full: improve error handling and error messages.

[ysiraichi]

This PR refactors the tensor_methods::full and tensor_methods::full_symint implementation by improving their error message, and returning a status type value.

Key Changes:

• Make tensor_methods::full and tensor_methods::full_symint return StatusOr<absl_nonnull XLATensorPtr>
• Improve error message on incompatible tensor shapes

Example:

torch.full((2, -2, 2), 1.5, device="xla")

Before:

Traceback (most recent call last):
  File "examples.py", line 5, in <module>
    torch.full((2, -2, 2), 1.5, device="xla")
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: Check failed: std::all_of(size.begin(), size.end(), [](int64_t dim) { return dim >= 0; }): Dimensions cannot be negative numbers (at torch_xla/csrc/tensor_methods.cpp:208)

Exception raised from operator& at torch_xla/csrc/runtime/tf_logging.cpp:26 (most recent call first):

After:

Traceback (most recent call last):
  File "example.py", line 5, in <module>
    torch.full((2, -2, 2), 1.5, device="xla")
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: full(): expected concrete sizes (i.e. non-symbolic) to be positive values, however found negative ones: [2, -2, 2]

Status Propagation Trace:
    From: CheckSizesArePositiveImpl at torch_xla/csrc/tensor_methods.cpp:427 (error: full(): expected concrete sizes (i.e. non-symbolic) to be positive values, however found negative ones: [2, -2, 2])
    From: full at torch_xla/csrc/tensor_methods.cpp:1818

Exception raised from MaybeThrow at torch_xla/csrc/status.cpp:121 (most recent call first):

[zhanyong-wan]

Thanks!

[ysiraichi]

The CI failure is unrelated to this PR. I'm landing this.

[ysiraichi]

Couldn't merge this PR because of the failing CI.
Tried merging it using the command line, but that also didn't work, since the master branch is protected.
Waiting for the issue to be resolved.