Make compilers agree on source position of thrown errors.

Make compilers agree on source position of thrown errors.

This makes the compilers agree on the source position of a message
generated by "throw new Error()", it points to the beginning of the
throw directive.

R=titzer@chromium.org
TEST=message/regress/regress-3995
BUG=v8:3995
LOG=N

Committed: https://crrev.com/10dd9ce8beef4e57bc0ad90cb336d8dfe7aedf64
Cr-Commit-Position: refs/heads/master@{#27775}

[Michael Starzinger, 2015-04-10 11:43:24]

mstarzinger@chromium.org changed reviewers:
+ titzer@chromium.org

[Michael Starzinger, 2015-04-10 11:49:11]

mstarzinger@chromium.org changed reviewers:
+ arv@chromium.org, verwaest@chromium.org

[Michael Starzinger, 2015-04-10 11:49:12]

Arv & Toon: Please advise.
Ben: PTAL.

https://codereview.chromium.org/1049703002/diff/40001/test/cctest/test-api.cc
File test/cctest/test-api.cc (right):

https://codereview.chromium.org/1049703002/diff/40001/test/cctest/test-api.cc...
test/cctest/test-api.cc:9798: strcmp("TypeError: object is not a function",
*exception_value1));
Toon: Can you please advise on this change. Does this indicate a regression?

https://codereview.chromium.org/1049703002/diff/40001/test/mjsunit/harmony/de...
File test/mjsunit/harmony/debug-step-into-constructor.js (right):

https://codereview.chromium.org/1049703002/diff/40001/test/mjsunit/harmony/de...
test/mjsunit/harmony/debug-step-into-constructor.js:27: constructor() {  // 1.
Arv: Can you please advise on this change. Does this indicate a regression?

[Toon Verwaest, 2015-04-10 13:05:58]

If you change positions, you will have to be careful that it doesn't affect
blink layout tests as well. If it does, you'll at least have to rebaseline.

https://codereview.chromium.org/1049703002/diff/40001/test/cctest/test-api.cc
File test/cctest/test-api.cc (right):

https://codereview.chromium.org/1049703002/diff/40001/test/cctest/test-api.cc...
test/cctest/test-api.cc:9798: strcmp("TypeError: object is not a function",
*exception_value1));
On 2015/04/10 11:49:11, Michael Starzinger wrote:
> Toon: Can you please advise on this change. Does this indicate a regression?

Yes, this indicates that the position of the error doesn't match anymore what
the CallPrinter uses to find the matching call target.

https://codereview.chromium.org/1049703002/diff/40001/test/webkit/exception-f...
File test/webkit/exception-for-nonobject-expected.txt (right):

https://codereview.chromium.org/1049703002/diff/40001/test/webkit/exception-f...
test/webkit/exception-for-nonobject-expected.txt:29: PASS new {}.undefined threw
exception TypeError: undefined is not a function.
Probably the same here?

[Michael Starzinger, 2015-04-10 13:56:39]

mstarzinger@chromium.org changed reviewers:
- arv@chromium.org

[Michael Starzinger, 2015-04-10 13:56:40]

https://codereview.chromium.org/1049703002/diff/40001/test/mjsunit/harmony/de...
File test/mjsunit/harmony/debug-step-into-constructor.js (right):

https://codereview.chromium.org/1049703002/diff/40001/test/mjsunit/harmony/de...
test/mjsunit/harmony/debug-step-into-constructor.js:27: constructor() {  // 1.
On 2015/04/10 11:49:12, Michael Starzinger wrote:
> Arv: Can you please advise on this change. Does this indicate a regression?

OK, I managed to find this regression and fixed it. Please disregard my previous
comment here. Arv, I took you off the reviewer list. Sorry for the noise on this
one!

[Michael Starzinger, 2015-04-10 14:08:27]

Thanks for the comments. PTAL.

https://codereview.chromium.org/1049703002/diff/40001/test/cctest/test-api.cc
File test/cctest/test-api.cc (right):

https://codereview.chromium.org/1049703002/diff/40001/test/cctest/test-api.cc...
test/cctest/test-api.cc:9798: strcmp("TypeError: object is not a function",
*exception_value1));
On 2015/04/10 13:05:58, Toon Verwaest wrote:
> On 2015/04/10 11:49:11, Michael Starzinger wrote:
> > Toon: Can you please advise on this change. Does this indicate a regression?
> 
> Yes, this indicates that the position of the error doesn't match anymore what
> the CallPrinter uses to find the matching call target.

Done.

https://codereview.chromium.org/1049703002/diff/40001/test/webkit/exception-f...
File test/webkit/exception-for-nonobject-expected.txt (right):

https://codereview.chromium.org/1049703002/diff/40001/test/webkit/exception-f...
test/webkit/exception-for-nonobject-expected.txt:29: PASS new {}.undefined threw
exception TypeError: undefined is not a function.
On 2015/04/10 13:05:58, Toon Verwaest wrote:
> Probably the same here?

Done.

[Toon Verwaest, 2015-04-10 15:24:28]

CallPrinter looks good. Uniformity ftw

[titzer, 2015-04-13 08:29:04]

lgtm

[Michael Starzinger, 2015-04-13 08:33:39]

The CQ bit was checked by mstarzinger@chromium.org

[commit-bot: I haz the power, 2015-04-13 08:33:50]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1049703002/80001

[commit-bot: I haz the power, 2015-04-13 09:02:57]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2015-04-13 09:03:04]

Patchset 5 (id:??) landed as
https://crrev.com/10dd9ce8beef4e57bc0ad90cb336d8dfe7aedf64
Cr-Commit-Position: refs/heads/master@{#27775}