PlzNavigate: inform the renderer that a navigation is a POST

PlzNavigate: inform the renderer that a navigation is a POST

This CL moves the is_post params from StartNavigatioNparams(not used in
PlzNavigate) to CommonNavigationParams(used in PlzNavigate). This ensures that
when a POST navigation commits, it will be properly recorder as a POST.

BUG=575210
TBR=nasko@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation

Committed: https://crrev.com/b39c17cacc5237edae1eac622754941ae82b74d3
Cr-Commit-Position: refs/heads/master@{#378204}

[clamy, 2016-02-09 14:52:31]

Description was changed from

==========
PlzNavigate: inform the renderer that a navigation is a POST

This CL moves the is_post params from StartNavigatioNparams(not used in
PlzNavigate) to CommonNavigationParams(used in PlzNavigate). This ensures that
when a POST navigation commits, it will be properly recorder as a POST.

BUG=575210
==========

to

==========
PlzNavigate: inform the renderer that a navigation is a POST

This CL moves the is_post params from StartNavigatioNparams(not used in
PlzNavigate) to CommonNavigationParams(used in PlzNavigate). This ensures that
when a POST navigation commits, it will be properly recorder as a POST.

BUG=575210
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

[clamy, 2016-02-12 12:26:30]

Description was changed from

==========
PlzNavigate: inform the renderer that a navigation is a POST

This CL moves the is_post params from StartNavigatioNparams(not used in
PlzNavigate) to CommonNavigationParams(used in PlzNavigate). This ensures that
when a POST navigation commits, it will be properly recorder as a POST.

BUG=575210
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

to

==========
PlzNavigate: inform the renderer that a navigation is a POST

This CL moves the is_post params from StartNavigatioNparams(not used in
PlzNavigate) to CommonNavigationParams(used in PlzNavigate). This ensures that
when a POST navigation commits, it will be properly recorder as a POST.

BUG=575210
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

[clamy, 2016-02-12 12:26:31]

clamy@chromium.org changed reviewers:
+ creis@chromium.org, nasko@chromium.org

[clamy, 2016-02-12 12:27:17]

@nasko, creis: PTAL

This is the first patch to properly support POST navigations in PlzNavigate.

[Charlie Reis, 2016-02-17 05:55:28]

I think this should work.  Can you run linux_site_isolation try jobs as well?

A few questions below to be safe.

https://codereview.chromium.org/1678303004/diff/60001/content/browser/loader/...
File content/browser/loader/resource_dispatcher_host_impl.cc (right):

https://codereview.chromium.org/1678303004/diff/60001/content/browser/loader/...
content/browser/loader/resource_dispatcher_host_impl.cc:1491: // PlzNavigate:
always set the method to GET when gaining access to the
nit: Capitalize Always

https://codereview.chromium.org/1678303004/diff/60001/content/browser/loader/...
content/browser/loader/resource_dispatcher_host_impl.cc:1492: // stream that
contains the response body of a navigation.
I think this comment needs a bit more elaboration.  It took me a while to
remember what is_navigation_stream_request was about, and why it would make
sense for it to be GET even if the actual navigation was POST.

(Also, I'm not sure what the concrete reason is-- what goes wrong if it's POST?)

https://codereview.chromium.org/1678303004/diff/60001/content/browser/loader/...
content/browser/loader/resource_dispatcher_host_impl.cc:2223:
info.common_params.is_post ? "POST" : "GET",
There might be some loss of generality here.  Was it ever possible for method to
be something else before, like PUT or HEAD?  Hopefully not.

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
File content/common/navigation_params.h (left):

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
content/common/navigation_params.h:138: // The request method: GET, POST, etc.
Again, I wonder about whether the "etc" was intentional or if GET and POST are
really the only things possible.

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
File content/common/navigation_params.h (right):

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
content/common/navigation_params.h:63: bool is_post);
I would think that moving this to CommonNavParams implies it gets used on the
commit side as well as on the request side (as opposed to RequestNavParams,
which get used both by default and in PlzNavigate).

