Combine PDF plugin into the Chromium binary.

Combine PDF plugin into the Chromium binary.

On Windows, this moves it to chrome_child.dll. Overall binary sizes is 4.5 MB smaller (chrome_child.dll gets 3.6 MB larger while we drop the 8.1 MB pdf.dll). On Mac, the binary is 6.6 MB smaller. On Linux, it's 7MB smaller. This is from official release builds, after stripping on Linux.

The size savings are because we don't ship duplicate versions of V8, and also the PDF plugin uses some of base and net.

This depends on OOP PDF, since otherwise the V8 isolates for the plugin and Blink interact badly. That got turned on a few weeks ago.

BUG=453844
Committed: https://crrev.com/dc0e4390c0364515fa360c5292a4389784580a48
Cr-Commit-Position: refs/heads/master@{#314575}

[jam, 2015-01-12 20:43:41]

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

[jam, 2015-01-12 20:59:09]

jam@chromium.org changed reviewers:
+ cpu@chromium.org
- scottmg@chromium.org

[jam, 2015-01-31 00:35:18]

Patchset #7 (id:120001) has been deleted

[jam, 2015-02-02 22:09:49]

Patchset #11 (id:220001) has been deleted

[jam, 2015-02-02 22:10:02]

Patchset #11 (id:220001) has been deleted

[jam, 2015-02-02 22:10:16]

Patchset #11 (id:220001) has been deleted

[cpu_(ooo_6.6-7.5), 2015-02-02 22:55:23]

lgtm

Note, I did not looked carefully at the gyp madness.

https://codereview.chromium.org/799643004/diff/240001/chrome/renderer/DEPS
File chrome/renderer/DEPS (right):

https://codereview.chromium.org/799643004/diff/240001/chrome/renderer/DEPS#ne...
chrome/renderer/DEPS:3: "+chrome/installer/util/util_constants.h",
why is this?

[jam, 2015-02-02 23:15:52]

On 2015/02/02 22:55:23, cpu wrote:
> lgtm

Thanks, although this isn't quite ready yet :) I'm planning on fixing other
stuff in another patch first (how PDF is patched) and then I'll tell you when
it's ready.

[jam, 2015-02-03 18:50:23]

Patchset #16 (id:340001) has been deleted

[jam, 2015-02-03 19:11:38]

Ok this is now ready. The patch itself didn't change much, just the dependent
changes in PDFium and Chrome (unifying the IAT patching) landed so the trybots
should all be green now.

https://codereview.chromium.org/799643004/diff/240001/chrome/renderer/DEPS
File chrome/renderer/DEPS (right):

https://codereview.chromium.org/799643004/diff/240001/chrome/renderer/DEPS#ne...
chrome/renderer/DEPS:3: "+chrome/installer/util/util_constants.h",
On 2015/02/02 22:55:23, cpu wrote:
> why is this?

it was because of
https://codereview.chromium.org/799643004/diff/240001/chrome/renderer/chrome_...

[cpu_(ooo_6.6-7.5), 2015-02-04 17:33:14]

lgtm

[jam, 2015-02-04 17:34:06]

The CQ bit was checked by jam@chromium.org

[commit-bot: I haz the power, 2015-02-04 17:35:48]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/799643004/380001

[commit-bot: I haz the power, 2015-02-04 17:40:52]

Committed patchset #17 (id:380001)

[commit-bot: I haz the power, 2015-02-04 17:41:51]

Patchset 17 (id:??) landed as
https://crrev.com/dc0e4390c0364515fa360c5292a4389784580a48
Cr-Commit-Position: refs/heads/master@{#314575}

[kelvinp, 2015-02-04 17:53:32]

A revert of this CL (patchset #17 id:380001) has been created in
https://codereview.chromium.org/899033002/ by kelvinp@chromium.org.

The reason for reverting is: Compile step failed on Official Chrome
See 

http://build.chromium.org/p/chromium.chrome/builders/Google%20Chrome%20Linux%...

ninja: error: 'libpdf.so', needed by
'google-chrome-stable-42.0.2296.0-1.x86_64.rpm', missing and no known rule to
make it.

[jochen (gone - plz use gerrit), 2015-02-11 08:30:00]

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

[jochen (gone - plz use gerrit), 2015-02-11 08:30:02]

https://codereview.chromium.org/799643004/diff/380001/pdf/pdf.gyp
File pdf/pdf.gyp (left):

https://codereview.chromium.org/799643004/diff/380001/pdf/pdf.gyp#oldcode19
pdf/pdf.gyp:19: '../ppapi/ppapi.gyp:ppapi_cpp',
this breaks compilation with disable_nacl=1, because nothing pulls in ppapi_cpp
and we don't have e.g. pp::Var

[Jiang Jiang, 2015-02-12 17:14:42]

On 2015/02/11 08:30:02, jochen (slow) wrote:
> https://codereview.chromium.org/799643004/diff/380001/pdf/pdf.gyp
> File pdf/pdf.gyp (left):
> 
> https://codereview.chromium.org/799643004/diff/380001/pdf/pdf.gyp#oldcode19
> pdf/pdf.gyp:19: '../ppapi/ppapi.gyp:ppapi_cpp',
> this breaks compilation with disable_nacl=1, because nothing pulls in
ppapi_cpp
> and we don't have e.g. pp::Var

Yup, that broke builds for Opera.

Do you mind if I add the missing dependency like
https://code.google.com/p/chromium/codesearch#chromium/src/components/nacl/re...
did?

[jam, 2015-02-12 21:45:39]

On 2015/02/12 17:14:42, Jiang Jiang wrote:
> On 2015/02/11 08:30:02, jochen (slow) wrote:
> > https://codereview.chromium.org/799643004/diff/380001/pdf/pdf.gyp
> > File pdf/pdf.gyp (left):
> > 
> > https://codereview.chromium.org/799643004/diff/380001/pdf/pdf.gyp#oldcode19
> > pdf/pdf.gyp:19: '../ppapi/ppapi.gyp:ppapi_cpp',
> > this breaks compilation with disable_nacl=1, because nothing pulls in
> ppapi_cpp
> > and we don't have e.g. pp::Var
> 
> Yup, that broke builds for Opera.
> 
> Do you mind if I add the missing dependency like
>
https://code.google.com/p/chromium/codesearch#chromium/src/components/nacl/re...
> did?

just saw this comment, thanks for the fix

[Jiang Jiang, 2015-02-17 14:05:47]

BTW, is it intentional for this change to make --disable-out-of-process-pdf
useless?

[jam, 2015-02-18 16:34:48]

On 2015/02/17 14:05:47, Jiang Jiang wrote:
> BTW, is it intentional for this change to make --disable-out-of-process-pdf
> useless?

yes, it's incompatible