[commit-bot: I haz the power, 2016-07-13 14:01:32]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-13 14:42:33]

Description was changed from

==========
Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1
==========

to

==========
Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1
==========

[commit-bot: I haz the power, 2016-07-13 14:42:34]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-07-13 14:44:43]

Description was changed from

==========
Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1
==========

to

==========
Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1

Committed: https://crrev.com/ae2a704ab712680c8b53a86e77eba78d0e591106
Cr-Commit-Position: refs/heads/master@{#405148}
==========

[commit-bot: I haz the power, 2016-07-13 14:44:44]

Patchset 1 (id:??) landed as
https://crrev.com/ae2a704ab712680c8b53a86e77eba78d0e591106
Cr-Commit-Position: refs/heads/master@{#405148}

[Dirk Pranke, 2016-07-13 22:15:01]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.chromium.org/2147083002/ by dpranke@chromium.org.

The reason for reverting is: I think this breaks some flash and eme-related
internal tests. See crbug.com/627959..

[waffles, 2016-07-14 00:04:07]

Description was changed from

==========
Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1

Committed: https://crrev.com/ae2a704ab712680c8b53a86e77eba78d0e591106
Cr-Commit-Position: refs/heads/master@{#405148}
==========

to

==========
Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1

Committed: https://crrev.com/ae2a704ab712680c8b53a86e77eba78d0e591106
Cr-Commit-Position: refs/heads/master@{#405148}
==========

[waffles, 2016-07-14 00:29:35]

On 2016/07/13 22:15:01, Dirk Pranke wrote:
> A revert of this CL (patchset #1 id:1) has been created in
> https://codereview.chromium.org/2147083002/ by mailto:dpranke@chromium.org.
> 
> The reason for reverting is: I think this breaks some flash and eme-related
> internal tests. See crbug.com/627959..

So, some more investigation indicates that there was a break introduced by an
earlier CL that this CL uncovered.
The breakage is fixed in this CL, which I will reland after going through the
test plan again.

[Will Harris, 2016-07-14 00:30:48]

On 2016/07/14 00:29:35, waffles wrote:
> On 2016/07/13 22:15:01, Dirk Pranke wrote:
> > A revert of this CL (patchset #1 id:1) has been created in
> > https://codereview.chromium.org/2147083002/ by mailto:dpranke@chromium.org.
> > 
> > The reason for reverting is: I think this breaks some flash and eme-related
> > internal tests. See crbug.com/627959..
> 
> So, some more investigation indicates that there was a break introduced by an
> earlier CL that this CL uncovered.
> The breakage is fixed in this CL, which I will reland after going through the
> test plan again.

I would like to subscribe to your newsletter (especially if the earlier CL is
mine...)

[waffles, 2016-07-14 00:33:29]

The CQ bit was checked by waffles@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-07-14 00:34:27]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[waffles, 2016-07-14 00:52:02]

On 2016/07/14 00:30:48, Will Harris wrote:
> On 2016/07/14 00:29:35, waffles wrote:
> > On 2016/07/13 22:15:01, Dirk Pranke wrote:
> > > A revert of this CL (patchset #1 id:1) has been created in
> > > https://codereview.chromium.org/2147083002/ by
mailto:dpranke@chromium.org.
> > > 
> > > The reason for reverting is: I think this breaks some flash and
eme-related
> > > internal tests. See crbug.com/627959..
> > 
> > So, some more investigation indicates that there was a break introduced by
an
> > earlier CL that this CL uncovered.
> > The breakage is fixed in this CL, which I will reland after going through
the
> > test plan again.
> 
> I would like to subscribe to your newsletter (especially if the earlier CL is
> mine...)

Nope, it's a long history of waffles@'s mistakes, where each fix introduces a
problem...
Starts with https://codereview.chromium.org/2041573002/ , where the problem is
that PathService::Override is called on the UI thread. Missed this one because I
was building/testing with the official release build and so DCHECKS were off.
So, tried to fix in
https://codereview.chromium.org/2116723002/diff/1/chrome/browser/component_up...
, but in that the .Append disappeared, which broke the Flash component
installer; however when I went to chrome://plugins and Adobe's about flash page,
everything worked fine. Since I mistakenly concluded that Flash was being loaded
by the component installer correctly (when in fact it was simply being loaded by
chrome_content_client.cc), I checked it in. The tests don't catch it for the
same reason that I missed it.
Then, halfway through making this change I rediscovered the code in
chrome_content_client.cc and decided to clean it up, which caused tests to
uncover the bug. I missed it in my earlier run through your test plan because I
was overconfident and looking at chrome://plugins and chrome://components to
verify that the component was loaded correctly, instead of going the full way to
Adobe's flash page.
(Ultimately, I discovered the problem at about the same time the tests did when
suddenly Flash wasn't working anymore in my synced-to-head changelists.)

The fix for the threading assert was merged into Beta, so I'll either have to
merge this into Beta or prepare another smaller CL to merge into Beta so that
component-updated Flash isn't broken there.

