-//\r
-// This software is now distributed according to\r
-// the Lesser Gnu Public License. Please see\r
-// http://www.gnu.org/copyleft/lesser.txt for\r
-// the details.\r
-// -- Happy Computing!\r
-//\r
-package com.stevesoft.pat;\r
-\r
-import java.io.*;\r
-import java.util.*;\r
-\r
-import com.stevesoft.pat.wrap.*;\r
-\r
-/** Matches a Unicode punctuation character. */\r
-class UnicodePunct\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && Prop.isPunct(s.charAt(from)) ? to : -1;\r
- }\r
-}\r
-\r
-/** Matches a Unicode white space character. */\r
-class UnicodeWhite\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && Prop.isWhite(s.charAt(from)) ? to : -1;\r
- }\r
-}\r
-\r
-/** Matches a character that is not a Unicode punctuation\r
- * character.\r
- */\r
-class NUnicodePunct\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && !Prop.isPunct(s.charAt(from)) ? to : -1;\r
- }\r
-}\r
-\r
-/** Matches a character that is not a\r
- * Unicode white space character.\r
- */\r
-class NUnicodeWhite\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && !Prop.isWhite(s.charAt(from)) ? to : -1;\r
- }\r
-}\r
-\r
-/** Matches a Unicode word character: an alphanumeric or underscore. */\r
-class UnicodeW\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- if (from >= s.length())\r
- {\r
- return -1;\r
- }\r
- char c = s.charAt(from);\r
- return (Prop.isAlphabetic(c) || Prop.isDecimalDigit(c) || c == '_') ? to :\r
- -1;\r
- }\r
-}\r
-\r
-/** Matches a character that is not a Unicode alphanumeric or underscore. */\r
-class NUnicodeW\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- if (from >= s.length())\r
- {\r
- return -1;\r
- }\r
- char c = s.charAt(from);\r
- return! (Prop.isAlphabetic(c) || Prop.isDecimalDigit(c) || c == '_') ? to :\r
- -1;\r
- }\r
-}\r
-\r
-/** Matches a Unicode decimal digit. */\r
-class UnicodeDigit\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && Prop.isDecimalDigit(s.charAt(from)) ? to : -1;\r
- }\r
-}\r
-\r
-/** Matches a character that is not a Unicode digit.*/\r
-class NUnicodeDigit\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && !Prop.isDecimalDigit(s.charAt(from)) ? to : -1;\r
- }\r
-}\r
-\r
-/** Matches a Unicode math character. */\r
-class UnicodeMath\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && Prop.isMath(s.charAt(from)) ? to : -1;\r
- }\r
-}\r
-\r
-/** Matches a non-math Unicode character. */\r
-class NUnicodeMath\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && !Prop.isMath(s.charAt(from)) ? to : -1;\r
- }\r
-}\r
-\r
-/** Matches a Unicode currency symbol. */\r
-class UnicodeCurrency\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && Prop.isCurrency(s.charAt(from)) ? to : -1;\r
- }\r
-}\r
-\r
-/** Matches a non-currency symbol Unicode character. */\r
-class NUnicodeCurrency\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && !Prop.isCurrency(s.charAt(from)) ? to : -1;\r
- }\r
-}\r
-\r
-/** Matches a Unicode alphabetic character. */\r
-class UnicodeAlpha\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && Prop.isAlphabetic(s.charAt(from)) ? to : -1;\r
- }\r
-}\r
-\r
-/** Matches a non-alphabetic Unicode character. */\r
