Fix "unreachable code" warnings (MSVC warning 4702) in Blink.

Fix "unreachable code" warnings (MSVC warning 4702) in Blink.

BUG=346399
TEST=none
R=thakis@chromium.org

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=169581

[Peter Kasting, 2014-03-18 17:40:47]

https://codereview.chromium.org/203743002/diff/1/Source/build/scripts/make_cs...
File Source/build/scripts/make_css_value_keywords.py (left):

https://codereview.chromium.org/203743002/diff/1/Source/build/scripts/make_cs...
Source/build/scripts/make_css_value_keywords.py:90: // two as only
QuirksOrUASheet is used in CSSValueKeyword.in
(ua_sheet_mode_values_keywords cannot be removed as it is used by cases in this
file which set the 'mode' property to 'UASheet')

[Nico, 2014-03-18 17:47:51]

https://codereview.chromium.org/203743002/diff/1/Source/build/scripts/make_cs...
File Source/build/scripts/make_css_value_keywords.py (left):

https://codereview.chromium.org/203743002/diff/1/Source/build/scripts/make_cs...
Source/build/scripts/make_css_value_keywords.py:95: return
isQuirksModeBehavior(mode);
I don't understand this part. Is this always dead? (+keishi, kenneth)

https://codereview.chromium.org/203743002/diff/1/Source/core/rendering/Render...
File Source/core/rendering/RenderBox.cpp (right):

https://codereview.chromium.org/203743002/diff/1/Source/core/rendering/Render...
Source/core/rendering/RenderBox.cpp:360: #if 1
Just delete this code; if Eric wants to pick this up again he can get it from
svn. (The FIXME has been here for several months.)

[eseidel, 2014-03-18 18:10:22]

https://codereview.chromium.org/203743002/diff/1/Source/core/rendering/Render...
File Source/core/rendering/RenderBox.cpp (right):

https://codereview.chromium.org/203743002/diff/1/Source/core/rendering/Render...
Source/core/rendering/RenderBox.cpp:360: #if 1
On 2014/03/18 17:47:51, Nico (on GMT time Mar 15 - 24) wrote:
> Just delete this code; if Eric wants to pick this up again he can get it from
> svn. (The FIXME has been here for several months.)

Sorry, this was an emergency fix done for a stable-blocker for M31, which was
rolling out tony's optimization.  We should just remove the whole codepath now. 
We have no plans to persue that optimization further.

[Peter Kasting, 2014-03-18 20:58:04]

https://codereview.chromium.org/203743002/diff/1/Source/build/scripts/make_cs...
File Source/build/scripts/make_css_value_keywords.py (left):

https://codereview.chromium.org/203743002/diff/1/Source/build/scripts/make_cs...
Source/build/scripts/make_css_value_keywords.py:95: return
isQuirksModeBehavior(mode);
On 2014/03/18 17:47:51, Nico (on GMT time Mar 15 - 24) wrote:
> I don't understand this part. Is this always dead? (+keishi, kenneth)

No keywords in CSSValueKeyword.in use it, and nothing in this file uses it
either; removing it doesn't cause any problems.  Leaving it causes problems
because the generated code has no cases, so we get code like this:

    ...
    case blah_ua_sheet:
    case xyz_ua_sheet:
        return isUASheetBehavior(mode);
        return isQuirksModeBehavior(mode);
    case blah_quirks_or_ua:
    ...

However, "quirks_mode_or_ua_sheet_mode_values_keywords" (below) is not dead, so
the isQuirksModeBehavior() routine cannot go away.

https://codereview.chromium.org/203743002/diff/1/Source/core/rendering/Render...
File Source/core/rendering/RenderBox.cpp (right):

https://codereview.chromium.org/203743002/diff/1/Source/core/rendering/Render...
Source/core/rendering/RenderBox.cpp:360: #if 1
On 2014/03/18 18:10:23, eseidel wrote:
> On 2014/03/18 17:47:51, Nico (on GMT time Mar 15 - 24) wrote:
> > Just delete this code; if Eric wants to pick this up again he can get it
from
> > svn. (The FIXME has been here for several months.)
> 
> Sorry, this was an emergency fix done for a stable-blocker for M31, which was
> rolling out tony's optimization.  We should just remove the whole codepath
now. 
> We have no plans to persue that optimization further.