[waffles, 2016-07-14 00:58:20]

On 2016/07/14 00:52:02, waffles wrote:
> On 2016/07/14 00:30:48, Will Harris wrote:
> > On 2016/07/14 00:29:35, waffles wrote:
> > > On 2016/07/13 22:15:01, Dirk Pranke wrote:
> > > > A revert of this CL (patchset #1 id:1) has been created in
> > > > https://codereview.chromium.org/2147083002/ by
> mailto:dpranke@chromium.org.
> > > > 
> > > > The reason for reverting is: I think this breaks some flash and
> eme-related
> > > > internal tests. See crbug.com/627959..
> > > 
> > > So, some more investigation indicates that there was a break introduced by
> an
> > > earlier CL that this CL uncovered.
> > > The breakage is fixed in this CL, which I will reland after going through
> the
> > > test plan again.
> > 
> > I would like to subscribe to your newsletter (especially if the earlier CL
is
> > mine...)
> 
> Nope, it's a long history of waffles@'s mistakes, where each fix introduces a
> problem...
> Starts with https://codereview.chromium.org/2041573002/ , where the problem is
> that PathService::Override is called on the UI thread. Missed this one because
I
> was building/testing with the official release build and so DCHECKS were off.
> So, tried to fix in
>
https://codereview.chromium.org/2116723002/diff/1/chrome/browser/component_up...
> , but in that the .Append disappeared, which broke the Flash component
> installer; however when I went to chrome://plugins and Adobe's about flash
page,
> everything worked fine. Since I mistakenly concluded that Flash was being
loaded
> by the component installer correctly (when in fact it was simply being loaded
by
> chrome_content_client.cc), I checked it in. The tests don't catch it for the
> same reason that I missed it.
> Then, halfway through making this change I rediscovered the code in
> chrome_content_client.cc and decided to clean it up, which caused tests to
> uncover the bug. I missed it in my earlier run through your test plan because
I
> was overconfident and looking at chrome://plugins and chrome://components to
> verify that the component was loaded correctly, instead of going the full way
to
> Adobe's flash page.
> (Ultimately, I discovered the problem at about the same time the tests did
when
> suddenly Flash wasn't working anymore in my synced-to-head changelists.)
> 
> The fix for the threading assert was merged into Beta, so I'll either have to
> merge this into Beta or prepare another smaller CL to merge into Beta so that
> component-updated Flash isn't broken there.

Actually, because of the merge issue, I think I will just split the Append into
a separate CL altogether and rebase this on top of that once its in.

[commit-bot: I haz the power, 2016-07-14 01:28:21]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-07-14 01:28:22]

Dry run: Try jobs failed on following builders:
  win_chromium_rel_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)

[waffles, 2016-07-14 20:23:09]

The CQ bit was checked by waffles@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-07-14 20:23:43]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-14 21:36:12]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-07-14 21:36:13]

Dry run: This issue passed the CQ dry run.

[waffles, 2016-07-14 22:04:17]

The CQ bit was checked by waffles@chromium.org

[waffles, 2016-07-14 22:04:17]

The patchset sent to the CQ was uploaded after l-g-t-m from
kerrnel@chromium.org, thakis@chromium.org, wfh@chromium.org
Link to the patchset: https://codereview.chromium.org/2130803003/#ps40001
(title: "Keep using chrome_content_client to load bundled Flash")

[commit-bot: I haz the power, 2016-07-14 22:04:55]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-14 22:11:59]

Description was changed from

==========
Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1

Committed: https://crrev.com/ae2a704ab712680c8b53a86e77eba78d0e591106
Cr-Commit-Position: refs/heads/master@{#405148}
==========

to

==========
Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1

Committed: https://crrev.com/ae2a704ab712680c8b53a86e77eba78d0e591106
Cr-Commit-Position: refs/heads/master@{#405148}
==========

[commit-bot: I haz the power, 2016-07-14 22:12:03]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-07-14 22:12:35]

CQ bit was unchecked.

[commit-bot: I haz the power, 2016-07-14 22:14:48]

Description was changed from

==========
Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1

Committed: https://crrev.com/ae2a704ab712680c8b53a86e77eba78d0e591106
Cr-Commit-Position: refs/heads/master@{#405148}
==========

to

==========
Rearrange Flash component registration.
(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1

Committed: https://crrev.com/ae2a704ab712680c8b53a86e77eba78d0e591106
Committed: https://crrev.com/66a5763d27e17e175c46fcb1036f73c2c91e5f7f
Cr-Original-Commit-Position: refs/heads/master@{#405148}
Cr-Commit-Position: refs/heads/master@{#405597}
==========

[commit-bot: I haz the power, 2016-07-14 22:14:50]

Patchset 3 (id:??) landed as
https://crrev.com/66a5763d27e17e175c46fcb1036f73c2c91e5f7f
Cr-Commit-Position: refs/heads/master@{#405597}