Examples of org.apache.oro.text.regex.PatternMatcherInput

org.apache.oro.text.regex.PatternMatcherInput
orm whatever processing on the result you want. } // Suppose we want to start searching from the beginning again with // a different pattern. // Just set the current offset to the begin offset. input.setCurrentOffset(input.getBeginOffset()); // Second search omitted // Suppose we're done with this input, but want to search another string. // There's no need to create another PatternMatcherInput instance. // We can just use the setInput() method. input.setInput(aNewInputString); @version @version@ @since 1.0 @see PatternMatcher

  public ArrayList<String> extractHrefs() {


    ArrayList<String> aHrefs = new ArrayList<String>();
  try {
      if (null==oFullHref) oFullHref = oCompiler.compile("<a ((accesskey|charset|class|coords|dir|hreflang|id|lang|name|rel|rev|shape|style|tabindex|target|title)\\s*=\\s*[\"']?([^'\"\\r\\n]+)[\"']?)* href\\s*=\\s*[\"']?([^'\"\\r\\n]+)[\"']?", Perl5Compiler.CASE_INSENSITIVE_MASK);
      PatternMatcherInput oPinpt = new PatternMatcherInput(sBody);
      while (oMatcher.contains(oPinpt, oFullHref)) {
      aHrefs.add(oMatcher.getMatch().group(4));
      if (oPinpt.endOfInput()) break;
      } // wend
    } catch (MalformedPatternException neverthrown) { } 
    return aHrefs;
  } // extractHrefs()

View Full Code Here

   */
  public ArrayList<String> extractLocalUrls()
    throws ArrayIndexOutOfBoundsException {


    ArrayList<String> aLocalUrls = new ArrayList<String>();
    PatternMatcherInput oPinpt;
    String sSrcUrl, sHrefUrl;
    
    try {
      synchronized(oCompiler) {
        if (null==oFullHref) oFullHref = oCompiler.compile("<a( (accesskey|charset|class|coords|dir|hreflang|id|lang|name|rel|rev|shape|style|tabindex|target|title)\\s*=\\s*[\"']?([^'\"\\r\\n]+)[\"']?)* href\\s*=\\s*[\"']?([^'\"\\r\\n]+)[\"']?", Perl5Compiler.CASE_INSENSITIVE_MASK);
        if (null==oGoodHref) oGoodHref = oCompiler.compile("<a( (accesskey|charset|class|coords|dir|hreflang|id|lang|name|rel|rev|shape|style|tabindex|target|title)\\s*=\\s*[\"']?([^'\"\\r\\n]+)[\"']?)* href\\s*=\\s*[\"']?(http://|https://|mailto:)\\w+([^'\"\\r\\n]+)[\"']?", Perl5Compiler.CASE_INSENSITIVE_MASK);
        if (null==oHostHref) oHostHref = oCompiler.compile("<a( (accesskey|charset|class|coords|dir|hreflang|id|lang|name|rel|rev|shape|style|tabindex|target|title)\\s*=\\s*[\"']?([^'\"\\r\\n]+)[\"']?)* href\\s*=\\s*[\"']?(http://|https://)localhost([^'\"\\r\\n]+)[\"']?", Perl5Compiler.CASE_INSENSITIVE_MASK);
        if (null==oFullSrc) oFullSrc = oCompiler.compile("<img( (align|alt|border|class|dir|height|hspace|id|ismap|lang|longdesc|style|title|usemap|vspace|width)\\s*=\\s*[\"']?([^'\"\\r\\n]+)[\"']?)* src\\s*=\\s*[\"']?([^'\"\\r\\n]+)[\"']?", Perl5Compiler.CASE_INSENSITIVE_MASK);
      if (null==oGoodSrc) oGoodSrc = oCompiler.compile("<img( (align|alt|border|class|dir|height|hspace|id|ismap|lang|longdesc|style|title|usemap|vspace|width)\\s*=\\s*[\"']?([^'\"\\r\\n]+)[\"']?)* src\\s*=\\s*[\"']?(cid:|http://|https://)([^'\"\\r\\n]+)[\"']?", Perl5Compiler.CASE_INSENSITIVE_MASK);
      if (null==oHostSrc) oHostSrc = oCompiler.compile("<img( (align|alt|border|class|dir|height|hspace|id|ismap|lang|longdesc|style|title|usemap|vspace|width)\\s*=\\s*[\"']?([^'\"\\r\\n]+)[\"']?)* src\\s*=\\s*[\"']?(http://localhost|https://localhost)([^'\"\\r\\n]+)[\"']?", Perl5Compiler.CASE_INSENSITIVE_MASK);
      }


