* within the nested {@link TextNormalizeContext} class.
 *
 * <p>The text normalization process includes:
 * <ul>
 *   <li>Treating ISO control characters and specified space characters as spaces.</li>
 *   <li>Appending alphanumeric characters (0-9, A-Z, a-z) to the buffer.</li>
 *   <li>Appending symbol characters (!-/, :-@, [-`, {-~) to the buffer.</li>

    }

    /**
     * Test name with special characters
     */
    public void test_name_specialCharacters() {
        // Test with spaces
        testTransformer.setName("name with spaces");
        assertEquals("name with spaces", testTransformer.getName());

        // Test with special characters
        testTransformer.setName("name-with_special.chars#123");

    protected int initialBufferSize = 10000;

    /**
     * If true, duplicated terms are replaced.
     */
    protected boolean replaceDuplication = false;

    /**
     * Space characters. Default includes common space characters.
     */
    protected int[] spaceChars = { '\u0020', '\u00a0', '\u3000', '\ufffd' };

    /**
     * Memory size.
     */
    protected int memorySize = 1024 * 1024; //1mb

    }

    /**
     * Returns whether space trimming is enabled.
     * @return True if space trimming is enabled, false otherwise.
     */
    public boolean isTrimSpace() {
        return trimSpaceEnabled;
    }

    /**
     * Sets whether space trimming is enabled.
     * @param trimSpace The trim space flag to set.
     */
    public void setTrimSpace(final boolean trimSpace) {

        value = "ftp://123.123.123.123/path with spaces/file.txt";
        ftpInfo = new FtpClient.FtpInfo(value, Constants.UTF_8);
        assertEquals(value, ftpInfo.toUrl());
        assertEquals("123.123.123.123:21", ftpInfo.getCacheKey());
        assertEquals("123.123.123.123", ftpInfo.getHost());
        assertEquals(21, ftpInfo.getPort());
        assertEquals("/path with spaces", ftpInfo.getParent());

        } catch (IllegalArgumentException e) {
            // Expected
        }

        try {
            CrawlerStatus.valueOf(" RUNNING "); // with spaces
            fail("Should throw IllegalArgumentException for value with spaces");
        } catch (IllegalArgumentException e) {
            // Expected
        }
    }

    /**
     * Test valueOf method with null
     */

        // Unicode characters
        testRule.setRuleId("ルール日本語");
        assertEquals("ルール日本語", testRule.getRuleId());

        // Whitespace
        testRule.setRuleId("  rule with spaces  ");
        assertEquals("  rule with spaces  ", testRule.getRuleId());
    }

    /**
     * Test multiple registrations of same rule
     */
    public void test_multipleRegistrations() {

    /**
     * A map of field rules.
     */
    protected Map<String, String> fieldRuleMap = new LinkedHashMap<>();

    /** a flag to trim a space characters. */
    protected boolean trimSpaceEnabled = true;

    /**
     * The charset name.
     */
    protected String charsetName = Constants.UTF_8;

    /**

        assertEquals("3", extractData.getValues("meta:character-count")[0]);
        assertEquals("じょうたい", extractData.getValues("cp:contentStatus")[0]);
        assertEquals("3", extractData.getValues("meta:character-count-with-spaces")[0]);
        assertEquals("たいとる", extractData.getValues("dc:title")[0]);
        assertEquals("2", extractData.getValues("extended-properties:TotalTime")[0]);

  </mime-type>
  <mime-type type="image/vnd.dxf;format=ascii">
    <sub-class-of type="image/vnd.dxf"/>
    <_comment>AutoCAD DXF in ASCII Text form</_comment>
    <magic priority="50">
      <!-- Variable number of spaces or possibly creator tool before the code groups -->
      <!-- allow 999 (comments) before section -->
      <match type="regex" value="(999\r?\n[^\r\n]{0,64}\\s+)?0\r?\nSECTION\r?\n" offset="0:32">