-class NUnicodeAlpha\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && !Prop.isAlphabetic(s.charAt(from)) ? to : -1;\r
- }\r
-}\r
-\r
-/** Matches an upper case Unicode character. */\r
-class UnicodeUpper\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && isUpper(s.charAt(from)) ? to : -1;\r
- }\r
-\r
- final boolean isUpper(char c)\r
- {\r
- return c == CaseMgr.toUpperCase(c) && c != CaseMgr.toLowerCase(c);\r
- }\r
-}\r
-\r
-/** Matches an upper case Unicode character. */\r
-class UnicodeLower\r
- extends UniValidator\r
-{\r
- public int validate(StringLike s, int from, int to)\r
- {\r
- return from < s.length() && isLower(s.charAt(from)) ? to : -1;\r
- }\r
-\r
- final boolean isLower(char c)\r
- {\r
- return c != CaseMgr.toUpperCase(c) && c == CaseMgr.toLowerCase(c);\r
- }\r
-}\r
-\r
-/**\r
- Regex provides the parser which constructs the linked list of\r
- Pattern classes from a String.\r
- <p>\r
- For the purpose of this documentation, the fact that java interprets the\r
- backslash will be ignored. In practice, however, you will need a\r
- double backslash to obtain a string that contains a single backslash\r
- character. Thus, the example pattern "\b" should really be typed\r
- as "\\b" inside java code.\r
- <p>\r
- Note that Regex is part of package "com.stevesoft.pat".\r
- To use it, simply import\r
- com.stevesoft.pat.Regex at the top of your file.\r
- <p>\r
- Regex is made with a constructor that takes a String that defines\r
- the regular expression. Thus, for example\r
- <pre>\r
- Regex r = new Regex("[a-c]*");\r
- </pre>\r
- matches any number of characters so long as the are 'a', 'b', or 'c').\r
- <p>\r
- To attempt to match the Pattern to a given string, you can use either\r
- the search(String) member function, or the matchAt(String,int position)\r
- member function. These functions return a boolean which tells you\r
- whether or not the thing worked, and sets the methods "charsMatched()"\r
- and "matchedFrom()" in the Regex object appropriately.\r
- <p>\r
- The portion of the string before the match can be obtained by the\r
- left() member, and the portion after the match can be obtained\r
- by the right() member.\r
- <p>\r
- Essentially, this package implements a syntax that is very much\r
- like the perl 5 regular expression syntax.\r
-\r
- Longer example:\r
- <pre>\r
- Regex r = new Regex("x(a|b)y");\r
- r.matchAt("xay",0);\r
- System.out.println("sub = "+r.stringMatched(1));\r
- </pre>\r
- The above would print "sub = a".\r
- <pre>\r
- r.left() // would return "x"\r
- r.right() // would return "y"\r
- </pre>\r
- <p>\r
- Differences between this package and perl5:<br>\r
- The extended Pattern for setting flags, is now supported,\r
- but the flags are different. "(?i)" tells the pattern to\r
- ignore case, "(?Q)" sets the "dontMatchInQuotes" flag, and\r
- "(?iQ)" sets them both. You can change the escape character.\r
- The pattern <pre>(?e=#)#d+</pre> is the same as <pre>\d+</pre>,\r
- but note that the sequence <pre>(?e=#)</pre> <b>must</b> occur\r
- at the very beginning of the pattern. There may be other small\r
- differences as well. I will either make my package conform\r
- or note them as I become aware of them.\r
- <p>\r
- This package supports additional patterns not in perl5:\r
