nradov / abnffuzzer Goto Github PK

Fuzzer only recognizes rules with 3 children (rule = elements). But RFC 5234 allows in paragraph 3.3 Incremental Alternatives (rule =/ elements).

The Fuzzer class can currently only parse ABNF rules from files that contain only those rules as per the ANTLR grammar. So users who want to test implementations of IETF RFCs first have to manually copy and paste the rules into a separate file.
It would improve usability if this library could take the URI for an RFC as an argument and automatically extract the ABNF rules out of the entire document. However this appears to be a difficult problem in pattern recognition; it's hard to tell whether a particular line contains an ABNF rule, narrative text, or part of the header / footer. The bap aex tool may offer a starting point, however it doesn't appear to produce correct output for some RFCs.

Seeing the following issue with https://tools.ietf.org/html/rfc7950

public class YangTest {

    /**
     * Test the {@link UUID#fromString(String)} method.
     * 
     * @throws IOException
     *             if an error occurs while reading a resource
     */
    @Test
    public void testFromFile() throws IOException {

        final Fuzzer f = new Fuzzer(
                YangTest.class.getResourceAsStream("/rfc7950.yang"));
        
    }

}

Results in

Results :

Tests in error:
  testFromFile(com.github.nradov.abnffuzzer.YangTest): No enum constant com.github.nradov.abnffuzzer.Terminal.Radix.s

Tests run: 11, Failures: 0, Errors: 1, Skipped: 0

Could you clarify why its failing please? Included the abnf in attached zip.

rfc7950.yang.zip