Allow equality operation on mixed double/smi arguments.

Allow equality operation on mixed double/smi arguments.
Fix a bug in relational operations in 64 bit mode: if ICData specifies double and Smi as possible arguments, we generate code for double and unbox or convert to double the inputs. These works only if smi can fit into the double. Current solution for 64-bit architecture is to disallow two smi-s as input to a polymorphic comparison instruction (equality, relational).

R=johnmccutchan@google.com

Committed: https://code.google.com/p/dart/source/detail?r=25522

[Cutch, 2013-07-25 21:03:35]

https://codereview.chromium.org/20468002/diff/12001/runtime/vm/flow_graph_opt...
File runtime/vm/flow_graph_optimizer.cc (right):

https://codereview.chromium.org/20468002/diff/12001/runtime/vm/flow_graph_opt...
runtime/vm/flow_graph_optimizer.cc:2665: const ICData& ic_data =
*comp->ic_data();
To match HandleEqualityCompare why don't you add:

ASSERT(ic_data.num_args_tested() == 2);
ASSERT(comp->operation_cid() == kIllegalCid);

https://codereview.chromium.org/20468002/diff/12001/runtime/vm/flow_graph_opt...
runtime/vm/flow_graph_optimizer.cc:2785: ASSERT(ic_data.num_args_tested() == 2);
Can this duplicate code in HandleEqualityCompare and HandleRelationalOp be moved
into a common function?

[srdjan, 2013-07-25 22:17:01]

https://codereview.chromium.org/20468002/diff/12001/runtime/vm/flow_graph_opt...
File runtime/vm/flow_graph_optimizer.cc (right):

https://codereview.chromium.org/20468002/diff/12001/runtime/vm/flow_graph_opt...
runtime/vm/flow_graph_optimizer.cc:2665: const ICData& ic_data =
*comp->ic_data();
On 2013/07/25 21:03:36, Cutch wrote:
> To match HandleEqualityCompare why don't you add:
> 
> ASSERT(ic_data.num_args_tested() == 2);
> ASSERT(comp->operation_cid() == kIllegalCid);

Factored out code. Done.

https://codereview.chromium.org/20468002/diff/12001/runtime/vm/flow_graph_opt...
runtime/vm/flow_graph_optimizer.cc:2785: ASSERT(ic_data.num_args_tested() == 2);
On 2013/07/25 21:03:36, Cutch wrote:
> Can this duplicate code in HandleEqualityCompare and HandleRelationalOp be
moved
> into a common function?

Done.

[Cutch, 2013-07-26 16:02:36]

lgtm

[srdjan, 2013-07-26 16:48:53]

Committed patchset #5 manually as r25522 (presubmit successful).