2 // This software is now distributed according to
3 // the Lesser Gnu Public License. Please see
4 // http://www.gnu.org/copyleft/lesser.txt for
8 package com.stevesoft.pat;
10 import jalview.util.MessageManager;
13 import java.io.FilenameFilter;
14 import java.util.BitSet;
15 import java.util.Hashtable;
17 import com.stevesoft.pat.wrap.StringWrap;
19 /** Matches a Unicode punctuation character. */
20 class UnicodePunct extends UniValidator
23 public int validate(StringLike s, int from, int to)
25 return from < s.length() && Prop.isPunct(s.charAt(from)) ? to : -1;
29 /** Matches a Unicode white space character. */
30 class UnicodeWhite extends UniValidator
33 public int validate(StringLike s, int from, int to)
35 return from < s.length() && Prop.isWhite(s.charAt(from)) ? to : -1;
40 * Matches a character that is not a Unicode punctuation character.
42 class NUnicodePunct extends UniValidator
45 public int validate(StringLike s, int from, int to)
47 return from < s.length() && !Prop.isPunct(s.charAt(from)) ? to : -1;
52 * Matches a character that is not a Unicode white space character.
54 class NUnicodeWhite extends UniValidator
57 public int validate(StringLike s, int from, int to)
59 return from < s.length() && !Prop.isWhite(s.charAt(from)) ? to : -1;
63 /** Matches a Unicode word character: an alphanumeric or underscore. */
64 class UnicodeW extends UniValidator
67 public int validate(StringLike s, int from, int to)
69 if (from >= s.length())
73 char c = s.charAt(from);
74 return (Prop.isAlphabetic(c) || Prop.isDecimalDigit(c) || c == '_') ? to
79 /** Matches a character that is not a Unicode alphanumeric or underscore. */
80 class NUnicodeW extends UniValidator
83 public int validate(StringLike s, int from, int to)
85 if (from >= s.length())
89 char c = s.charAt(from);
90 return !(Prop.isAlphabetic(c) || Prop.isDecimalDigit(c) || c == '_')
96 /** Matches a Unicode decimal digit. */
97 class UnicodeDigit extends UniValidator
100 public int validate(StringLike s, int from, int to)
102 return from < s.length() && Prop.isDecimalDigit(s.charAt(from)) ? to
107 /** Matches a character that is not a Unicode digit. */
108 class NUnicodeDigit extends UniValidator
111 public int validate(StringLike s, int from, int to)
113 return from < s.length() && !Prop.isDecimalDigit(s.charAt(from)) ? to
118 /** Matches a Unicode math character. */
119 class UnicodeMath extends UniValidator
122 public int validate(StringLike s, int from, int to)
124 return from < s.length() && Prop.isMath(s.charAt(from)) ? to : -1;
128 /** Matches a non-math Unicode character. */
129 class NUnicodeMath extends UniValidator
132 public int validate(StringLike s, int from, int to)
134 return from < s.length() && !Prop.isMath(s.charAt(from)) ? to : -1;
138 /** Matches a Unicode currency symbol. */
139 class UnicodeCurrency extends UniValidator
142 public int validate(StringLike s, int from, int to)
144 return from < s.length() && Prop.isCurrency(s.charAt(from)) ? to : -1;
148 /** Matches a non-currency symbol Unicode character. */
149 class NUnicodeCurrency extends UniValidator
152 public int validate(StringLike s, int from, int to)
154 return from < s.length() && !Prop.isCurrency(s.charAt(from)) ? to : -1;
158 /** Matches a Unicode alphabetic character. */
159 class UnicodeAlpha extends UniValidator
162 public int validate(StringLike s, int from, int to)
164 return from < s.length() && Prop.isAlphabetic(s.charAt(from)) ? to : -1;
168 /** Matches a non-alphabetic Unicode character. */
169 class NUnicodeAlpha extends UniValidator
172 public int validate(StringLike s, int from, int to)
174 return from < s.length() && !Prop.isAlphabetic(s.charAt(from)) ? to
179 /** Matches an upper case Unicode character. */
180 class UnicodeUpper extends UniValidator
183 public int validate(StringLike s, int from, int to)
185 return from < s.length() && isUpper(s.charAt(from)) ? to : -1;
188 final boolean isUpper(char c)
190 return c == CaseMgr.toUpperCase(c) && c != CaseMgr.toLowerCase(c);
194 /** Matches an upper case Unicode character. */
195 class UnicodeLower extends UniValidator
198 public int validate(StringLike s, int from, int to)
200 return from < s.length() && isLower(s.charAt(from)) ? to : -1;
203 final boolean isLower(char c)
205 return c != CaseMgr.toUpperCase(c) && c == CaseMgr.toLowerCase(c);
210 * Regex provides the parser which constructs the linked list of Pattern classes
213 * For the purpose of this documentation, the fact that java interprets the
214 * backslash will be ignored. In practice, however, you will need a double
215 * backslash to obtain a string that contains a single backslash character.
216 * Thus, the example pattern "\b" should really be typed as "\\b" inside java
219 * Note that Regex is part of package "com.stevesoft.pat". To use it, simply
220 * import com.stevesoft.pat.Regex at the top of your file.
222 * Regex is made with a constructor that takes a String that defines the regular
223 * expression. Thus, for example
226 * Regex r = new Regex("[a-c]*");
229 * matches any number of characters so long as the are 'a', 'b', or 'c').
231 * To attempt to match the Pattern to a given string, you can use either the
232 * search(String) member function, or the matchAt(String,int position) member
233 * function. These functions return a boolean which tells you whether or not the
234 * thing worked, and sets the methods "charsMatched()" and "matchedFrom()" in
235 * the Regex object appropriately.