- <center>\r
- <table border=1>\r
- <tr><td>(?@())</td><td>Group</td><td>This matches all characters between\r
- the '(' character and the balancing ')' character. Thus, it will\r
- match "()" as well as "(())". The balancing characters are\r
- arbitrary, thus (?@{}) matches on "{}" and "{{}}".</td>\r
- <tr><td>(?<1)</td><td>Backup</td><td>Moves the pointer backwards within the text.\r
- This allows you to make a "look behind." It fails if it\r
- attempts to move to a position before the beginning of the string.\r
- "x(?<1)" is equivalent to "(?=x)". The number, 1 in this example,\r
- is the number of characters to move backwards.</td>\r
- </table>\r
- </center>\r
- </dl>\r
- @author Steven R. Brandt\r
- @version package com.stevesoft.pat, release 1.5.3\r
- @see Pattern\r
- */\r
-public class Regex\r
- extends RegRes implements FilenameFilter\r
-{\r
- /** BackRefOffset gives the identity number of the first\r
- pattern. Version 1.0 used zero, version 1.1 uses 1 to be\r
- more compatible with perl. */\r
- static int BackRefOffset = 1;\r
- private static Pattern none = new NoPattern();\r
- Pattern thePattern = none;\r
- patInt minMatch = new patInt(0);\r
-\r
- static Hashtable validators = new Hashtable();\r
- static\r
- {\r
- define("p", "(?>1)", new UnicodePunct());\r
- define("P", "(?>1)", new NUnicodePunct());\r
- define("s", "(?>1)", new UnicodeWhite());\r
- define("S", "(?>1)", new NUnicodeWhite());\r
- define("w", "(?>1)", new UnicodeW());\r
- define("W", "(?>1)", new NUnicodeW());\r
- define("d", "(?>1)", new UnicodeDigit());\r
- define("D", "(?>1)", new NUnicodeDigit());\r
- define("m", "(?>1)", new UnicodeMath());\r
- define("M", "(?>1)", new NUnicodeMath());\r
- define("c", "(?>1)", new UnicodeCurrency());\r
- define("C", "(?>1)", new NUnicodeCurrency());\r
- define("a", "(?>1)", new UnicodeAlpha());\r
- define("A", "(?>1)", new NUnicodeAlpha());\r
- define("uc", "(?>1)", new UnicodeUpper());\r
- define("lc", "(?>1)", new UnicodeLower());\r
- }\r
-\r
- /** Set the dontMatch in quotes flag. */\r
- public void setDontMatchInQuotes(boolean b)\r
- {\r
- dontMatchInQuotes = b;\r
- }\r
-\r
- /** Find out if the dontMatchInQuotes flag is enabled. */\r
- public boolean getDontMatchInQuotes()\r
- {\r
- return dontMatchInQuotes;\r
- }\r
-\r
- boolean dontMatchInQuotes = false;\r
-\r
- /** Set the state of the ignoreCase flag. If set to true, then\r
- the pattern matcher will ignore case when searching for a\r
- match. */\r
- public void setIgnoreCase(boolean b)\r
- {\r
- ignoreCase = b;\r
- }\r
-\r
- /** Get the state of the ignoreCase flag. Returns true if we\r
- are ignoring the case of the pattern, false otherwise. */\r
- public boolean getIgnoreCase()\r
- {\r
- return ignoreCase;\r
- }\r
-\r
- boolean ignoreCase = false;\r
-\r
- static boolean defaultMFlag = false;\r
- /** Set the default value of the m flag. If it\r
- is set to true, then the MFlag will be on\r
- for any regex search executed. */\r
- public static void setDefaultMFlag(boolean mFlag)\r
- {\r
- defaultMFlag = mFlag;\r
- }\r
-\r
- /** Get the default value of the m flag. If it\r
- is set to true, then the MFlag will be on\r
- for any regex search executed. */\r
- public static boolean getDefaultMFlag()\r
