[Test][SYCL] Add test to verify SPIR kernel argument names

[elizabethandrews]

This is a follow-up to PR#5772. This test verifies that
user specified names are retained in SPIR kernel argument.
Please note that the names were already retained (prior to
PR#5772) when kernel was specified as a functor. PR#5722 added
code to ensure the same behavior when kernel is specified using
lambda.

Signed-off-by: Elizabeth Andrews [email protected]

[elizabethandrews]

When working on test I realized compiler generated name can be improved for base classes as well. At the moment, they all have same name _arg__base. I will work on it as a follow-up PR.

[smanna12]

LGTM

[keryell]

Why adding a test would be a [NFC]?

[bader]

Why adding a test would be a [NFC]?

I guess there might be different interpretations of what "functional change" means.

• If it means only patches changing compiler tools behavior, the test changes are all non-functional.
• If it means patches changing compiler tools validation results, then adding new tests can be considered as a functional change.

I think I've seen some folks are using [Test] tag for such patches, which is more specific than [NFC].

[keryell]

* If it means patches changing compiler tools validation results, then adding new tests can be considered as a functional change.

Yes, adding a test add a new function to the global testing infrastructure.

[bader]

Changed [NFC] to [Test]. @keryell, are you okay with that change?

[keryell]

Changed [NFC] to [Test]. @keryell, are you okay with that change?

I am very supportive to this great improvement to the quality of this project ! :-)

[elizabethandrews]

Why adding a test would be a [NFC]?

I guess there might be different interpretations of what "functional change" means.

• If it means only patches changing compiler tools behavior, the test changes are all non-functional.
• If it means patches changing compiler tools validation results, then adding new tests can be considered as a functional change.

I think I've seen some folks are using [Test] tag for such patches, which is more specific than [NFC].

Thanks for the information. I didn't know we were supposed to use [Test] tag for tests. I used [NFC] because PR didn't change existing compiler behavior