237 * The portion of the string before the match can be obtained by the left()
238 * member, and the portion after the match can be obtained by the right()
241 * Essentially, this package implements a syntax that is very much like the perl
242 * 5 regular expression syntax.
247 * Regex r = new Regex("x(a|b)y");
248 * r.matchAt("xay", 0);
249 * System.out.println("sub = " + r.stringMatched(1));
252 * The above would print "sub = a".
255 * r.left() // would return "x"
256 * r.right() // would return "y"
260 * Differences between this package and perl5:<br>
261 * The extended Pattern for setting flags, is now supported, but the flags are
262 * different. "(?i)" tells the pattern to ignore case, "(?Q)" sets the
263 * "dontMatchInQuotes" flag, and "(?iQ)" sets them both. You can change the
264 * escape character. The pattern
276 * , but note that the sequence
282 * <b>must</b> occur at the very beginning of the pattern. There may be other
283 * small differences as well. I will either make my package conform or note them
284 * as I become aware of them.
286 * This package supports additional patterns not in perl5: <center> <table *
291 * <td>This matches all characters between the '(' character and the balancing
292 * ')' character. Thus, it will match "()" as well as "(())". The balancing
293 * characters are arbitrary, thus (?@{}) matches on "{}" and "{{}}".</td>
297 * <td>Moves the pointer backwards within the text. This allows you to make a
298 * "look behind." It fails if it attempts to move to a position before the
299 * beginning of the string. "x(?<1)" is equivalent to "(?=x)". The number, 1
300 * in this example, is the number of characters to move backwards.</td>
305 * @author Steven R. Brandt
306 * @version package com.stevesoft.pat, release 1.5.3
309 public class Regex extends RegRes implements FilenameFilter
312 * BackRefOffset gives the identity number of the first pattern. Version 1.0
313 * used zero, version 1.1 uses 1 to be more compatible with perl.
315 static int BackRefOffset = 1;
317 private static Pattern none = new NoPattern();
319 Pattern thePattern = none;
321 patInt minMatch = new patInt(0);
323 static Hashtable validators = new Hashtable();
326 define("p", "(?>1)", new UnicodePunct());
327 define("P", "(?>1)", new NUnicodePunct());
328 define("s", "(?>1)", new UnicodeWhite());
329 define("S", "(?>1)", new NUnicodeWhite());
330 define("w", "(?>1)", new UnicodeW());
331 define("W", "(?>1)", new NUnicodeW());
332 define("d", "(?>1)", new UnicodeDigit());
333 define("D", "(?>1)", new NUnicodeDigit());
334 define("m", "(?>1)", new UnicodeMath());
335 define("M", "(?>1)", new NUnicodeMath());
336 define("c", "(?>1)", new UnicodeCurrency());
337 define("C", "(?>1)", new NUnicodeCurrency());
338 define("a", "(?>1)", new UnicodeAlpha());
339 define("A", "(?>1)", new NUnicodeAlpha());
340 define("uc", "(?>1)", new UnicodeUpper());
341 define("lc", "(?>1)", new UnicodeLower());
344 /** Set the dontMatch in quotes flag. */
345 public void setDontMatchInQuotes(boolean b)
347 dontMatchInQuotes = b;
350 /** Find out if the dontMatchInQuotes flag is enabled. */
351 public boolean getDontMatchInQuotes()
353 return dontMatchInQuotes;
356 boolean dontMatchInQuotes = false;
359 * Set the state of the ignoreCase flag. If set to true, then the pattern
360 * matcher will ignore case when searching for a match.
362 public void setIgnoreCase(boolean b)
368 * Get the state of the ignoreCase flag. Returns true if we are ignoring the
369 * case of the pattern, false otherwise.
371 public boolean getIgnoreCase()
376 boolean ignoreCase = false;
378 static boolean defaultMFlag = false;
381 * Set the default value of the m flag. If it is set to true, then the MFlag
382 * will be on for any regex search executed.
384 public static void setDefaultMFlag(boolean mFlag)
386 defaultMFlag = mFlag;
390 * Get the default value of the m flag. If it is set to true, then the MFlag
391 * will be on for any regex search executed.
393 public static boolean getDefaultMFlag()
399 * Initializes the object without a Pattern. To supply a Pattern use
402 * @see com.stevesoft.pat.Regex#compile(java.lang.String)
409 * Create and compile a Regex, but do not throw any exceptions. If you wish to
410 * have exceptions thrown for syntax errors, you must use the Regex(void)
411 * constructor to create the Regex object, and then call the compile method.
412 * Therefore, you should only call this method when you know your pattern is
413 * right. I will probably become more like
415 * @see com.stevesoft.pat.Regex#search(java.lang.String)
416 * @see com.stevesoft.pat.Regex#compile(java.lang.String)
418 public Regex(String s)
423 } catch (RegSyntax rs)
428 ReplaceRule rep = null;
431 * Create and compile both a Regex and a ReplaceRule.
433 * @see com.stevesoft.pat.ReplaceRule
434 * @see com.stevesoft.pat.Regex#compile(java.lang.String)
436 public Regex(String s, String rp)
439 rep = ReplaceRule.perlCode(rp);
443 * Create and compile a Regex, but give it the ReplaceRule specified. This
444 * allows the user finer control of the Replacement process, if that is
447 * @see com.stevesoft.pat.ReplaceRule
448 * @see com.stevesoft.pat.Regex#compile(java.lang.String)
450 public Regex(String s, ReplaceRule rp)
457 * Change the ReplaceRule of this Regex by compiling a new one using String
460 public void setReplaceRule(String rp)
462 rep = ReplaceRule.perlCode(rp);
463 repr = null; // Clear Replacer history
466 /** Change the ReplaceRule of this Regex to rp. */
467 public void setReplaceRule(ReplaceRule rp)
473 * Test to see if a custom defined rule exists.