- {\r
- return defaultMFlag;\r
- }\r
-\r
- /** Initializes the object without a Pattern. To supply a Pattern\r
- use compile(String s).\r
- @see com.stevesoft.pat.Regex#compile(java.lang.String)\r
- */\r
- public Regex()\r
- {}\r
-\r
- /** Create and compile a Regex, but do not throw any exceptions.\r
- If you wish to have exceptions thrown for syntax errors,\r
- you must use the Regex(void) constructor to create the\r
- Regex object, and then call the compile method. Therefore, you\r
- should only call this method when you know your pattern is right.\r
- I will probably become more like\r
- @see com.stevesoft.pat.Regex#search(java.lang.String)\r
- @see com.stevesoft.pat.Regex#compile(java.lang.String)\r
- */\r
- public Regex(String s)\r
- {\r
- try\r
- {\r
- compile(s);\r
- }\r
- catch (RegSyntax rs)\r
- {}\r
- }\r
-\r
- ReplaceRule rep = null;\r
- /** Create and compile both a Regex and a ReplaceRule.\r
- @see com.stevesoft.pat.ReplaceRule\r
- @see com.stevesoft.pat.Regex#compile(java.lang.String)\r
- */\r
- public Regex(String s, String rp)\r
- {\r
- this(s);\r
- rep = ReplaceRule.perlCode(rp);\r
- }\r
-\r
- /** Create and compile a Regex, but give it the ReplaceRule\r
- specified. This allows the user finer control of the\r
- Replacement process, if that is desired.\r
- @see com.stevesoft.pat.ReplaceRule\r
- @see com.stevesoft.pat.Regex#compile(java.lang.String)\r
- */\r
- public Regex(String s, ReplaceRule rp)\r
- {\r
- this(s);\r
- rep = rp;\r
- }\r
-\r
- /** Change the ReplaceRule of this Regex by compiling\r
- a new one using String rp. */\r
- public void setReplaceRule(String rp)\r
- {\r
- rep = ReplaceRule.perlCode(rp);\r
- repr = null; // Clear Replacer history\r
- }\r
-\r
- /** Change the ReplaceRule of this Regex to rp. */\r
- public void setReplaceRule(ReplaceRule rp)\r
- {\r
- rep = rp;\r
- }\r
-\r
- /** Test to see if a custom defined rule exists.\r
- @see com.stevesoft.pat#define(java.lang.String,java.lang.String,Validator)\r
- */\r
- public static boolean isDefined(String nm)\r
- {\r
- return validators.get(nm) != null;\r
- }\r
-\r
- /** Removes a custom defined rule.\r
- @see com.stevesoft.pat#define(java.lang.String,java.lang.String,Validator)\r
- */\r
- public static void undefine(String nm)\r
- {\r
- validators.remove(nm);\r
- }\r
-\r
- /** Defines a method to create a new rule. See test/deriv2.java\r
- and test/deriv3.java for examples of how to use it. */\r
- public static void define(String nm, String pat, Validator v)\r
- {\r
- v.pattern = pat;\r
- validators.put(nm, v);\r
- }\r
-\r
- /** Defines a shorthand for a pattern. The pattern will be\r
- invoked by a string that has the form "(??"+nm+")".\r
- */\r
- public static void define(String nm, String pat)\r
- {\r
- validators.put(nm, pat);\r
- }\r
-\r
- /** Get the current ReplaceRule. */\r
- public ReplaceRule getReplaceRule()\r
- {\r
- return rep;\r
- }\r
-\r
- Replacer repr = null;\r
- final Replacer _getReplacer()\r
- {\r
- return repr == null ? repr = new Replacer() : repr;\r
- }\r
-\r
- public Replacer getReplacer()\r
- {\r
- if (repr == null)\r
- {\r
- repr = new Replacer();\r
- }\r
- repr.rh.me = this;\r
- repr.rh.prev = null;\r
- return repr;\r
- }\r
-\r
