THRIFT-5910: Add UUID support in Ruby

[kpumuk]

Currently in draft state until CI fixes are merged in my other pull requests.

This PR adds UUID support to Ruby. The UUID is stored as a string, as this is a common format returned by SecureRandom.uuid in Ruby. This adds a little bit overhead on writing, as the format of the UUID is verified there.

• Did you create an Apache Jira ticket? THRIFT-5910
• If a ticket exists: Does your pull request title follow the pattern "THRIFT-NNNN: describe my issue"?
• Did you squash your changes to a single commit? (not required, but preferred)
• Did you do your best to avoid breaking changes? If one was needed, did you label the Jira ticket with "Breaking-Change"?
• If your change does not involve any code, include [skip ci] anywhere in the commit message to free up build resources.

[CJCombrink]

Please look at my findings here #3144 to ensure that when the ruby code (your implementation) sends a UUID to the .NET client/server that the values match on the other side. Not an echo test since that only ensures that the client and server has the same understanding, this must specifically be one implementation sending it and the other implementation correctly extracting it

There still is no unit tests to confirm this but it would be great if you can do a manual check to ensure that it corresponds to ensure there is a common understanding of the underlying data layout.

[kpumuk]

I did a full cross-test suite verification with C++, Java, Rust, and Go. Will review your findings as well

Edit: I see what you mean, will check what needs to be updated on Ruby cross-tests side. Thank you!

Edit2: And you were right:

  1) BinaryProtocol it should behave like a binary protocol should conform protocol
     Failure/Error: expect(uuid).to eq("00112233-4455-6677-8899-aabbccddeeff")

       expected: #<Encoding:ASCII-8BIT> "00112233-4455-6677-8899-aabbccddeeff"
            got: #<Encoding:UTF-8> "8899aabb-ccdd-eeff-0011-223344556677"

Great finding, thank you!

[CJCombrink]

Unfortunately it is a manual test, run the test and manually check the output of the client and server.

And assume the .net implementation is correct

[kpumuk]

I added unit tests that match protocol conformity based on @Jens-G 's 3e6be73

[kpumuk]

Found another inconsistency while testing constants THRIFT-5911:

typedef uuid myUUID

struct thing {
  1: i32  hello,
  2: i32  goodbye
  3: uuid id
  4: myUUID my_id
  5: optional myUUID my_optional_id
}

// uuids are accepted with or without curly braces
const uuid GEN_UUID =  '00000000-4444-CCCC-ffff-0123456789ab'
const uuid GEN_GUID = '{00112233-4455-6677-8899-aaBBccDDeeFF}'

const myUUID MY_UUID =  '00000000-4444-CCCC-ffff-0123456789ab'
const myUUID MY_GUID = '{00112233-4455-6677-8899-aaBBccDDeeFF}'

produces

  GEN_UUID = %q"00000000-4444-CCCC-ffff-0123456789ab"

  GEN_GUID = %q"00112233-4455-6677-8899-aaBBccDDeeFF"

  MY_UUID = %q"00000000-4444-CCCC-ffff-0123456789ab"

  MY_GUID = %q"{00112233-4455-6677-8899-aaBBccDDeeFF}"

Seems like a bit of an issue with the compiler.

[kpumuk]

Seems like a bit of an issue with the compiler.

This has been addressed in #3256