475 * @see com.stevesoft.pat#define(java.lang.String,java.lang.String,Validator)
477 public static boolean isDefined(String nm)
479 return validators.get(nm) != null;
483 * Removes a custom defined rule.
485 * @see com.stevesoft.pat#define(java.lang.String,java.lang.String,Validator)
487 public static void undefine(String nm)
489 validators.remove(nm);
493 * Defines a method to create a new rule. See test/deriv2.java and
494 * test/deriv3.java for examples of how to use it.
496 public static void define(String nm, String pat, Validator v)
499 validators.put(nm, v);
503 * Defines a shorthand for a pattern. The pattern will be invoked by a string
504 * that has the form "(??"+nm+")".
506 public static void define(String nm, String pat)
508 validators.put(nm, pat);
511 /** Get the current ReplaceRule. */
512 public ReplaceRule getReplaceRule()
517 Replacer repr = null;
519 final Replacer _getReplacer()
521 return repr == null ? repr = new Replacer() : repr;
524 public Replacer getReplacer()
528 repr = new Replacer();
536 * Replace the first occurence of this pattern in String s according to the
539 * @see com.stevesoft.pat.ReplaceRule
540 * @see com.stevesoft.pat.Regex#getReplaceRule()
542 public String replaceFirst(String s)
544 return _getReplacer().replaceFirstRegion(s, this, 0, s.length())
549 * Replace the first occurence of this pattern in String s beginning with
550 * position pos according to the ReplaceRule.
552 * @see com.stevesoft.pat.ReplaceRule
553 * @see com.stevesoft.pat.Regex#getReplaceRule()
555 public String replaceFirstFrom(String s, int pos)
557 return _getReplacer().replaceFirstRegion(s, this, pos, s.length())
562 * Replace the first occurence of this pattern in String s beginning with
563 * position start and ending with end according to the ReplaceRule.
565 * @see com.stevesoft.pat.ReplaceRule
566 * @see com.stevesoft.pat.Regex#getReplaceRule()
568 public String replaceFirstRegion(String s, int start, int end)
570 return _getReplacer().replaceFirstRegion(s, this, start, end)
575 * Replace all occurences of this pattern in String s according to the
578 * @see com.stevesoft.pat.ReplaceRule
579 * @see com.stevesoft.pat.Regex#getReplaceRule()
581 public String replaceAll(String s)
583 return _getReplacer().replaceAllRegion(s, this, 0, s.length())
587 public StringLike replaceAll(StringLike s)
589 return _getReplacer().replaceAllRegion(s, this, 0, s.length());
593 * Replace all occurences of this pattern in String s beginning with position
594 * pos according to the ReplaceRule.
596 * @see com.stevesoft.pat.ReplaceRule
597 * @see com.stevesoft.pat.Regex#getReplaceRule()
599 public String replaceAllFrom(String s, int pos)
601 return _getReplacer().replaceAllRegion(s, this, pos, s.length())
606 * Replace all occurences of this pattern in String s beginning with position
607 * start and ending with end according to the ReplaceRule.
609 * @see com.stevesoft.pat.ReplaceRule
610 * @see com.stevesoft.pat.Regex#getReplaceRule()
612 public String replaceAllRegion(String s, int start, int end)
614 return _getReplacer().replaceAllRegion(s, this, start, end).toString();
617 /** Essentially clones the Regex object */
618 public Regex(Regex r)
621 dontMatchInQuotes = r.dontMatchInQuotes;
623 ignoreCase = r.ignoreCase;
631 rep = (ReplaceRule) r.rep.clone();
634 * try { compile(r.toString()); } catch(RegSyntax r_) {}
636 thePattern = r.thePattern.clone(new Hashtable());
637 minMatch = r.minMatch;
642 * By default, the escape character is the backslash, but you can make it
643 * anything you want by setting this variable.
645 public char esc = Pattern.ESC;
648 * This method compiles a regular expression, making it possible to call the
649 * search or matchAt methods.
651 * @exception com.stevesoft.pat.RegSyntax
652 * is thrown if a syntax error is encountered in the pattern. For
653 * example, "x{3,1}" or "*a" are not valid patterns.
654 * @see com.stevesoft.pat.Regex#search
655 * @see com.stevesoft.pat.Regex#matchAt
657 public void compile(String prepat) throws RegSyntax
659 String postpat = parsePerl.codify(prepat, true);
660 String pat = postpat == null ? prepat : postpat;
663 dontMatchInQuotes = false;
664 Rthings mk = new Rthings(this);
670 minMatch = new patInt(0);
671 StrPos sp = new StrPos(pat, 0);
672 if (sp.incMatch("(?e="))
678 newpat = reEscape(pat.substring(6), newEsc, Pattern.ESC);
681 else if (esc != Pattern.ESC)
683 newpat = reEscape(pat, esc, Pattern.ESC);
685 thePattern = _compile(newpat, mk);
686 numSubs_ = mk.val - offset;
691 * If a Regex is compared against a Regex, a check is done to see that the
692 * patterns are equal as well as the most recent match. If a Regex is compare
693 * with a RegRes, only the result of the most recent match is compared.