- /** Replace the first occurence of this pattern in String s\r
- according to the ReplaceRule.\r
- @see com.stevesoft.pat.ReplaceRule\r
- @see com.stevesoft.pat.Regex#getReplaceRule()\r
- */\r
- public String replaceFirst(String s)\r
- {\r
- return _getReplacer().replaceFirstRegion(s, this, 0, s.length()).toString();\r
- }\r
-\r
- /** Replace the first occurence of this pattern in String s\r
- beginning with position pos according to the ReplaceRule.\r
- @see com.stevesoft.pat.ReplaceRule\r
- @see com.stevesoft.pat.Regex#getReplaceRule()\r
- */\r
- public String replaceFirstFrom(String s, int pos)\r
- {\r
- return _getReplacer().replaceFirstRegion(s, this, pos, s.length()).toString();\r
- }\r
-\r
- /** Replace the first occurence of this pattern in String s\r
- beginning with position start and ending with end\r
- according to the ReplaceRule.\r
- @see com.stevesoft.pat.ReplaceRule\r
- @see com.stevesoft.pat.Regex#getReplaceRule()\r
- */\r
- public String replaceFirstRegion(String s, int start, int end)\r
- {\r
- return _getReplacer().replaceFirstRegion(s, this, start, end).toString();\r
- }\r
-\r
- /** Replace all occurences of this pattern in String s\r
- according to the ReplaceRule.\r
- @see com.stevesoft.pat.ReplaceRule\r
- @see com.stevesoft.pat.Regex#getReplaceRule()\r
- */\r
- public String replaceAll(String s)\r
- {\r
- return _getReplacer().replaceAllRegion(s, this, 0, s.length()).toString();\r
- }\r
-\r
- public StringLike replaceAll(StringLike s)\r
- {\r
- return _getReplacer().replaceAllRegion(s, this, 0, s.length());\r
- }\r
-\r
- /** Replace all occurences of this pattern in String s\r
- beginning with position pos according to the ReplaceRule.\r
- @see com.stevesoft.pat.ReplaceRule\r
- @see com.stevesoft.pat.Regex#getReplaceRule()\r
- */\r
- public String replaceAllFrom(String s, int pos)\r
- {\r
- return _getReplacer().replaceAllRegion(s, this, pos, s.length()).toString();\r
- }\r
-\r
- /** Replace all occurences of this pattern in String s\r
- beginning with position start and ending with end\r
- according to the ReplaceRule.\r
- @see com.stevesoft.pat.ReplaceRule\r
- @see com.stevesoft.pat.Regex#getReplaceRule()\r
- */\r
- public String replaceAllRegion(String s, int start, int end)\r
- {\r
- return _getReplacer().replaceAllRegion(s, this, start, end).toString();\r
- }\r
-\r
- /** Essentially clones the Regex object */\r
- public Regex(Regex r)\r
- {\r
- super( (RegRes) r);\r
- dontMatchInQuotes = r.dontMatchInQuotes;\r
- esc = r.esc;\r
- ignoreCase = r.ignoreCase;\r
- gFlag = r.gFlag;\r
- if (r.rep == null)\r
- {\r
- rep = null;\r
- }\r
- else\r
- {\r
- rep = (ReplaceRule) r.rep.clone();\r
- }\r
- /* try {\r
- compile(r.toString());\r
- } catch(RegSyntax r_) {} */\r
- thePattern = r.thePattern.clone(new Hashtable());\r
- minMatch = r.minMatch;\r
- skipper = r.skipper;\r
- }\r
-\r
- /** By default,\r
- the escape character is the backslash, but you can\r
- make it anything you want by setting this variable. */\r
- public char esc = Pattern.ESC;\r
- /** This method compiles a regular expression, making it\r
- possible to call the search or matchAt methods.\r
- @exception com.stevesoft.pat.RegSyntax\r
- is thrown if a syntax error is encountered\r
- in the pattern.\r
