eof: Enable compilation to EOF for all EVMVersionRestrictedTestCase tests by default

[rodiazet]

Depends on:
#15818 Merged
#15657 Merged
#15658 Merged
#15659 Merged
#15660 Merged
#15661 Merged
#15662 Merged
#15663 Merged
#15665 Merged

[github-actions]

Thank you for your contribution to the Solidity compiler! A team member will follow up shortly.

If you haven't read our contributing guidelines and our review checklist before, please do it now, this makes the reviewing process and accepting your contribution smoother.

If you have any questions or need our help, feel free to post them in the PR or talk to us directly on the #solidity-dev channel on Matrix.

[cameel]

Duplicating compileViaYul here may be a bad idea. This makes the setting available even in tests where it does not make sense. And, even worse, in some where it would make sense but they don't check it so it has no effect. And in those that define it later a conflicting definition here might have some weird side-effects. On top of that it will be just confusing - here EOF depends on compileViaYul, but in SyntaxTest we already have a default for compileViaYul that depends on EOF.

You need this change only because of semantic tests, right? It's only there that it defaults to also. If that's the case then a better solution would be to have the default in SemanticTest be dependent on EOF, just like we already did in SyntaxTest. Make it also in non-EOF runs and true in EOF runs.

[rodiazet]

Fix. I also not run a test if compileViaYul flag is set to false and command option enables EOF

[rodiazet]

But doing this I found one problem with test configuration check. We don't verify that a test settings are correct. In case where compileViaYul: fasle and 'bytecodeFormat: >=EOFv1' the test never runs and we don't get any error.

[cameel]

We don't verify that a test settings are correct.

True. We should have checks against this. Or use TestCaseReader::enumSetting() instead of stringSetting().

Can you refactor then into enum settings (in a separate PR)?

[cameel]

Wait, but both semantic and syntax test have compileViaYulAllowedValues and check the value against it. Doesn't it catch the error?

[cameel]

I also not run a test if compileViaYul flag is set to false and command option enables EOF

Sounds fine, but please put this in a separate commit. It's a separate fix.

[rodiazet]

I mean something different by verifying a test settings. We do verify the values separately but we should also check that i.e. compileViaYul: false then bytecodeFormat cannot be >=EOFv1. Every test like this never runs.

[cameel]

Ah, we don't verify that they conflict. True, we should add a check against that.

Do you want to do it here, of should I merge already?

[cameel]

Oh, I see #15819. I guess you can do it there.

[cameel]

All dependencies merged. Please rebase for a final check and we're done!