Fix confusion over on-disk object id sizes during instance serialization.

Fix confusion over on-disk object id sizes during instance serialization.

To avoid confusion in the future the WritePointer methods have been
renamed to the more descriptive WriteObjectId.

Committed: https://code.google.com/p/dart/source/detail?r=16782

[siva, 2013-01-08 02:18:01]

lgtm

https://codereview.chromium.org/11778028/diff/1/runtime/vm/heap_profiler.cc
File runtime/vm/heap_profiler.cc (right):

https://codereview.chromium.org/11778028/diff/1/runtime/vm/heap_profiler.cc#n...
runtime/vm/heap_profiler.cc:599: sub.Write32(num_instance_fields *
kObjectIdSize);
I am assuming that Write32 will truncate the value beyond 32 bits do we should
probably assert here that
ASSERT((num_instance_fields * kObjectSize) < kMaxInt32);

[siva, 2013-01-08 02:18:02]

lgtm

[cshapiro, 2013-01-08 02:53:04]

https://codereview.chromium.org/11778028/diff/1/runtime/vm/heap_profiler.cc
File runtime/vm/heap_profiler.cc (right):

https://codereview.chromium.org/11778028/diff/1/runtime/vm/heap_profiler.cc#n...
runtime/vm/heap_profiler.cc:599: sub.Write32(num_instance_fields *
kObjectIdSize);
That will not quite work :-)

Write32 takes a uint32_t argument so kMaxInt32 needs to be a kMaxUint32 and the
check needs to be <= not < as above.  It is, of course, legal for the value to
range up to and including the limit.

Also, it is always true that an uint32_t is going to be less than 0xFFFFFFFF so
to make this check work the multiply needs to be done in higher precision.

Anyway, I fixed the code to do that.  Presumably all we care about is having the
product be in range so I did not bother with the possibility of overflow above
where the fields are counted.