Extended support of various special chars

[AlexeyKuznetsov-DD]

It is not a feature request, more like a brainstorm, I have no ideas how it can looks like to be human readable,
but with my first PR (see method testMappingToJsonStringArguments) where I tried to use this lib I faced the test with a lot of special chars, and I have feeling that it should be eligible for conversion into table test.

Any thought on this?

P.S. PR with how it was converted from Spock to Java without this lib.
P.S.S. my PR when I applied @TableTest

[nchaugen]

What kind of special chars are we talking about?

[AlexeyKuznetsov-DD]

So far, I want to translate Groovy Spock to @TableTest:

def "test mapping to JSON string: input"() {
    when:
    String escaped = JsonMapper.toJson((String) string)

    then:
    escaped == expected

    where:
    string                   | expected
    null                     | ""
    ""                       | ""
    ((char) 4096).toString() | '"\\u1000"'
    ((char) 256).toString()  | '"\\u0100"'
    ((char) 128).toString()  | '"\\u0080"'
    "\b"                     | '"\\b"'
    "\t"                     | '"\\t"'
    "\n"                     | '"\\n"'
    "\f"                     | '"\\f"'
    "\r"                     | '"\\r"'
    '"'                      | '"\\\""'
    '/'                      | '"\\/"'
    '\\'                     | '"\\\\"'
    "a"                      | '"a"'
  }

[nchaugen]

I see. You can use a @TypeConverter for this. Something like this:

public record EscapedString(String value) {

    @TypeConverter
    public static EscapedString parseEscapedString(String value) {
        if (value == null) return new EscapedString(null);
        if (value.isEmpty()) return new EscapedString(value);
        if (value.startsWith("\\u") && value.length() == 6) {
            return new EscapedString(
                String.valueOf((char) Integer.parseInt(value.substring(2), 16))
            );
        }
        String unescaped = switch (value) {
            case "\\b" -> "\b";
            case "\\t" -> "\t";
            case "\\n" -> "\n";
            case "\\f" -> "\f";
            case "\\r" -> "\r";
            case "\\\\" -> "\\";
            default -> value;
        };
        return new EscapedString(unescaped);
    }
}

The test would then become:

@TypeConverterSources(EscapedString.class)
public class StringToJsonTest {
    
    @TableTest("""
          Scenario         | string   | Json?
          null input       |          | ''
          empty string     | ''       | '""'
          unicode 4096     | \\u1000  | '"\\u1000"'
          unicode 256      | \\u0100  | '"\\u0100"'
          unicode 128      | \\u0080  | '"\\u0080"'
          backspace        | \\b      | '"\\b"'
          tab              | \\t      | '"\\t"'
          newline          | \\n      | '"\\n"'
          form feed        | \\f      | '"\\f"'
          carriage return  | \\r      | '"\\r"'
          double quote     | '"'      | '"\\""'
          forward slash    | /        | '"\\/"'
          backslash        | \\\\     | '"\\\\"'
          letter a         | a        | '"a"'
          """)
    void shouldMapStringToJson(EscapedString string, String json) {
        assertEquals(json, JsonMapper.toJson(string.value()));
    }

If you do not need to reuse the EscapedString in other tests you can also make it local to the test class. Then you do not need the @TypeConverterSources annotation, but you would have to pull the parseEscapedString type converter out into the test method class.

You can read more about custom converters here: https://tabletest.org/reference/type-conversion/#custom-converter-methods

[AlexeyKuznetsov-DD]

Oh! Thanks for this example!