696 public boolean equals(Object o)
698 if (o instanceof Regex)
700 if (toString().equals(o.toString()))
702 return super.equals(o);
711 return super.equals(o);
715 /** A clone by any other name would smell as sweet. */
717 public Object clone()
719 return new Regex(this);
722 /** Return a clone of the underlying RegRes object. */
723 public RegRes result()
725 return (RegRes) super.clone();
728 // prep sets global variables of class
729 // Pattern so that it can access them
730 // during an attempt at a match
731 Pthings pt = new Pthings();
733 final Pthings prep(StringLike s)
736 pt.lastPos = matchedTo();
741 if ((s == null ? null : s.unwrap()) != (src == null ? null
747 pt.dotDoesntMatchCR = dotDoesntMatchCR && (!sFlag);
748 pt.mFlag = (mFlag | defaultMFlag);
749 pt.ignoreCase = ignoreCase;
751 if (pt.marks != null)
753 for (int i = 0; i < pt.marks.length; i++)
759 pt.nMarks = numSubs_;
761 if (dontMatchInQuotes)
773 * Attempt to match a Pattern beginning at a specified location within the
776 * @see com.stevesoft.pat.Regex#search
778 public boolean matchAt(String s, int start_pos)
780 return _search(s, start_pos, start_pos);
784 * Attempt to match a Pattern beginning at a specified location within the
787 * @see com.stevesoft.pat.Regex#search
789 public boolean matchAt(StringLike s, int start_pos)
791 return _search(s, start_pos, start_pos);
795 * Search through a String for the first occurrence of a match.
797 * @see com.stevesoft.pat.Regex#searchFrom
798 * @see com.stevesoft.pat.Regex#matchAt
800 public boolean search(String s)
804 throw new NullPointerException(MessageManager
805 .getString("exception.null_string_given_to_regex_search"));
807 return _search(s, 0, s.length());
810 public boolean search(StringLike sl)
814 throw new NullPointerException(MessageManager.getString(
815 "exception.null_string_like_given_to_regex_search"));
817 return _search(sl, 0, sl.length());
820 public boolean reverseSearch(String s)
824 throw new NullPointerException(MessageManager.getString(
825 "exception.null_string_given_to_regex_reverse_search"));
827 return _reverseSearch(s, 0, s.length());
830 public boolean reverseSearch(StringLike sl)
834 throw new NullPointerException(MessageManager.getString(
835 "exception.null_string_like_given_to_regex_reverse_search"));
837 return _reverseSearch(sl, 0, sl.length());
841 * Search through a String for the first occurence of a match, but start at
848 public boolean searchFrom(String s, int start)
852 throw new NullPointerException(MessageManager.getString(
853 "exception.null_string_like_given_to_regex_search_from"));
855 return _search(s, start, s.length());
858 public boolean searchFrom(StringLike s, int start)
862 throw new NullPointerException(MessageManager.getString(
863 "exception.null_string_like_given_to_regex_search_from"));
865 return _search(s, start, s.length());
869 * Search through a region of a String for the first occurence of a match.
871 public boolean searchRegion(String s, int start, int end)
875 throw new NullPointerException(MessageManager.getString(
876 "exception.null_string_like_given_to_regex_search_region"));
878 return _search(s, start, end);
882 * Set this to change the default behavior of the "." pattern. By default it
883 * now matches perl's behavior and fails to match the '\n' character.
885 public static boolean dotDoesntMatchCR = true;
891 boolean gFlag = false;
893 /** Set the 'g' flag */
894 public void setGFlag(boolean b)
899 /** Get the state of the 'g' flag. */
900 public boolean getGFlag()
905 boolean sFlag = false;
907 /** Get the state of the sFlag */
908 public boolean getSFlag()
913 boolean mFlag = false;
915 /** Get the state of the sFlag */
916 public boolean getMFlag()
921 final boolean _search(String s, int start, int end)
923 return _search(new StringWrap(s), start, end);
926 final boolean _search(StringLike s, int start, int end)
928 if (gFlag && gFlagto > 0 && gFlags != null
929 && s.unwrap() == gFlags.unwrap())
935 Pthings pt = prep(s);
937 int up = (minMatch == null ? end : end - minMatch.i);
939 if (up < start && end >= start)
946 for (int i = start; i <= up; i++)
948 charsMatched_ = thePattern.matchAt(s, i, pt);
949 if (charsMatched_ >= 0)
951 matchFrom_ = thePattern.mfrom;
953 gFlagto = matchFrom_ + charsMatched_;
955 return didMatch_ = true;
962 for (int i = start; i <= up; i++)
964 i = skipper.find(src, i, up);
967 charsMatched_ = matchFrom_ = -1;
968 return didMatch_ = false;
970 charsMatched_ = thePattern.matchAt(s, i, pt);
971 if (charsMatched_ >= 0)
973 matchFrom_ = thePattern.mfrom;
975 gFlagto = matchFrom_ + charsMatched_;
977 return didMatch_ = true;
981 return didMatch_ = false;
985 * final boolean _search(LongStringLike s,long start,long end) { if(gFlag &&
986 * gFlagto > 0 && s==gFlags) start = gFlagto; gFlags = null;
988 * Pthings pt=prep(s);
990 * int up = end;//(minMatch == null ? end : end-minMatch.i);
992 * if(up < start && end >= start) up = start;
994 * if(skipper == null) { for(long i=start;i<=up;i++) { charsMatched_ =
995 * thePattern.matchAt(s,i,pt); if(charsMatched_ >= 0) { matchFrom_ =
996 * thePattern.mfrom; marks = pt.marks; gFlagto = matchFrom_+charsMatched_;
997 * return didMatch_=true; } } } else { pt.no_check = true; for(long
998 * i=start;i<=up;i++) { i = skipper.find(src,i,up); if(i<0) { charsMatched_ =
999 * matchFrom_ = -1; return didMatch_ = false; } charsMatched_ =
1000 * thePattern.matchAt(s,i,pt); if(charsMatched_ >= 0) { matchFrom_ =
1001 * thePattern.mfrom; marks = pt.marks; gFlagto = matchFrom_+charsMatched_;
1002 * gFlags = s; return didMatch_=true; } else { i = s.adjustIndex(i); up =