Is it used during commit?  Or is there another reason to move it?  Maybe it's so
that it can go browser to renderer (like RequestNavParams) and renderer to
browser (like BeginNavParams)?

[clamy, 2016-02-17 12:39:39]

Thanks!

https://codereview.chromium.org/1678303004/diff/60001/content/browser/loader/...
File content/browser/loader/resource_dispatcher_host_impl.cc (right):

https://codereview.chromium.org/1678303004/diff/60001/content/browser/loader/...
content/browser/loader/resource_dispatcher_host_impl.cc:1492: // stream that
contains the response body of a navigation.
On 2016/02/17 05:55:28, Charlie Reis (catching up) wrote:
> I think this comment needs a bit more elaboration.  It took me a while to
> remember what is_navigation_stream_request was about, and why it would make
> sense for it to be GET even if the actual navigation was POST.
> 
> (Also, I'm not sure what the concrete reason is-- what goes wrong if it's
POST?)

I've added a sentence that explains what happens if we don't do it. Is it now
clearer?

https://codereview.chromium.org/1678303004/diff/60001/content/browser/loader/...
content/browser/loader/resource_dispatcher_host_impl.cc:2223:
info.common_params.is_post ? "POST" : "GET",
On 2016/02/17 05:55:28, Charlie Reis (catching up) wrote:
> There might be some loss of generality here.  Was it ever possible for method
to
> be something else before, like PUT or HEAD?  Hopefully not.

I changed the boolean to a string, this should not be an issue now.

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
File content/common/navigation_params.h (left):

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
content/common/navigation_params.h:138: // The request method: GET, POST, etc.
On 2016/02/17 05:55:28, Charlie Reis (catching up) wrote:
> Again, I wonder about whether the "etc" was intentional or if GET and POST are
> really the only things possible.

Changed the boolean to a string.

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
File content/common/navigation_params.h (right):

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
content/common/navigation_params.h:63: bool is_post);
On 2016/02/17 05:55:28, Charlie Reis (catching up) wrote:
> I would think that moving this to CommonNavParams implies it gets used on the
> commit side as well as on the request side (as opposed to RequestNavParams,
> which get used both by default and in PlzNavigate).
> 
> Is it used during commit?  Or is there another reason to move it?  Maybe it's
so
> that it can go browser to renderer (like RequestNavParams) and renderer to
> browser (like BeginNavParams)?

It is used both at request time in the browser, and at commit time in the
renderer:
- request time: we need to know the method to do the HTTP request
- commit time: we need to send in DidCommitProvisionalLoad whether the
navigation was a POST or not

[Charlie Reis, 2016-02-17 21:52:06]

Thanks.  Still trying to understand why the new parameter is in CommonNavParams
(or what it means for it to be there), but otherwise this seems fine.

https://codereview.chromium.org/1678303004/diff/60001/content/browser/loader/...
File content/browser/loader/resource_dispatcher_host_impl.cc (right):

https://codereview.chromium.org/1678303004/diff/60001/content/browser/loader/...
content/browser/loader/resource_dispatcher_host_impl.cc:1492: // stream that
contains the response body of a navigation.
On 2016/02/17 12:39:38, clamy wrote:
> On 2016/02/17 05:55:28, Charlie Reis (catching up) wrote:
> > I think this comment needs a bit more elaboration.  It took me a while to
> > remember what is_navigation_stream_request was about, and why it would make
> > sense for it to be GET even if the actual navigation was POST.
> > 
> > (Also, I'm not sure what the concrete reason is-- what goes wrong if it's
> POST?)
> 
> I've added a sentence that explains what happens if we don't do it. Is it now
> clearer?

Yes, thanks.

https://codereview.chromium.org/1678303004/diff/60001/content/browser/loader/...
content/browser/loader/resource_dispatcher_host_impl.cc:2223:
info.common_params.is_post ? "POST" : "GET",
On 2016/02/17 12:39:38, clamy wrote:
> On 2016/02/17 05:55:28, Charlie Reis (catching up) wrote:
> > There might be some loss of generality here.  Was it ever possible for
method
> to
> > be something else before, like PUT or HEAD?  Hopefully not.
> 
> I changed the boolean to a string, this should not be an issue now.

