Add validation to CI

[palemieux]

No description provided.

[himorin]

I personally would propose to have action repository within w3c organization for (far) future maintenance and continuity. I'm happy to setup tools repository, if needed.

[nigelmegitt]

I personally would propose to have action repository within w3c organization for (far) future maintenance and continuity. I'm happy to setup tools repository, if needed.

Which action do you mean? This could be a very useful thing, if it's reusable generally.

[himorin]

I personally would propose to have action repository within w3c organization for (far) future maintenance and continuity. I'm happy to setup tools repository, if needed.

Which action do you mean? This could be a very useful thing, if it's reusable generally.

ah, actually was not action, but toolkit used for main part of this CI
https://github.com/w3c/imsc-tests/pull/113/changes#diff-a5c0f08bc43f53a0451cad4a4a2e435ce55ad34183350dfae5e4b2c7dca6360bR16

I haven't checked well for which part of toolkit is used for validation, but I somehow afraid that PR would be failed with unsupported repository or something..
right after writing until here, I've noticed that target tool should be from Glenn,,, so should be fine just to use??

[nigelmegitt]

I've noticed that target tool should be from Glenn,,, so should be fine just to use??

Yes, unless @skynavga has any reasons not to use ttv to validate the files, I think it should be fine.

[nigelmegitt]

Noting that the ttv model for imsc1 doesn't seem to accept the value in the ttp:profile attribute - seems like there might be an issue with the verifier itself there.

Also noting that we expect all the test TTML files to be valid, because they're all presentation tests, i.e. there are (intentionally) no tests that check a validator catches invalid content.

From a tests perspective, ttp:profile is not required in IMSC 1.0.1 documents, and is actually prohibited in EBU-TT-D documents, and many of the failing files are testing features that we would expect to work in EBU-TT-D.

So one way to address the failing validation for a subset of files is to remove ttp:profile from the sources; however that doesn't solve it for other cases, and the validator should accept the parameter.

It's possible that there are other validation approaches we could take that would be less semantically discriminating but still catch syntactic errors, e.g. include a script that does XSD validation based on the schema in the spec at the relevant release tag, or find an alternative validator.