soltest pass for EOF

[cameel]

We already have some tests for EOF, but the whole test suite doesn't pass yet, so we were not running them in CI. This often results in failing tests not being noticed in EOF PRs.

This PR adds a soltest pass with --eof-version 1 in CI but with a temporary blacklist for tests that are not expected to pass yet.

Note that syntax and semantic tests are not blacklisted. They all pass because for them (and any other tests based on EVMVersionRestrictedTestCase) EOF is excluded unless explicitly requested via EOFVersion: >=EOFv1 setting. This will allow us to enable them on a case-by-case basis for now and only when all of them are ready we'll flip the default to include EOF.

[r0qs]

LGTM, but have you tested with EOF version 1?

[r0qs]

This change actually made me wonder if we should also consider changing OPTIMIZE and ABI_ENCODER_V1 to integers, even though they are used as boolean values.

[cameel]

Would not hurt. The variables should also be quoted. And we should validate the values - currently anything other than 1 means "optimizer disabled", even "true".

I'm going to leave it as is here though. There were just too many small things like this in the script that could be improved and I didn't want to get sidetracked by cleaning it up :)

[r0qs]

LGTM, but have you tested with EOF version 1?

Oh, my mistake, I see it in the EOF_VERSIONS array.

[cameel]

Yeah, it's now running for EOF v1 too. That was the whole point of the PR :)