Ok, thanks.  (Out of curiosity, was it possible for it to be anything else? 
This question comes up from time to time and I don't know the answer.)

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
File content/common/navigation_params.h (right):

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
content/common/navigation_params.h:63: bool is_post);
On 2016/02/17 12:39:39, clamy wrote:
> On 2016/02/17 05:55:28, Charlie Reis (catching up) wrote:
> > I would think that moving this to CommonNavParams implies it gets used on
the
> > commit side as well as on the request side (as opposed to RequestNavParams,
> > which get used both by default and in PlzNavigate).
> > 
> > Is it used during commit?  Or is there another reason to move it?  Maybe
it's
> so
> > that it can go browser to renderer (like RequestNavParams) and renderer to
> > browser (like BeginNavParams)?
> 
> It is used both at request time in the browser, and at commit time in the
> renderer:
> - request time: we need to know the method to do the HTTP request
> - commit time: we need to send in DidCommitProvisionalLoad whether the
> navigation was a POST or not

Ah, I'm missing the second part.  Where do we use it in
DidCommitProvisionalLoad?

Actually, I see a different is_post in
FrameHostMsg_DidCommitProvisionalLoad_Params, which gets assigned in
SendDidCommitProvisionalLoad.  Should that somehow be replaced by this one? 
(Then again, that's from the previous patch set.  Does the new |method|
parameter get used on commit, and how?  Does it ever get sent from renderer to
browser?)

Apologies if I'm misunderstanding the way CommonNavParams gets used.  The part
that's confusing me is that the comment above says CommonNavParams is "Used by
all navigation IPCs," but it doesn't seem to be sent in DidCommit.

https://codereview.chromium.org/1678303004/diff/100001/content/renderer/rende...
File content/renderer/render_frame_impl.cc (left):

https://codereview.chromium.org/1678303004/diff/100001/content/renderer/rende...
content/renderer/render_frame_impl.cc:5462:
request.setHTTPMethod(WebString::fromUTF8("POST"));
Sanity check: How does this happen now?  The CreateURLRequestForNavigation call
on line 5370?

[clamy, 2016-02-22 12:48:36]

Thanks!

https://codereview.chromium.org/1678303004/diff/60001/content/browser/loader/...
File content/browser/loader/resource_dispatcher_host_impl.cc (right):

https://codereview.chromium.org/1678303004/diff/60001/content/browser/loader/...
content/browser/loader/resource_dispatcher_host_impl.cc:2223:
info.common_params.is_post ? "POST" : "GET",
On 2016/02/17 21:52:06, Charlie Reis (slow til Mar 7) wrote:
> On 2016/02/17 12:39:38, clamy wrote:
> > On 2016/02/17 05:55:28, Charlie Reis (catching up) wrote:
> > > There might be some loss of generality here.  Was it ever possible for
> method
> > to
> > > be something else before, like PUT or HEAD?  Hopefully not.
> > 
> > I changed the boolean to a string, this should not be an issue now.
> 
> Ok, thanks.  (Out of curiosity, was it possible for it to be anything else? 
> This question comes up from time to time and I don't know the answer.)

Honestly I don't know. I know in some places in the code we assume it, and it's
not possible to have a browser-initiated navigation with a method other than
POST or GET. However, I'm not sure that this is true for render-initiated
navigations.

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
File content/common/navigation_params.h (right):