      oPinpt = new PatternMatcherInput(sBody);
      while (oMatcher.contains(oPinpt, oFullSrc)) {
      sSrcUrl = oMatcher.getMatch().toString();
      if (!oMatcher.matches(sSrcUrl, oGoodSrc))
      aLocalUrls.add(sSrcUrl);
      if (oPinpt.endOfInput()) break;
      } // wend


      oPinpt.setCurrentOffset(oPinpt.getBeginOffset());
      while (oMatcher.contains(oPinpt, oFullSrc)) {
      sSrcUrl = oMatcher.getMatch().toString();
    if (oMatcher.matches(sSrcUrl, oHostSrc))
      aLocalUrls.add(sSrcUrl);
      if (oPinpt.endOfInput()) break;
      } // wend
      
      oPinpt.setCurrentOffset(oPinpt.getBeginOffset());
      while (oMatcher.contains(oPinpt, oFullHref)) {
    sHrefUrl = oMatcher.getMatch().toString();
      if (!oMatcher.matches(sHrefUrl, oGoodHref))
      aLocalUrls.add(sHrefUrl);
      if (oPinpt.endOfInput()) break;
      } // wend


      oPinpt.setCurrentOffset(oPinpt.getBeginOffset());
      while (oMatcher.contains(oPinpt, oFullHref)) {
    sHrefUrl = oMatcher.getMatch().toString();
    if (oMatcher.matches(sHrefUrl, oHostHref))
      aLocalUrls.add(sHrefUrl);    
      if (oPinpt.endOfInput()) break;
      } // wend
    } catch (MalformedPatternException neverthrown) { } 


  return aLocalUrls;
  } // extractLocalUrls

View Full Code Here

  


  protected Zone accept(String pathinfo, int start)
  {
    if (_pattern != null) {
      PatternMatcherInput input = new PatternMatcherInput(pathinfo, start, pathinfo.length()-start);
      Perl5Matcher matcher = new Perl5Matcher();
      if (matcher.contains(input, _pattern)) {
        return resolveZone(pathinfo, start);
      }
    } else {

View Full Code Here

    private int _index = 0;


    public StringPatternMatcher(Pattern pattern, String string)
    {
      _pattern = pattern;
      _input = new PatternMatcherInput(string);
    }

View Full Code Here

    if (oMatcher.contains(sHtml, oPattern)) {
      sBaseHref = Gadgets.chomp(oMatcher.getMatch().group(3),"/");
      if (DebugFile.trace) DebugFile.writeln("<base href="+sBaseHref+">");
    }
  
      PatternMatcherInput oMatchInput = new PatternMatcherInput(sHtml);
    oPattern = oCompiler.compile("\\x20(src=|background=|background-image:url\\x28)(\"|')?([^'\"\\r\\n]+)(\"|')?(\\x20|\\x29|/|>)", Perl5Compiler.CASE_INSENSITIVE_MASK);
    StringSubstitution oSrcSubs = new StringSubstitution();
    int nMatches = 0;
    while (oMatcher.contains(oMatchInput, oPattern)) {
      nMatches++;
      String sMatch = oMatcher.getMatch().toString();
      String sAttr = oMatcher.getMatch().group(1);
      String sQuo = oMatcher.getMatch().group(2);
      if (sQuo==null) sQuo = "";
      String sSrc = oMatcher.getMatch().group(3);
      if (DebugFile.trace) DebugFile.writeln("Source file found at "+sSrc);
      String sEnd = oMatcher.getMatch().group(5);
      if (!oFiles.contains(sSrc)) oFiles.add(sSrc);
      String sFilename = sSrc.substring(sSrc.replace('\\','/').lastIndexOf('/')+1);
      if (DebugFile.trace)
        DebugFile.writeln("StringSubstitution.setSubstitution("+sMatch+" replace with "+sMatch.substring(0,sAttr.length()+1)+sQuo+sFilename+sQuo+sEnd+")");
      oSrcSubs.setSubstitution(sMatch.substring(0,sAttr.length()+1)+sQuo+sFilename+sQuo+sEnd);
        sHtml = Util.substitute(oReplacer, oCompiler.compile(sMatch), oSrcSubs, sHtml, Util.SUBSTITUTE_ALL);
    } //wend
    