- For example, "x{3,1}" or "*a" are not valid\r
- patterns.\r
- @see com.stevesoft.pat.Regex#search\r
- @see com.stevesoft.pat.Regex#matchAt\r
- */\r
- public void compile(String prepat)\r
- throws RegSyntax\r
- {\r
- String postpat = parsePerl.codify(prepat, true);\r
- String pat = postpat == null ? prepat : postpat;\r
- minMatch = null;\r
- ignoreCase = false;\r
- dontMatchInQuotes = false;\r
- Rthings mk = new Rthings(this);\r
- int offset = mk.val;\r
- String newpat = pat;\r
- thePattern = none;\r
- p = null;\r
- or = null;\r
- minMatch = new patInt(0);\r
- StrPos sp = new StrPos(pat, 0);\r
- if (sp.incMatch("(?e="))\r
- {\r
- char newEsc = sp.c;\r
- sp.inc();\r
- if (sp.match(')'))\r
- {\r
- newpat = reEscape(pat.substring(6),\r
- newEsc, Pattern.ESC);\r
- }\r
- }\r
- else if (esc != Pattern.ESC)\r
- {\r
- newpat = reEscape(pat, esc, Pattern.ESC);\r
- }\r
- thePattern = _compile(newpat, mk);\r
- numSubs_ = mk.val - offset;\r
- mk.set(this);\r
- }\r
-\r
- /* If a Regex is compared against a Regex, a check is\r
- done to see that the patterns are equal as well as\r
- the most recent match. If a Regex is compare with\r
- a RegRes, only the result of the most recent match\r
- is compared. */\r
- public boolean equals(Object o)\r
- {\r
- if (o instanceof Regex)\r
- {\r
- if (toString().equals(o.toString()))\r
- {\r
- return super.equals(o);\r
- }\r
- else\r
- {\r
- return false;\r
- }\r
- }\r
- else\r
- {\r
- return super.equals(o);\r
- }\r
- }\r
-\r
- /** A clone by any other name would smell as sweet. */\r
- public Object clone()\r
- {\r
- return new Regex(this);\r
- }\r
-\r
- /** Return a clone of the underlying RegRes object. */\r
- public RegRes result()\r
- {\r
- return (RegRes)super.clone();\r
- }\r
-\r
- // prep sets global variables of class\r
- // Pattern so that it can access them\r
- // during an attempt at a match\r
- Pthings pt = new Pthings();\r
- final Pthings prep(StringLike s)\r
- {\r
- //if(gFlag)\r
- pt.lastPos = matchedTo();\r
- if (pt.lastPos < 0)\r
- {\r
- pt.lastPos = 0;\r
- }\r
- if ( (s == null ? null : s.unwrap()) != (src == null ? null : s.unwrap()))\r
- {\r
- pt.lastPos = 0;\r
- }\r
- src = s;\r
- pt.dotDoesntMatchCR = dotDoesntMatchCR && (!sFlag);\r
- pt.mFlag = (mFlag | defaultMFlag);\r
- pt.ignoreCase = ignoreCase;\r
- pt.no_check = false;\r
- if (pt.marks != null)\r
- {\r
- for (int i = 0; i < pt.marks.length; i++)\r
- {\r
- pt.marks[i] = -1;\r
- }\r
- }\r
- pt.marks = null;\r
- pt.nMarks = numSubs_;\r
- pt.src = s;\r
- if (dontMatchInQuotes)\r
- {\r
- setCbits(s, pt);\r
- }\r
- else\r
- {\r
- pt.cbits = null;\r
- }\r
- return pt;\r
- }\r
-\r
- /** Attempt to match a Pattern beginning\r
- at a specified location within the string.\r
- @see com.stevesoft.pat.Regex#search\r
- */\r
- public boolean matchAt(String s, int start_pos)\r
- {\r
- return _search(s, start_pos, start_pos);\r
- }\r
-\r
- /** Attempt to match a Pattern beginning\r
- at a specified location within the StringLike.\r
- @see com.stevesoft.pat.Regex#search\r
- */\r
- public boolean matchAt(StringLike s, int start_pos)\r
- {\r
- return _search(s, start_pos, start_pos);\r
- }\r
-\r
- /** Search through a String for the first\r
- occurrence of a match.\r