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
content/common/navigation_params.h:63: bool is_post);
On 2016/02/17 21:52:06, Charlie Reis (slow til Mar 7) wrote:
> On 2016/02/17 12:39:39, clamy wrote:
> > On 2016/02/17 05:55:28, Charlie Reis (catching up) wrote:
> > > I would think that moving this to CommonNavParams implies it gets used on
> the
> > > commit side as well as on the request side (as opposed to
RequestNavParams,
> > > which get used both by default and in PlzNavigate).
> > > 
> > > Is it used during commit?  Or is there another reason to move it?  Maybe
> it's
> > so
> > > that it can go browser to renderer (like RequestNavParams) and renderer to
> > > browser (like BeginNavParams)?
> > 
> > It is used both at request time in the browser, and at commit time in the
> > renderer:
> > - request time: we need to know the method to do the HTTP request
> > - commit time: we need to send in DidCommitProvisionalLoad whether the
> > navigation was a POST or not
> 
> Ah, I'm missing the second part.  Where do we use it in
> DidCommitProvisionalLoad?
> 
> Actually, I see a different is_post in
> FrameHostMsg_DidCommitProvisionalLoad_Params, which gets assigned in
> SendDidCommitProvisionalLoad.  Should that somehow be replaced by this one? 
> (Then again, that's from the previous patch set.  Does the new |method|
> parameter get used on commit, and how?  Does it ever get sent from renderer to
> browser?)
> 
> Apologies if I'm misunderstanding the way CommonNavParams gets used.  The part
> that's confusing me is that the comment above says CommonNavParams is "Used by
> all navigation IPCs," but it doesn't seem to be sent in DidCommit.

Yes it seems the comment above CommonNavParams is a bit misleading. They don't
get used in DidCommitProvisionalLoad, since DidCommitProvisionalLoad already
uses a NavigationParameters struct as a base. We could probably unify that, but
it has not been done yet.

The is_post parameter in DidCommitProvisionalLoad is how the information about
the method gets used at commit time (we get it from the method of the WebRequest
that was initially set using the method param from CommonNavigationParams).

I'm not changing it right now, since I plan to make a next patch that will move
the knowledge of whether was a POST or not to the FrameNavigationEntry, and I
thought I would change the is_post parameter to method at that time.

https://codereview.chromium.org/1678303004/diff/100001/content/renderer/rende...
File content/renderer/render_frame_impl.cc (left):

https://codereview.chromium.org/1678303004/diff/100001/content/renderer/rende...
content/renderer/render_frame_impl.cc:5462:
request.setHTTPMethod(WebString::fromUTF8("POST"));
On 2016/02/17 21:52:06, Charlie Reis (slow til Mar 7) wrote:
> Sanity check: How does this happen now?  The CreateURLRequestForNavigation
call
> on line 5370?

Yes.

[Charlie Reis, 2016-02-27 00:21:49]

Sorry for the delay!

LGTM if you agree with the reasoning below.

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
File content/common/navigation_params.h (right):

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
content/common/navigation_params.h:63: bool is_post);
On 2016/02/22 12:48:36, clamy wrote:
> On 2016/02/17 21:52:06, Charlie Reis (slow til Mar 7) wrote:
> > On 2016/02/17 12:39:39, clamy wrote:
> > > On 2016/02/17 05:55:28, Charlie Reis (catching up) wrote:
> > > > I would think that moving this to CommonNavParams implies it gets used
on
> > the
> > > > commit side as well as on the request side (as opposed to
> RequestNavParams,
> > > > which get used both by default and in PlzNavigate).
> > > > 
> > > > Is it used during commit?  Or is there another reason to move it?  Maybe
> > it's
> > > so
> > > > that it can go browser to renderer (like RequestNavParams) and renderer
to
> > > > browser (like BeginNavParams)?
> > > 
> > > It is used both at request time in the browser, and at commit time in the
> > > renderer:
> > > - request time: we need to know the method to do the HTTP request
> > > - commit time: we need to send in DidCommitProvisionalLoad whether the
> > > navigation was a POST or not
> > 
> > Ah, I'm missing the second part.  Where do we use it in
> > DidCommitProvisionalLoad?
> > 
> > Actually, I see a different is_post in
> > FrameHostMsg_DidCommitProvisionalLoad_Params, which gets assigned in
> > SendDidCommitProvisionalLoad.  Should that somehow be replaced by this one? 
> > (Then again, that's from the previous patch set.  Does the new |method|
> > parameter get used on commit, and how?  Does it ever get sent from renderer
to
> > browser?)
> > 
> > Apologies if I'm misunderstanding the way CommonNavParams gets used.  The
part
> > that's confusing me is that the comment above says CommonNavParams is "Used
by
> > all navigation IPCs," but it doesn't seem to be sent in DidCommit.
> 
> Yes it seems the comment above CommonNavParams is a bit misleading. They don't
> get used in DidCommitProvisionalLoad, since DidCommitProvisionalLoad already
> uses a NavigationParameters struct as a base. We could probably unify that,
but
> it has not been done yet.