    oMatchInput = new PatternMatcherInput(sHtml);
    oPattern = oCompiler.compile("<link\\x20+(rel=(\"|')?stylesheet(\"|')?\\x20+)?(type=(\"|')?text/css(\"|')?\\x20+)?href=(\"|')?([^'\"\\r\\n]+)(\"|')?");
    while (oMatcher.contains(oMatchInput, oPattern)) {
      nMatches++;
      String sMatch = oMatcher.getMatch().toString();
      String sSrc = oMatcher.getMatch().group(8);

View Full Code Here

        Perl5Matcher localMatcher = JMeterUtils.getMatcher();
        // The headers and body are divided by a blank line
        String regularExpression = "^.$"; 
        Pattern pattern = JMeterUtils.getPattern(regularExpression, Perl5Compiler.READ_ONLY_MASK | Perl5Compiler.CASE_INSENSITIVE_MASK | Perl5Compiler.MULTILINE_MASK);
        
        PatternMatcherInput input = new PatternMatcherInput(stringToCheck);
        while(localMatcher.contains(input, pattern)) {
            MatchResult match = localMatcher.getMatch();
            return match.beginOffset(0);
        }
        // No divider was found

View Full Code Here

        }


        List<MatchResult> collectAllMatches = new ArrayList<MatchResult>();
        try {
            PatternMatcher matcher = JMeterUtils.getMatcher();
            PatternMatcherInput input = new PatternMatcherInput(textToMatch);
            while (matcher.contains(input, searchPattern)) {
                MatchResult match = matcher.getMatch();
                collectAllMatches.add(match);
            }
        } catch (NumberFormatException e) {//TODO: can this occur?

View Full Code Here

        List<String> pieces = new ArrayList<String>();
        // String or Integer
        List<Object> combined = new LinkedList<Object>();
        PatternMatcher matcher = JMeterUtils.getMatcher();
        Util.split(pieces, matcher, templatePattern, rawTemplate);
        PatternMatcherInput input = new PatternMatcherInput(rawTemplate);
        boolean startsWith = isFirstElementGroup(rawTemplate);
        if (startsWith) {
            pieces.remove(0);// Remove initial empty entry
        }
        Iterator<String> iter = pieces.iterator();

View Full Code Here

    }


    private String process(String textToParse) {


        Perl5Matcher matcher = new Perl5Matcher();
        PatternMatcherInput input = new PatternMatcherInput(textToParse);


        PatternCacheLRU pcLRU = new PatternCacheLRU();
        Pattern pattern = pcLRU.getPattern(regexpField.getText(), Perl5Compiler.READ_ONLY_MASK);
        List<MatchResult> matches = new LinkedList<MatchResult>();
        while (matcher.contains(input, pattern)) {

View Full Code Here

     */
    @Override
    public Iterator<URL> getEmbeddedResourceURLs(byte[] html, URL baseUrl, URLCollection urls) {


        Perl5Matcher matcher = JMeterUtils.getMatcher();
        PatternMatcherInput input = localInput.get();
        // TODO: find a way to avoid the cost of creating a String here --
        // probably a new PatternMatcherInput working on a byte[] would do
        // better.
        input.setInput(new String(html)); // TODO - charset?
        Pattern pattern=JMeterUtils.getPatternCache().getPattern(
                REGEXP,
                Perl5Compiler.CASE_INSENSITIVE_MASK
                | Perl5Compiler.SINGLELINE_MASK
                | Perl5Compiler.READ_ONLY_MASK);

View Full Code Here

0 1 2 3 4 5 6 7 8 9

TOP

Related Classes of org.apache.oro.text.regex.PatternMatcherInput

anvil.core.AnyUtils$StringPatternMatcher

anvil.server.Zone

com.alibaba.toolkit.util.regex.MatchContext

com.alibaba.toolkit.util.regex.MatchItem

com.knowgate.dfs.FileSystem

com.knowgate.hipermail.HtmlMimeBodyPart

jmt.engine.jwat.input.AllInputLoader

jmt.engine.jwat.input.IntervalInputLoader

jmt.engine.jwat.input.RndInputLoader

org.apache.ddlutils.platform.TestMSSqlPlatform

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.