1003 * s.adjustEnd(i); } } } return didMatch_=false; }
1006 boolean _reverseSearch(String s, int start, int end)
1008 return _reverseSearch(new StringWrap(s), start, end);
1011 boolean _reverseSearch(StringLike s, int start, int end)
1013 if (gFlag && gFlagto > 0 && s.unwrap() == gFlags.unwrap())
1018 Pthings pt = prep(s);
1019 for (int i = end; i >= start; i--)
1021 charsMatched_ = thePattern.matchAt(s, i, pt);
1022 if (charsMatched_ >= 0)
1024 matchFrom_ = thePattern.mfrom;
1026 gFlagto = matchFrom_ - 1;
1028 return didMatch_ = true;
1031 return didMatch_ = false;
1034 // This routine sets the cbits variable
1035 // of class Pattern. Cbits is true for
1036 // the bit corresponding to a character inside
1038 static StringLike lasts = null;
1040 static BitSet lastbs = null;
1042 static void setCbits(StringLike s, Pthings pt)
1049 BitSet bs = new BitSet(s.length());
1051 boolean setBit = false;
1052 for (int i = 0; i < s.length(); i++)
1058 char c = s.charAt(i);
1059 if (!setBit && c == '"')
1065 else if (!setBit && c == '\'')
1071 else if (setBit && c == qc)
1075 else if (setBit && c == '\\' && i + 1 < s.length())
1084 pt.cbits = lastbs = bs;
1088 // Wanted user to over-ride this in alpha version,
1089 // but it wasn't really necessary because of this trick:
1094 return getClass().getDeclaredConstructor().newInstance();
1095 } catch (InstantiationException ie)
1098 } catch (IllegalAccessException iae)
1101 } catch (ReflectiveOperationException roe)
1108 * Only needed for creating your own extensions of Regex. This method adds the
1109 * next Pattern in the chain of patterns or sets the Pattern if it is the
1112 protected void add(Pattern p2)
1126 * You only need to use this method if you are creating your own extentions to
1127 * Regex. compile1 compiles one Pattern element, it can be over-ridden to
1128 * allow the Regex compiler to understand new syntax. See deriv.java for an
1129 * example. This routine is the heart of class Regex. Rthings has one integer
1130 * member called intValue, it is used to keep track of the number of ()'s in
1133 * @exception com.stevesoft.pat.RegSyntax
1134 * is thrown when a nonsensensical pattern is supplied. For
1135 * example, a pattern beginning with *.
1137 protected void compile1(StrPos sp, Rthings mk) throws RegSyntax
1142 add(matchBracket(sp));
1144 else if (sp.match('|'))
1152 p = new NullPattern();
1157 else if (sp.incMatch("(?<"))
1159 patInt i = sp.getPatInt();
1162 RegSyntaxError.endItAll("No int after (?<");
1164 add(new Backup(i.intValue()));
1167 RegSyntaxError.endItAll("No ) after (?<");
1170 else if (sp.incMatch("(?>"))
1172 patInt i = sp.getPatInt();
1175 RegSyntaxError.endItAll("No int after (?>");
1177 add(new Backup(-i.intValue()));
1180 RegSyntaxError.endItAll("No ) after (?<");
1183 else if (sp.incMatch("(?@"))
1191 RegSyntaxError.endItAll("(?@ does not have closing paren");
1193 add(new Group(op, cl));
1195 else if (sp.incMatch("(?#"))
1197 while (!sp.match(')'))
1202 else if (sp.dontMatch && sp.c == 'w')
1204 // Regex r = new Regex();
1205 // r._compile("[a-zA-Z0-9_]",mk);
1206 // add(new Goop("\\w",r.thePattern));
1207 Bracket b = new Bracket(false);
1208 b.addOr(new Range('a', 'z'));
1209 b.addOr(new Range('A', 'Z'));
1210 b.addOr(new Range('0', '9'));
1211 b.addOr(new oneChar('_'));
1214 else if (sp.dontMatch && sp.c == 'G')
1218 else if (sp.dontMatch && sp.c == 's')
1220 // Regex r = new Regex();
1221 // r._compile("[ \t\n\r\b]",mk);
1222 // add(new Goop("\\s",r.thePattern));
1223 Bracket b = new Bracket(false);
1224 b.addOr(new oneChar((char) 32));
1225 b.addOr(new Range((char) 8, (char) 10));
1226 b.addOr(new oneChar((char) 13));
1229 else if (sp.dontMatch && sp.c == 'd')
1231 // Regex r = new Regex();
1232 // r._compile("[0-9]",mk);
1233 // add(new Goop("\\d",r.thePattern));
1234 Range digit = new Range('0', '9');
1235 digit.printBrackets = true;
1238 else if (sp.dontMatch && sp.c == 'W')
1240 // Regex r = new Regex();
1241 // r._compile("[^a-zA-Z0-9_]",mk);
1242 // add(new Goop("\\W",r.thePattern));
1243 Bracket b = new Bracket(true);
1244 b.addOr(new Range('a', 'z'));
1245 b.addOr(new Range('A', 'Z'));
1246 b.addOr(new Range('0', '9'));
1247 b.addOr(new oneChar('_'));
1250 else if (sp.dontMatch && sp.c == 'S')
1252 // Regex r = new Regex();
1253 // r._compile("[^ \t\n\r\b]",mk);
1254 // add(new Goop("\\S",r.thePattern));
1255 Bracket b = new Bracket(true);
1256 b.addOr(new oneChar((char) 32));
1257 b.addOr(new Range((char) 8, (char) 10));
1258 b.addOr(new oneChar((char) 13));
1261 else if (sp.dontMatch && sp.c == 'D')
1263 // Regex r = new Regex();
1264 // r._compile("[^0-9]",mk);
1265 // add(new Goop("\\D",r.thePattern));
1266 Bracket b = new Bracket(true);
1267 b.addOr(new Range('0', '9'));
1270 else if (sp.dontMatch && sp.c == 'B')
1272 Regex r = new Regex();
1273 r._compile("(?!" + back_slash + "b)", mk);
1276 else if (isOctalString(sp))
1280 d = 8 * d + sp.c - '0';
1281 StrPos sp2 = new StrPos(sp);
1283 if (isOctalDigit(sp2, false))
1286 d = 8 * d + sp.c - '0';
1288 add(new oneChar((char) d));
1290 else if (sp.dontMatch && sp.c >= '1' && sp.c <= '9')
1292 int iv = sp.c - '0';
1293 StrPos s2 = new StrPos(sp);
1295 if (!s2.dontMatch && s2.c >= '0' && s2.c <= '9')
1297 iv = 10 * iv + (s2.c - '0');
1300 add(new BackMatch(iv));
1302 else if (sp.dontMatch && sp.c == 'b')
1304 add(new Boundary());
1306 else if (sp.match('\b'))
1308 add(new Boundary());
1310 else if (sp.match('$'))
1314 else if (sp.dontMatch && sp.c == 'Z')
1316 add(new End(false));
1318 else if (sp.match('.'))
