Heap snapshot doesn't detect the fact that an old object was overriden by new one.

Heap snapshot doesn't detect the fact that an old object was overriden by new one.

It is a test only patch. I have no solution yet.

BUG=V8:2189
TEST=

Committed: https://code.google.com/p/v8/source/detail?r=13709

[yurys, 2013-02-21 09:00:30]

I like that we can test this issue but how can we prove that the test won't
flake?

https://chromiumcodereview.appspot.com/12321042/diff/1/test/cctest/test-heap-...
File test/cctest/test-heap-profiler.cc (right):

https://chromiumcodereview.appspot.com/12321042/diff/1/test/cctest/test-heap-...
test/cctest/test-heap-profiler.cc:397: // FIXME: We have problem with address
reuse.
"Object ids should be unique but it is not so at the moment."

[loislo, 2013-02-21 09:14:28]

On 2013/02/21 09:00:30, Yury Semikhatsky wrote:
> I like that we can test this issue but how can we prove that the test won't
> flake?
> 

I have no data about that but for 10k objects the reuse counter is ~4k.
I'll fix the flakiness if it happens.

>
https://chromiumcodereview.appspot.com/12321042/diff/1/test/cctest/test-heap-...
> File test/cctest/test-heap-profiler.cc (right):
> 
>
https://chromiumcodereview.appspot.com/12321042/diff/1/test/cctest/test-heap-...
> test/cctest/test-heap-profiler.cc:397: // FIXME: We have problem with address
> reuse.
> "Object ids should be unique but it is not so at the moment."

done

[yurys, 2013-02-21 12:14:10]

lgtm

[Yang, 2013-02-21 13:14:52]

On 2013/02/21 12:14:10, Yury Semikhatsky wrote:
> lgtm

LGTM.

[loislo, 2013-02-21 13:16:24]

Committed patchset #2 manually as r13709 (presubmit successful).