[SPIR-V][vk::SampledTexture] #11. Support vk::SampledTexture1D and vk::SampledTexture1DArray type.

[luciechoi]

Part of #7979

The function definitions are equivalent to that of Texture1D counterparts, just without Sampler argument.

[luciechoi]

(@Keenuts @s-perron next SampledTexture PR for review)

[Keenuts]

Overall looks good, but many overloads are not tested. I stopped checking after a few, but overall we should have at least one test for each overload.

[luciechoi]

Overall looks good, but many overloads are not tested. I stopped checking after a few, but overall we should have at least one test for each overload.

Hmmm do you mean every function variant should have a test..? The overloads are tested with SampledTexture2D type, and adding a 1D type doesn't require code changes.

I took a look at adding these, but noticed that the existing test coverage for the other texture types (like Texture2D, Texture2DMS, and their array counterparts) also doesn't exhaustively test every single overload.

To keep the scope of this PR manageable and maintain parity with regular Texture types, I've mirrored the existing level of coverage for the 1D types. Let me know if there's a specific overload variant you are concerned about and I'd be happy to add a test just for that one.

[Keenuts]

Hmmm do you mean every function variant should have a test..? The overloads are tested with SampledTexture2D type, and adding a 1D type doesn't require code changes.

If something is not tested, something can be removed accidentally without us noticing.
That's why some projects implement mutation testing or similar.
The test does not have to be complex, but we should make sure each overload is at least present through a test.

[s-perron]

I agree that we need to haves testing for all for all of the new functions, even if the code generating them is the same. I think you are working on that.

We can also improve some string checking, which can be slow.

[luciechoi]

@Keenuts @s-perron

I've added the tests for all function overloads, PTAL!

[s-perron]

I think we should try to make the string matching good. I don't think it is hard to read the code if you use startswith for an initial check, drop the prefix, and then check the rest.

[luciechoi]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[Keenuts]

Looks overall OK, some code nits, and a more generic remark on tests.

[Keenuts]

Seems like if we do name.drop_front(StringLiteral("Sampledtexture").size()) we get a compilation-time 14 which is self-explanatory.

[luciechoi]

Sounds good to get the text size as a compile time constant.