Commit 0626ebdb authored by Ari Trachtenberg's avatar Ari Trachtenberg
Browse files

broadened file check to look only at printable ASCII

parent ab055aca

src/edu/bu/ec504/hw1p3/Tester.java

+14 −2
Original line number Diff line number Diff line
@@ -99,8 +99,8 @@ public class Tester { 
        // ... also based on http://www.exampledepot.com/egs/java.io/CopyFile.html

        String sf1, sf2;

        try {

            sf1 = readFile(file1); sf1 = sf1.replaceAll("[\n\r]","");

            sf2 = readFile(file2); sf2 = sf2.replaceAll("[\n\r]","");

            sf1 = cleanString(readFile(file1));

            sf2 = cleanString(readFile(file2));

        }

        catch (IOException e) {return false;} // any exceptions are treated as a lack of a match
@@ -108,6 +108,18 @@ public class Tester { 


    }



    /**

     * Removes all non-printable characters from a string.

     * Based on:  https://howtodoinjava.com/regex/java-clean-ascii-text-non-printable-chars/

     * @param str

     */

    private static String cleanString(String str) {

        str=str.replaceAll("[^\\x00-\\x7F]", ""); // non-ASCII

        str = str.replaceAll("[\\p{Cntrl}&&[^\r\n\t]]", ""); // ASCII control characters - see http://www.unicode.org/reports/tr18/#Compatibility_Properties

        str = str.replaceAll("\\p{C}", ""); // non-printable characters - see http://www.unicode.org/reports/tr18/#General_Category_Property

        return str;

    }



    /**

     * Reads a file into a string

     * @param file1 the file to read