Various minor cctest fixes to make ASAN a bit happier.

Various minor cctest fixes to make ASAN a bit happier.

   * Running with ASAN needs more stack, so don't set resource constraints too
     tight.

   * Checking boot time memory usage doesn't make sense when running with ASAN,
     it eats tons of memory for itself.

   * Fixed a malloc/delete[] mismatch: Not surprisingly, the pointer wrapped by
     a SmartArrayPointer should better be allocated by, well, NewArray...

Even with these 3 fixes, we still have a few failures when running our test
suite with ASAN. Most of them are either timeouts or failures caused by greatly
increased stack usage.

R=mstarzinger@chromium.org

Committed: http://code.google.com/p/v8/source/detail?r=14674

[Michael Starzinger, 2013-05-15 08:23:48]

LGTM with a nit.

https://codereview.chromium.org/15096011/diff/1/test/cctest/test-api.cc
File test/cctest/test-api.cc (right):

https://codereview.chromium.org/15096011/diff/1/test/cctest/test-api.cc#newco...
test/cctest/test-api.cc:15887: static const int stack_breathing_room = 256 *
1024;
nit: Let's use the "KB" constant from globals.h instead of 1024 here.

[Sven Panne, 2013-05-15 08:59:02]

Landing

https://codereview.chromium.org/15096011/diff/1/test/cctest/test-api.cc
File test/cctest/test-api.cc (right):

https://codereview.chromium.org/15096011/diff/1/test/cctest/test-api.cc#newco...
test/cctest/test-api.cc:15887: static const int stack_breathing_room = 256 *
1024;
On 2013/05/15 08:23:48, Michael Starzinger wrote:
> nit: Let's use the "KB" constant from globals.h instead of 1024 here.

Done.

https://codereview.chromium.org/15096011/diff/1/test/cctest/test-mark-compact.cc
File test/cctest/test-mark-compact.cc (right):

https://codereview.chromium.org/15096011/diff/1/test/cctest/test-mark-compact...
test/cctest/test-mark-compact.cc:474: !(defined(__has_feature) &&
__has_feature(address_sanitizer))
I had to do this check slightly different, otherwise GCC doesn't like it. :-(

[Sven Panne, 2013-05-15 08:59:34]

Committed patchset #2 manually as r14674 (presubmit successful).