Speedup #generate_json by using case/when/end

[eregon]

• if/elsif comparing obj.class is significantly slower: https://github.com/ruby/json/pull/668#issuecomment-2450747190
• The only case where an exact class check is needed so far is for String (https://github.com/ruby/json/issues/667).
• Before:

  $ ruby -Ilib:ext benchmark/standalone.rb dump pure JSON::Pure::Generator
  truffleruby 24.2.0-dev-07b978e4, like ruby 3.2.4, Oracle GraalVM JVM [x86_64-linux] Calculating -------------------------------------
    JSON.dump(obj)      6.426k (± 5.9%) i/s  (155.62 μs/i) -     32.395k in   5.064479s
    JSON.dump(obj)      6.380k (± 7.4%) i/s  (156.73 μs/i) -     31.806k in   5.021304s
    JSON.dump(obj)      6.276k (±10.5%) i/s  (159.33 μs/i) -     31.217k in   5.060762s
    JSON.dump(obj)      6.450k (± 7.0%) i/s  (155.05 μs/i) -     32.395k in   5.059538s
    JSON.dump(obj)      6.413k (± 6.2%) i/s  (155.93 μs/i) -     32.395k in   5.081573s

• After:

  $ ruby -Ilib:ext benchmark/standalone.rb dump pure
  JSON::Pure::Generator
  truffleruby 24.2.0-dev-07b978e4, like ruby 3.2.4, Oracle GraalVM JVM [x86_64-linux]
  Calculating -------------------------------------
    JSON.dump(obj)      8.237k (± 5.0%) i/s  (121.41 μs/i) -     41.600k in   5.069507s
    JSON.dump(obj)      8.179k (± 5.1%) i/s  (122.26 μs/i) -     40.768k in   5.002035s
    JSON.dump(obj)      8.147k (± 7.9%) i/s  (122.74 μs/i) -     40.768k in   5.044840s
    JSON.dump(obj)      8.137k (± 6.9%) i/s  (122.90 μs/i) -     40.768k in   5.048690s
    JSON.dump(obj)      8.112k (±10.2%) i/s  (123.27 μs/i) -     39.936k in   5.023502s

[byroot]

• The only case where an exact class check is needed so far is for String

That's the one case where we had a failing test for it, but does that mean we should special case it?

I'm open to say only String subclasses have that behavior, but then all 3 implementations should match behavior. Could you add some tests to make sure we converge them?

Also I'm very surprised .class is this expensive on Truffle, I wouldn't have expected it.

[eregon]

I'll rebase and see if the added tests fail on this PR or if they are fine.

[eregon]

Tests seem to pass after rebase from a local test run.