It's fine if this isn't used at commit time, but then it seems like it would
belong in RequestNavigationParams (i.e., "Parameters used both in the current
architecture and PlzNavigate" and sent only from browser to renderer).

I'll assume that it still belongs in CommonNavParams because it's used in both
browser-to-renderer (RequestNavParams) and renderer-to-browser (BeignNavParams)
cases.  Feel free to proceed if that's correct.


> 
> The is_post parameter in DidCommitProvisionalLoad is how the information about
> the method gets used at commit time (we get it from the method of the
WebRequest
> that was initially set using the method param from CommonNavigationParams).
> 
> I'm not changing it right now, since I plan to make a next patch that will
move
> the knowledge of whether was a POST or not to the FrameNavigationEntry, and I
> thought I would change the is_post parameter to method at that time.

I'm cautious about moving that from NavEntry to FrameNavEntry because of the
possible implications to serialized session history on disk.  There's probably a
safe way to proceed and we can discuss that separately, though; no need to hold
this up.

[clamy, 2016-02-29 10:54:02]

Thanks!

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
File content/common/navigation_params.h (right):

https://codereview.chromium.org/1678303004/diff/60001/content/common/navigati...
content/common/navigation_params.h:63: bool is_post);
On 2016/02/27 00:21:49, Charlie Reis (OOO til Mar 7) wrote:
> On 2016/02/22 12:48:36, clamy wrote:
> > On 2016/02/17 21:52:06, Charlie Reis (slow til Mar 7) wrote:
> > > On 2016/02/17 12:39:39, clamy wrote:
> > > > On 2016/02/17 05:55:28, Charlie Reis (catching up) wrote:
> > > > > I would think that moving this to CommonNavParams implies it gets used
> on
> > > the
> > > > > commit side as well as on the request side (as opposed to
> > RequestNavParams,
> > > > > which get used both by default and in PlzNavigate).
> > > > > 
> > > > > Is it used during commit?  Or is there another reason to move it? 
Maybe
> > > it's
> > > > so
> > > > > that it can go browser to renderer (like RequestNavParams) and
renderer
> to
> > > > > browser (like BeginNavParams)?
> > > > 
> > > > It is used both at request time in the browser, and at commit time in
the
> > > > renderer:
> > > > - request time: we need to know the method to do the HTTP request
> > > > - commit time: we need to send in DidCommitProvisionalLoad whether the
> > > > navigation was a POST or not
> > > 
> > > Ah, I'm missing the second part.  Where do we use it in
> > > DidCommitProvisionalLoad?
> > > 
> > > Actually, I see a different is_post in
> > > FrameHostMsg_DidCommitProvisionalLoad_Params, which gets assigned in
> > > SendDidCommitProvisionalLoad.  Should that somehow be replaced by this
one? 
> > > (Then again, that's from the previous patch set.  Does the new |method|
> > > parameter get used on commit, and how?  Does it ever get sent from
renderer
> to
> > > browser?)
> > > 
> > > Apologies if I'm misunderstanding the way CommonNavParams gets used.  The
> part
> > > that's confusing me is that the comment above says CommonNavParams is
"Used
> by
> > > all navigation IPCs," but it doesn't seem to be sent in DidCommit.
> > 
> > Yes it seems the comment above CommonNavParams is a bit misleading. They
don't
> > get used in DidCommitProvisionalLoad, since DidCommitProvisionalLoad already
> > uses a NavigationParameters struct as a base. We could probably unify that,
> but
> > it has not been done yet.
> 
> It's fine if this isn't used at commit time, but then it seems like it would
> belong in RequestNavigationParams (i.e., "Parameters used both in the current
> architecture and PlzNavigate" and sent only from browser to renderer).
> 
> I'll assume that it still belongs in CommonNavParams because it's used in both
> browser-to-renderer (RequestNavParams) and renderer-to-browser
(BeignNavParams)
> cases.  Feel free to proceed if that's correct.

