Remove overzealous check from %OptimizeFunctionOnNextCall.

Remove overzealous check from %OptimizeFunctionOnNextCall.

R=titzer@chromium.org

Committed: https://crrev.com/79748e3f7ced568355d00ea7d145790290e6df32
Cr-Commit-Position: refs/heads/master@{#26142}

[Michael Starzinger, 2015-01-19 12:19:22]

Reason for the change is, that the following two invocation behave unexpectedly
and surprisingly different:

1) $ d8 --allow-natives-syntax -e "function f() { try {} catch(e) {} } f();
%OptimizeFunctionOnNextCall(f); f();" --turbo-filter=* --trace-opt

2) $ d8 --allow-natives-syntax -e "function f() { try {} catch(e) {} }
%OptimizeFunctionOnNextCall(f); f();" --turbo-filter=* --trace-opt

Here 1) will not enter the optimizing pipeline, because optimization has been
disabled on the SharedFunctionInfo already, whereas 2) will enter the optimizing
pipeline and let the pipeline bailout on it's own. I think the pipeline should
have the final say in this matter.

[titzer, 2015-01-19 12:25:16]

On 2015/01/19 12:19:22, Michael Starzinger wrote:
> Reason for the change is, that the following two invocation behave
unexpectedly
> and surprisingly different:
> 
> 1) $ d8 --allow-natives-syntax -e "function f() { try {} catch(e) {} } f();
> %OptimizeFunctionOnNextCall(f); f();" --turbo-filter=* --trace-opt
> 
> 2) $ d8 --allow-natives-syntax -e "function f() { try {} catch(e) {} }
> %OptimizeFunctionOnNextCall(f); f();" --turbo-filter=* --trace-opt
> 
> Here 1) will not enter the optimizing pipeline, because optimization has been
> disabled on the SharedFunctionInfo already, whereas 2) will enter the
optimizing
> pipeline and let the pipeline bailout on it's own. I think the pipeline should
> have the final say in this matter.

Can we make the examples above into some mjsunit tests?

[Michael Starzinger, 2015-01-19 14:35:27]

Added test cases. PTAL.

https://codereview.chromium.org/863443003/diff/20001/test/mjsunit/compiler/op...
File test/mjsunit/compiler/opt-next-call-turbo.js (right):

https://codereview.chromium.org/863443003/diff/20001/test/mjsunit/compiler/op...
test/mjsunit/compiler/opt-next-call-turbo.js:22: // TODO(mstarzinger): Still not
optimized, make sure it is.
Note that this is sill not working as inteded, you can now see why compiler.cc
is very much "full of surprises". Will take care of that in a follow-up CL.

[titzer, 2015-01-19 14:37:27]

lgtm

[Michael Starzinger, 2015-01-19 15:24:01]

The CQ bit was checked by mstarzinger@chromium.org

[commit-bot: I haz the power, 2015-01-19 15:24:16]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/863443003/20001

[commit-bot: I haz the power, 2015-01-19 15:51:57]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-01-19 15:52:05]

Patchset 2 (id:??) landed as
https://crrev.com/79748e3f7ced568355d00ea7d145790290e6df32
Cr-Commit-Position: refs/heads/master@{#26142}