1322 else if (sp.incMatch("(??"))
1324 StringBuffer sb = new StringBuffer();
1325 StringBuffer sb2 = new StringBuffer();
1326 while (!sp.match(')') && !sp.match(':'))
1331 if (sp.incMatch(":"))
1333 while (!sp.match(')'))
1339 String sbs = sb.toString();
1340 if (validators.get(sbs) instanceof String)
1342 String pat = (String) validators.get(sbs);
1343 Regex r = newRegex();
1344 Rthings rth = new Rthings(this);
1345 rth.noBackRefs = true;
1346 r._compile(pat, rth);
1351 Custom cm = new Custom(sb.toString());
1354 Validator v2 = cm.v.arg(sb2.toString());
1357 v2.argsave = sb2.toString();
1358 String p = cm.v.pattern;
1362 Regex r = newRegex();
1363 Rthings rth = new Rthings(this);
1364 rth.noBackRefs = true;
1365 r._compile(cm.v.pattern, rth);
1366 cm.sub = r.thePattern;
1367 cm.sub.add(new CustomEndpoint(cm));
1368 cm.sub.setParent(cm);
1373 else if (sp.match('('))
1376 Regex r = newRegex();
1379 if (sp.incMatch("?:"))
1383 else if (sp.incMatch("?="))
1385 r.or = new lookAhead(false);
1387 else if (sp.incMatch("?!"))
1389 r.or = new lookAhead(true);
1391 else if (sp.match('?'))
1398 mk.ignoreCase = true;
1402 mk.dontMatchInQuotes = true;
1406 mk.optimizeMe = true;
1421 } while (!sp.match(')') && !sp.eos);
1424 if (sp.eos) // throw new RegSyntax
1426 RegSyntaxError.endItAll("Unclosed ()");
1430 { // just ordinary parenthesis
1431 r.or = mk.noBackRefs ? new Or() : new OrMark(mk.val++);
1435 add(r._compile(sp, mk));
1438 else if (sp.match('^'))
1440 add(new Start(true));
1442 else if (sp.dontMatch && sp.c == 'A')
1444 add(new Start(false));
1446 else if (sp.match('*'))
1448 addMulti(new patInt(0), new patInf());
1450 else if (sp.match('+'))
1452 addMulti(new patInt(1), new patInf());
1454 else if (sp.match('?'))
1456 addMulti(new patInt(0), new patInt(1));
1458 else if (sp.match('{'))
1460 boolean bad = false;
1461 StrPos sp2 = new StrPos(sp);
1462 // StringBuffer sb = new StringBuffer();
1464 patInt i1 = sp.getPatInt();
1475 * RegSyntaxError.endItAll( "String \"{"+i2+ "\" should be followed
1487 i2 = sp.getPatInt();
1490 if (i1 == null || i2 == null)
1493 * throw new RegSyntax("Badly formatted Multi: " +"{"+i1+","+i2+"}");
1500 add(new oneChar(sp.c));
1507 else if (sp.escMatch('x') && next2Hex(sp))
1510 int d = getHexDigit(sp);
1512 d = 16 * d + getHexDigit(sp);
1513 add(new oneChar((char) d));
1515 else if (sp.escMatch('c'))
1518 if (sp.c < Ctrl.cmap.length)
1520 add(new oneChar(Ctrl.cmap[sp.c]));
1524 add(new oneChar(sp.c));
1527 else if (sp.escMatch('f'))
1529 add(new oneChar((char) 12));
1531 else if (sp.escMatch('a'))
1533 add(new oneChar((char) 7));
1535 else if (sp.escMatch('t'))
1537 add(new oneChar('\t'));
1539 else if (sp.escMatch('n'))
1541 add(new oneChar('\n'));
1543 else if (sp.escMatch('r'))
1545 add(new oneChar('\r'));
1547 else if (sp.escMatch('b'))
1549 add(new oneChar('\b'));
1551 else if (sp.escMatch('e'))
1553 add(new oneChar((char) 27));
1557 add(new oneChar(sp.c));
1560 RegSyntaxError.endItAll("Unmatched right paren in pattern");
1565 // compiles all Pattern elements, internal method
1566 private Pattern _compile(String pat, Rthings mk) throws RegSyntax
1569 sFlag = mFlag = ignoreCase = gFlag = false;
1570 StrPos sp = new StrPos(pat, 0);
1571 thePattern = _compile(sp, mk);
1580 Pattern _compile(StrPos sp, Rthings mk) throws RegSyntax
1582 while (!(sp.eos || (or != null && sp.match(')'))))
1591 else if (sp.eos && mk.parenLevel != 0)
1593 RegSyntaxError.endItAll("Unclosed Parenthesis! lvl=" + mk.parenLevel);
1599 p = new NullPattern();
1604 return p == null ? new NullPattern() : p;
1607 // add a multi object to the end of the chain
1608 // which applies to the last object
1609 void addMulti(patInt i1, patInt i2) throws RegSyntax
1611 Pattern last, last2;