That's indeed the case.

> 
> 
> > 
> > The is_post parameter in DidCommitProvisionalLoad is how the information
about
> > the method gets used at commit time (we get it from the method of the
> WebRequest
> > that was initially set using the method param from CommonNavigationParams).
> > 
> > I'm not changing it right now, since I plan to make a next patch that will
> move
> > the knowledge of whether was a POST or not to the FrameNavigationEntry, and
I
> > thought I would change the is_post parameter to method at that time.
> 
> I'm cautious about moving that from NavEntry to FrameNavEntry because of the
> possible implications to serialized session history on disk.  There's probably
a
> safe way to proceed and we can discuss that separately, though; no need to
hold
> this up.

OK, I probably won't get started started on this right away though, so there
will be time to see how we can safely proceed.

[clamy, 2016-02-29 11:35:40]

Description was changed from

==========
PlzNavigate: inform the renderer that a navigation is a POST

This CL moves the is_post params from StartNavigatioNparams(not used in
PlzNavigate) to CommonNavigationParams(used in PlzNavigate). This ensures that
when a POST navigation commits, it will be properly recorder as a POST.

BUG=575210
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

to

==========
PlzNavigate: inform the renderer that a navigation is a POST

This CL moves the is_post params from StartNavigatioNparams(not used in
PlzNavigate) to CommonNavigationParams(used in PlzNavigate). This ensures that
when a POST navigation commits, it will be properly recorder as a POST.

BUG=575210
TBR=nasko@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

[clamy, 2016-02-29 11:35:46]

The CQ bit was checked by clamy@chromium.org

[clamy, 2016-02-29 11:35:46]

The patchset sent to the CQ was uploaded after l-g-t-m from creis@chromium.org
Link to the patchset: https://codereview.chromium.org/1678303004/#ps120001
(title: "Rebase")

[commit-bot: I haz the power, 2016-02-29 11:36:04]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1678303004/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1678303004/120001

[commit-bot: I haz the power, 2016-02-29 13:33:37]

Description was changed from

==========
PlzNavigate: inform the renderer that a navigation is a POST

This CL moves the is_post params from StartNavigatioNparams(not used in
PlzNavigate) to CommonNavigationParams(used in PlzNavigate). This ensures that
when a POST navigation commits, it will be properly recorder as a POST.

BUG=575210
TBR=nasko@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

to

==========
PlzNavigate: inform the renderer that a navigation is a POST

This CL moves the is_post params from StartNavigatioNparams(not used in
PlzNavigate) to CommonNavigationParams(used in PlzNavigate). This ensures that
when a POST navigation commits, it will be properly recorder as a POST.

BUG=575210
TBR=nasko@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

[commit-bot: I haz the power, 2016-02-29 13:33:38]

Committed patchset #7 (id:120001)

[commit-bot: I haz the power, 2016-02-29 13:35:34]

Description was changed from

==========
PlzNavigate: inform the renderer that a navigation is a POST

This CL moves the is_post params from StartNavigatioNparams(not used in
PlzNavigate) to CommonNavigationParams(used in PlzNavigate). This ensures that
when a POST navigation commits, it will be properly recorder as a POST.

BUG=575210
TBR=nasko@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

to

==========
PlzNavigate: inform the renderer that a navigation is a POST

This CL moves the is_post params from StartNavigatioNparams(not used in
PlzNavigate) to CommonNavigationParams(used in PlzNavigate). This ensures that
when a POST navigation commits, it will be properly recorder as a POST.

BUG=575210
TBR=nasko@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation

Committed: https://crrev.com/b39c17cacc5237edae1eac622754941ae82b74d3
Cr-Commit-Position: refs/heads/master@{#378204}
==========

[commit-bot: I haz the power, 2016-02-29 13:35:35]

Patchset 7 (id:??) landed as
https://crrev.com/b39c17cacc5237edae1eac622754941ae82b74d3
Cr-Commit-Position: refs/heads/master@{#378204}