- @see com.stevesoft.pat.Regex#searchFrom\r
- @see com.stevesoft.pat.Regex#matchAt\r
- */\r
- public boolean search(String s)\r
- {\r
- if (s == null)\r
- {\r
- throw new NullPointerException("Null String Given to Regex.search");\r
- }\r
- return _search(s, 0, s.length());\r
- }\r
-\r
- public boolean search(StringLike sl)\r
- {\r
- if (sl == null)\r
- {\r
- throw new NullPointerException("Null StringLike Given to Regex.search");\r
- }\r
- return _search(sl, 0, sl.length());\r
- }\r
-\r
- public boolean reverseSearch(String s)\r
- {\r
- if (s == null)\r
- {\r
- throw new NullPointerException("Null String Given to Regex.reverseSearch");\r
- }\r
- return _reverseSearch(s, 0, s.length());\r
- }\r
-\r
- public boolean reverseSearch(StringLike sl)\r
- {\r
- if (sl == null)\r
- {\r
- throw new NullPointerException(\r
- "Null StringLike Given to Regex.reverseSearch");\r
- }\r
- return _reverseSearch(sl, 0, sl.length());\r
- }\r
-\r
- /** Search through a String for the first\r
- occurence of a match, but start at position <pre>start</pre>*/\r
- public boolean searchFrom(String s, int start)\r
- {\r
- if (s == null)\r
- {\r
- throw new NullPointerException("Null String Given to Regex.searchFrom");\r
- }\r
- return _search(s, start, s.length());\r
- }\r
-\r
- public boolean searchFrom(StringLike s, int start)\r
- {\r
- if (s == null)\r
- {\r
- throw new NullPointerException("Null String Given to Regex.searchFrom");\r
- }\r
- return _search(s, start, s.length());\r
- }\r
-\r
- /** Search through a region of a String\r
- for the first occurence of a match. */\r
- public boolean searchRegion(String s, int start, int end)\r
- {\r
- if (s == null)\r
- {\r
- throw new NullPointerException("Null String Given to Regex.searchRegion");\r
- }\r
- return _search(s, start, end);\r
- }\r
-\r
- /** Set this to change the default behavior of the "." pattern.\r
- By default it now matches perl's behavior and fails to\r
- match the '\n' character. */\r
- public static boolean dotDoesntMatchCR = true;\r
- StringLike gFlags;\r
- int gFlagto = 0;\r
- boolean gFlag = false;\r
- /** Set the 'g' flag */\r
- public void setGFlag(boolean b)\r
- {\r
- gFlag = b;\r
- }\r
-\r
- /** Get the state of the 'g' flag. */\r
- public boolean getGFlag()\r
- {\r
- return gFlag;\r
- }\r
-\r
- boolean sFlag = false;\r
- /** Get the state of the sFlag */\r
- public boolean getSFlag()\r
- {\r
- return sFlag;\r
- }\r
-\r
- boolean mFlag = false;\r
- /** Get the state of the sFlag */\r
- public boolean getMFlag()\r
- {\r
- return mFlag;\r
- }\r
-\r
- final boolean _search(String s, int start, int end)\r
- {\r
- return _search(new StringWrap(s), start, end);\r
- }\r
-\r
- final boolean _search(StringLike s, int start, int end)\r
- {\r
- if (gFlag && gFlagto > 0 && gFlags != null && s.unwrap() == gFlags.unwrap())\r
- {\r
- start = gFlagto;\r
- }\r
- gFlags = null;\r
-\r
- Pthings pt = prep(s);\r
-\r
- int up = (minMatch == null ? end : end - minMatch.i);\r
-\r
- if (up < start && end >= start)\r
- {\r
- up = start;\r
- }\r
-\r
- if (skipper == null)\r
- {\r
- for (int i = start; i <= up; i++)\r
- {\r
- charsMatched_ = thePattern.matchAt(s, i, pt);\r
- if (charsMatched_ >= 0)\r
- {\r
- matchFrom_ = thePattern.mfrom;\r