1612 for (last = p; last != null && last.next != null; last = last.next)
1616 if (last == null || last == p)
1622 for (last2 = p; last2.next != last; last2 = last2.next)
1627 if (last instanceof Multi && i1.intValue() == 0 && i2.intValue() == 1)
1629 ((Multi) last).matchFewest = true;
1631 else if (last instanceof FastMulti && i1.intValue() == 0
1632 && i2.intValue() == 1)
1634 ((FastMulti) last).matchFewest = true;
1636 else if (last instanceof DotMulti && i1.intValue() == 0
1637 && i2.intValue() == 1)
1639 ((DotMulti) last).matchFewest = true;
1641 else if (last instanceof Multi || last instanceof DotMulti
1642 || last instanceof FastMulti)
1644 throw new RegSyntax("Syntax error.");
1646 else if (last2 == null)
1648 p = mkMulti(i1, i2, p);
1652 last2.next = mkMulti(i1, i2, last);
1656 final static Pattern mkMulti(patInt lo, patInt hi, Pattern p)
1659 if (p instanceof Any && p.next == null)
1661 return new DotMulti(lo, hi);
1663 return RegOpt.safe4fm(p) ? (Pattern) new FastMulti(lo, hi, p)
1664 : (Pattern) new Multi(lo, hi, p);
1667 // process the bracket operator
1668 Pattern matchBracket(StrPos sp) throws RegSyntax
1673 ret = new Bracket(true);
1678 ret = new Bracket(false);
1682 // throw new RegSyntax
1683 RegSyntaxError.endItAll("Unmatched []");
1686 while (!sp.eos && !sp.match(']'))
1688 StrPos s1 = new StrPos(sp);
1690 StrPos s1_ = new StrPos(s1);
1692 if (s1.match('-') && !s1_.match(']'))
1694 StrPos s2 = new StrPos(s1);
1698 ret.addOr(new Range(sp.c, s2.c));
1703 else if (sp.escMatch('Q'))
1706 while (!sp.escMatch('E'))
1708 ret.addOr(new oneChar(sp.c));
1712 else if (sp.escMatch('d'))
1714 ret.addOr(new Range('0', '9'));
1716 else if (sp.escMatch('s'))
1718 ret.addOr(new oneChar((char) 32));
1719 ret.addOr(new Range((char) 8, (char) 10));
1720 ret.addOr(new oneChar((char) 13));
1722 else if (sp.escMatch('w'))
1724 ret.addOr(new Range('a', 'z'));
1725 ret.addOr(new Range('A', 'Z'));
1726 ret.addOr(new Range('0', '9'));
1727 ret.addOr(new oneChar('_'));
1729 else if (sp.escMatch('D'))
1731 ret.addOr(new Range((char) 0, (char) 47));
1732 ret.addOr(new Range((char) 58, (char) 65535));
1734 else if (sp.escMatch('S'))
1736 ret.addOr(new Range((char) 0, (char) 7));
1737 ret.addOr(new Range((char) 11, (char) 12));
1738 ret.addOr(new Range((char) 14, (char) 31));
1739 ret.addOr(new Range((char) 33, (char) 65535));
1741 else if (sp.escMatch('W'))
1743 ret.addOr(new Range((char) 0, (char) 64));
1744 ret.addOr(new Range((char) 91, (char) 94));
1745 ret.addOr(new oneChar((char) 96));
1746 ret.addOr(new Range((char) 123, (char) 65535));
1748 else if (sp.escMatch('x') && next2Hex(sp))
1751 int d = getHexDigit(sp);
1753 d = 16 * d + getHexDigit(sp);
1754 ret.addOr(new oneChar((char) d));
1756 else if (sp.escMatch('a'))
1758 ret.addOr(new oneChar((char) 7));
1760 else if (sp.escMatch('f'))
1762 ret.addOr(new oneChar((char) 12));
1764 else if (sp.escMatch('e'))
1766 ret.addOr(new oneChar((char) 27));
1768 else if (sp.escMatch('n'))
1770 ret.addOr(new oneChar('\n'));
1772 else if (sp.escMatch('t'))
1774 ret.addOr(new oneChar('\t'));
1776 else if (sp.escMatch('r'))
1778 ret.addOr(new oneChar('\r'));
1780 else if (sp.escMatch('c'))
1783 if (sp.c < Ctrl.cmap.length)
1785 ret.addOr(new oneChar(Ctrl.cmap[sp.c]));
1789 ret.addOr(new oneChar(sp.c));
1792 else if (isOctalString(sp))
1796 d = 8 * d + sp.c - '0';
1797 StrPos sp2 = new StrPos(sp);
1799 if (isOctalDigit(sp2, false))
1802 d = 8 * d + sp.c - '0';
1804 ret.addOr(new oneChar((char) d));
1808 ret.addOr(new oneChar(sp.c));
1816 * Converts the stored Pattern to a String -- this is a decompile. Note that
1817 * \t and \n will really print out here, Not just the two character
1818 * representations. Also be prepared to see some strange output if your
1819 * characters are not printable.
