GC plugin: consider references equal to raw pointers.

GC plugin: consider references equal to raw pointers.

Unmanaged raw pointers to managed(GC) classes aren't safe across GCs,
hence fields with raw pointer types to such GCed classes are identified
and reported as errors or warnings.

Extend that checking to also apply to reference pointer types.

R=haraken
BUG=420515
Committed: https://crrev.com/1a33d92721acdfc37d71fd891eade8fbddd6c7dd
Cr-Commit-Position: refs/heads/master@{#342615}

[sof, 2015-08-07 07:05:31]

sigbjornf@opera.com changed reviewers:
+ oilpan-reviews@chromium.org

[sof, 2015-08-07 07:05:31]

please take a look.

Will separately follow up on the build errors that this triggers on the Blink
codebase.

[haraken, 2015-08-07 07:11:25]

Thanks, LGTM

[sof, 2015-08-07 07:50:45]

Thanks, extended it a bit so as to propagate the notion of reference type into
the error messages generated.

[haraken, 2015-08-07 08:08:51]

Still LGTM

[sof, 2015-08-07 08:24:09]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2015-08-07 08:24:15]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1274403002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1274403002/20001

[commit-bot: I haz the power, 2015-08-07 08:28:03]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-08-07 08:28:40]

Patchset 2 (id:??) landed as
https://crrev.com/70fe23896f81cd66ba1a205b0336f91c64e3fa4e
Cr-Commit-Position: refs/heads/master@{#342320}

[Nico, 2015-08-07 16:45:25]

A revert of this CL (patchset #2 id:20001) has been created in
https://codereview.chromium.org/1275383002/ by thakis@chromium.org.

The reason for reverting is: Turned the clang/win tot bots red:
http://build.chromium.org/p/chromium.fyi/builders/ClangToTWin/builds/2327/ste...

..\..\third_party\WebKit\Source\core/css/parser/CSSSupportsParser.h(13,1) : 
error: [blink-gc] Class 'CSSSupportsParser' contains invalid fields.
class CSSSupportsParser {
^
..\..\third_party\WebKit\Source\core/css/parser/CSSSupportsParser.h(31,5) : 
note: [blink-gc] Stack-allocated field 'm_parser' declared here:
    CSSParserImpl& m_parser;
    ^.

[sof, 2015-08-10 14:18:25]

With Blink r200242 in place, relanding.

[sof, 2015-08-10 14:18:32]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2015-08-10 14:18:50]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1274403002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1274403002/20001

[commit-bot: I haz the power, 2015-08-10 14:37:39]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-08-10 14:38:15]

Patchset 2 (id:??) landed as
https://crrev.com/1a33d92721acdfc37d71fd891eade8fbddd6c7dd
Cr-Commit-Position: refs/heads/master@{#342615}

[Nico, 2015-08-10 16:58:45]

Still doesn't build:
http://build.chromium.org/p/chromium.fyi/builders/ClangToTWin%28dll%29/builds...

..\..\third_party\WebKit\Source\core/css/parser/CSSParserObserverWrapper.h(12,1)
:  error: [blink-gc] Class 'CSSParserObserverWrapper' contains invalid fields.
class CSSParserObserverWrapper {
^
..\..\third_party\WebKit\Source\core/css/parser/CSSParserObserverWrapper.h(39,5)
:  note: [blink-gc] Stack-allocated field 'm_observer' declared here:
    CSSParserObserver& m_observer;
    ^

Maybe you didn't wait for that blink change to roll in? I'll wait another build
before reverting, but if this was the case please only land plugin changes
without them being behind a flag if the codebase is known to build with the
plugin change and if it's very unlikely that the codebase will regress until the
next clang roll. Otherwise, plugin changes should always be behind flags.