FindNonUTF8Char.groovy

import groovy.transform.*
import java.nio.*
import java.nio.charset.*

@Field CharsetDecoder cs = Charset.forName('UTF-8').newDecoder()
InputStream.metaClass.eachChunk << { int preferredChunkSize, Closure closure ->
    delegate.eachByte(preferredChunkSize) { buffer, bytesRead ->
        if (bytesRead == preferredChunkSize) {
            closure(buffer)
        } else if (bytesRead > 0) {
            byte[] data = new byte[bytesRead]
            System.arraycopy(buffer, 0, data, 0, bytesRead)
            closure(data)
        }
    }
}
int pad = 10
String all = new File(/C:\dev\cp\PRPS\admin\wsdl_all.txt/).text
new File(/C:\dev\cp\PRPS\admin\wsdl_all_fixed)2.txt/) << all
int counter = 0
new File(/C:\dev\cp\PRPS\admin\wsdl_all_fixed.txt/).withInputStream { is ->
    is.eachChunk(2) { bytes ->
        if (!isValidUTF8(bytes)) {
           println "${new String(bytes, 'windows-1252')} is an illegal character at index $counter: ${all[(counter - pad)..(counter + pad)]}"       
        }
        counter += 1
    }
}

boolean isValidUTF8( byte[] input ) {
    try {
        cs.decode(ByteBuffer.wrap(input));
        return true;
    } catch(CharacterCodingException e){
        return false;
    }       
}