1822 public String toString()
1824 if (false && thePattern == null)
1830 StringBuffer sb = new StringBuffer();
1831 if (esc != Pattern.ESC)
1837 if (gFlag || mFlag || !dotDoesntMatchCR || sFlag || ignoreCase
1838 || dontMatchInQuotes || optimized())
1849 if (sFlag || !dotDoesntMatchCR)
1853 if (dontMatchInQuotes)
1867 String patstr = thePattern.toString();
1868 if (esc != Pattern.ESC)
1870 patstr = reEscape(patstr, Pattern.ESC, esc);
1873 return sb.toString();
1877 // Re-escape Pattern, allows us to use a different escape
1879 static String reEscape(String s, char oldEsc, char newEsc)
1881 if (oldEsc == newEsc)
1886 StringBuffer sb = new StringBuffer();
1887 for (i = 0; i < s.length(); i++)
1889 if (s.charAt(i) == oldEsc && i + 1 < s.length())
1891 if (s.charAt(i + 1) == oldEsc)
1898 sb.append(s.charAt(i + 1));
1902 else if (s.charAt(i) == newEsc)
1909 sb.append(s.charAt(i));
1912 return sb.toString();
1916 * This method implements FilenameFilter, allowing one to use a Regex to
1917 * search through a directory using File.list. There is a FileRegex now that
1920 * @see com.stevesoft.pat.FileRegex
1923 public boolean accept(File dir, String s)
1928 /** The version of this package */
1929 final static public String version()
1931 return "lgpl release 1.5.3";
1935 * Once this method is called, the state of variables ignoreCase and
1936 * dontMatchInQuotes should not be changed as the results will be
1937 * unpredictable. However, search and matchAt will run more quickly. Note that
1938 * you can check to see if the pattern has been optimized by calling the
1939 * optimized() method.
1941 * This method will attempt to rewrite your pattern in a way that makes it
1942 * faster (not all patterns execute at the same speed). In general, "(?: ...
1943 * )" will be faster than "( ... )" so if you don't need the backreference,
1944 * you should group using the former pattern.
1946 * It will also introduce new pattern elements that you can't get to
1947 * otherwise, for example if you have a large table of strings, i.e. the
1948 * months of the year "(January|February|...)" optimize() will make a
1949 * Hashtable that takes it to the next appropriate pattern element --
1950 * eliminating the need for a linear search.
1952 * @see com.stevesoft.pat.Regex#optimized
1953 * @see com.stevesoft.pat.Regex#ignoreCase
1954 * @see com.stevesoft.pat.Regex#dontMatchInQuotes
1955 * @see com.stevesoft.pat.Regex#matchAt
1956 * @see com.stevesoft.pat.Regex#search
1958 public void optimize()
1960 if (optimized() || thePattern == null)
1964 minMatch = new patInt(0); // thePattern.countMinChars();
1965 thePattern = RegOpt.opt(thePattern, ignoreCase, dontMatchInQuotes);
1966 skipper = Skip.findSkip(this);
1967 // RegOpt.setParents(this);
1974 * This function returns true if the optimize method has been called.
1976 public boolean optimized()
1978 return minMatch != null;
1982 * A bit of syntactic surgar for those who want to make their code look more
1983 * perl-like. To use this initialize your Regex object by saying:
1986 * Regex r1 = Regex.perlCode("s/hello/goodbye/");
1987 * Regex r2 = Regex.perlCode("s'fish'frog'i");
1988 * Regex r3 = Regex.perlCode("m'hello');
1991 * The i for ignoreCase is supported in this syntax, as well as m, s, and x.
1992 * The g flat is a bit of a special case.
1994 * If you wish to replace all occurences of a pattern, you do not put a 'g' in
1995 * the perlCode, but call Regex's replaceAll method.
1997 * If you wish to simply and only do a search for r2's pattern, you can do
1998 * this by calling the searchFrom method method repeatedly, or by calling
1999 * search repeatedly if the g flag is set.
2001 * Note: Currently perlCode does <em>not</em> support the (?e=#) syntax for
2002 * changing the escape character.
2005 public static Regex perlCode(String s)
2007 // this file is big enough, see parsePerl.java
2008 // for this function.
2009 return parsePerl.parse(s);
2012 static final char back_slash = '\\';
2015 * Checks to see if there are only literal and no special pattern elements in
2018 public boolean isLiteral()
2020 Pattern x = thePattern;
2023 if (x instanceof oneChar)
2027 else if (x instanceof Skipped)
2041 * You only need to know about this if you are inventing your own pattern
2044 public patInt countMinChars()
2046 return thePattern.countMinChars();
2050 * You only need to know about this if you are inventing your own pattern
2053 public patInt countMaxChars()
2055 return thePattern.countMaxChars();
2058 boolean isHexDigit(StrPos sp)
2060 boolean r = !sp.eos && !sp.dontMatch
2061 && ((sp.c >= '0' && sp.c <= '9') || (sp.c >= 'a' && sp.c <= 'f')
2062 || (sp.c >= 'A' && sp.c <= 'F'));
2066 boolean isOctalDigit(StrPos sp, boolean first)
2068 boolean r = !sp.eos && !(first ^ sp.dontMatch) && sp.c >= '0'
2073 int getHexDigit(StrPos sp)
2075 if (sp.c >= '0' && sp.c <= '9')
2079 if (sp.c >= 'a' && sp.c <= 'f')
2081 return sp.c - 'a' + 10;
2083 return sp.c - 'A' + 10;
2086 boolean next2Hex(StrPos sp)
2088 StrPos sp2 = new StrPos(sp);
2090 if (!isHexDigit(sp2))
2095 if (!isHexDigit(sp2))
2102 boolean isOctalString(StrPos sp)
2104 if (!isOctalDigit(sp, true))
2108 StrPos sp2 = new StrPos(sp);
2110 if (!isOctalDigit(sp2, false))