Eric, would you be willing to make this change separately?  I'm happy to either
wait on it or to land this as-is, whatever you prefer.  I looked briefly at just
removing this function, but one of the callers has scary-sounding "this check is
likely wrong!" sorts of language that makes me gunshy about touching it myself
when I understand nothing about what it's doing.

[Nico, 2014-03-19 10:20:05]

lgtm with minor code deletion

Thanks for explaining!

https://codereview.chromium.org/203743002/diff/1/Source/build/scripts/make_cs...
File Source/build/scripts/make_css_value_keywords.py (left):

https://codereview.chromium.org/203743002/diff/1/Source/build/scripts/make_cs...
Source/build/scripts/make_css_value_keywords.py:95: return
isQuirksModeBehavior(mode);
On 2014/03/18 20:58:04, Peter Kasting wrote:
> On 2014/03/18 17:47:51, Nico (on GMT time Mar 15 - 24) wrote:
> > I don't understand this part. Is this always dead? (+keishi, kenneth)
> 
> No keywords in CSSValueKeyword.in use it, and nothing in this file uses it
> either; removing it doesn't cause any problems.  Leaving it causes problems
> because the generated code has no cases, so we get code like this:
> 
>     ...
>     case blah_ua_sheet:
>     case xyz_ua_sheet:
>         return isUASheetBehavior(mode);
>         return isQuirksModeBehavior(mode);
>     case blah_quirks_or_ua:
>     ...
> 
> However, "quirks_mode_or_ua_sheet_mode_values_keywords" (below) is not dead,
so
> the isQuirksModeBehavior() routine cannot go away.

Ah, now I understand your comment about lines 89,90, thanks :-)

https://codereview.chromium.org/203743002/diff/1/Source/core/rendering/Render...
File Source/core/rendering/RenderBox.cpp (right):

https://codereview.chromium.org/203743002/diff/1/Source/core/rendering/Render...
Source/core/rendering/RenderBox.cpp:360: #if 1
On 2014/03/18 20:58:04, Peter Kasting wrote:
> On 2014/03/18 18:10:23, eseidel wrote:
> > On 2014/03/18 17:47:51, Nico (on GMT time Mar 15 - 24) wrote:
> > > Just delete this code; if Eric wants to pick this up again he can get it
> from
> > > svn. (The FIXME has been here for several months.)
> > 
> > Sorry, this was an emergency fix done for a stable-blocker for M31, which
was
> > rolling out tony's optimization.  We should just remove the whole codepath
> now. 
> > We have no plans to persue that optimization further.
> 
> Eric, would you be willing to make this change separately?  I'm happy to
either
> wait on it or to land this as-is, whatever you prefer.  I looked briefly at
just
> removing this function, but one of the callers has scary-sounding "this check
is
> likely wrong!" sorts of language that makes me gunshy about touching it myself
> when I understand nothing about what it's doing.

You can change the function body to "
  // FIXME: remove function and callers
  return false;
"

[Peter Kasting, 2014-03-19 20:50:00]

https://codereview.chromium.org/203743002/diff/1/Source/core/rendering/Render...
File Source/core/rendering/RenderBox.cpp (right):

https://codereview.chromium.org/203743002/diff/1/Source/core/rendering/Render...
Source/core/rendering/RenderBox.cpp:360: #if 1
On 2014/03/19 10:20:05, Nico (on GMT time Mar 15 - 24) wrote:
> On 2014/03/18 20:58:04, Peter Kasting wrote:
> > On 2014/03/18 18:10:23, eseidel wrote:
> > > On 2014/03/18 17:47:51, Nico (on GMT time Mar 15 - 24) wrote:
> > > > Just delete this code; if Eric wants to pick this up again he can get it
> > from
> > > > svn. (The FIXME has been here for several months.)
> > > 
> > > Sorry, this was an emergency fix done for a stable-blocker for M31, which
> was
> > > rolling out tony's optimization.  We should just remove the whole codepath
> > now. 
> > > We have no plans to persue that optimization further.
> > 
> > Eric, would you be willing to make this change separately?  I'm happy to
> either
> > wait on it or to land this as-is, whatever you prefer.  I looked briefly at
> just
> > removing this function, but one of the callers has scary-sounding "this
check
> is
> > likely wrong!" sorts of language that makes me gunshy about touching it
myself
> > when I understand nothing about what it's doing.
> 
> You can change the function body to "
>   // FIXME: remove function and callers
>   return false;
> "

Done.

[Peter Kasting, 2014-03-19 20:50:04]

Committed patchset #2 manually as r169581 (presubmit successful).