- marks = pt.marks;\r
- gFlagto = matchFrom_ + charsMatched_;\r
- gFlags = s;\r
- return didMatch_ = true;\r
- }\r
- }\r
- }\r
- else\r
- {\r
- pt.no_check = true;\r
- for (int i = start; i <= up; i++)\r
- {\r
- i = skipper.find(src, i, up);\r
- if (i < 0)\r
- {\r
- charsMatched_ = matchFrom_ = -1;\r
- return didMatch_ = false;\r
- }\r
- charsMatched_ = thePattern.matchAt(s, i, pt);\r
- if (charsMatched_ >= 0)\r
- {\r
- matchFrom_ = thePattern.mfrom;\r
- marks = pt.marks;\r
- gFlagto = matchFrom_ + charsMatched_;\r
- gFlags = s;\r
- return didMatch_ = true;\r
- }\r
- }\r
- }\r
- return didMatch_ = false;\r
- }\r
-\r
- /*final boolean _search(LongStringLike s,long start,long end) {\r
- if(gFlag && gFlagto > 0 && s==gFlags)\r
- start = gFlagto;\r
- gFlags = null;\r
-\r
- Pthings pt=prep(s);\r
-\r
- int up = end;//(minMatch == null ? end : end-minMatch.i);\r
-\r
- if(up < start && end >= start) up = start;\r
-\r
- if(skipper == null) {\r
- for(long i=start;i<=up;i++) {\r
- charsMatched_ = thePattern.matchAt(s,i,pt);\r
- if(charsMatched_ >= 0) {\r
- matchFrom_ = thePattern.mfrom;\r
- marks = pt.marks;\r
- gFlagto = matchFrom_+charsMatched_;\r
- return didMatch_=true;\r
- }\r
- }\r
- } else {\r
- pt.no_check = true;\r
- for(long i=start;i<=up;i++) {\r
- i = skipper.find(src,i,up);\r
- if(i<0) {\r
- charsMatched_ = matchFrom_ = -1;\r
- return didMatch_ = false;\r
- }\r
- charsMatched_ = thePattern.matchAt(s,i,pt);\r
- if(charsMatched_ >= 0) {\r
- matchFrom_ = thePattern.mfrom;\r
- marks = pt.marks;\r
- gFlagto = matchFrom_+charsMatched_;\r
- gFlags = s;\r
- return didMatch_=true;\r
- } else {\r
- i = s.adjustIndex(i);\r
- up = s.adjustEnd(i);\r
- }\r
- }\r
- }\r
- return didMatch_=false;\r
- }*/\r
-\r
- boolean _reverseSearch(String s, int start, int end)\r
- {\r
- return _reverseSearch(new StringWrap(s), start, end);\r
- }\r
-\r
- boolean _reverseSearch(StringLike s, int start, int end)\r
- {\r
- if (gFlag && gFlagto > 0 && s.unwrap() == gFlags.unwrap())\r
- {\r
- end = gFlagto;\r
- }\r
- gFlags = null;\r
- Pthings pt = prep(s);\r
- for (int i = end; i >= start; i--)\r
- {\r
- charsMatched_ = thePattern.matchAt(s, i, pt);\r
- if (charsMatched_ >= 0)\r
- {\r
- matchFrom_ = thePattern.mfrom;\r
- marks = pt.marks;\r
- gFlagto = matchFrom_ - 1;\r
- gFlags = s;\r
- return didMatch_ = true;\r
- }\r
- }\r
- return didMatch_ = false;\r
- }\r
-\r
- // This routine sets the cbits variable\r
- // of class Pattern. Cbits is true for\r
- // the bit corresponding to a character inside\r
- // a set of quotes.\r
- static StringLike lasts = null;\r
- static BitSet lastbs = null;\r
- static void setCbits(StringLike s, Pthings pt)\r
- {\r
- if (s == lasts)\r
- {\r
- pt.cbits = lastbs;\r
- return;\r
- }\r
- BitSet bs = new BitSet(s.length());\r
- char qc = ' ';\r
- boolean setBit = false;\r
- for (int i = 0; i < s.length(); i++)\r
- {\r
- if (setBit)\r
- {\r
- bs.set(i);\r
- }\r
- char c = s.charAt(i);\r
- if (!setBit && c == '"')\r
- {\r
- qc = c;\r
- setBit = true;\r
- bs.set(i);\r
- }\r
- else if (!setBit && c == '\'')\r
- {\r
- qc = c;\r
- setBit = true;\r
- bs.set(i);\r
- }\r
- else if (setBit && c == qc)\r
- {\r
- setBit = false;\r
- }\r
- else if (setBit && c == '\\' && i + 1 < s.length())\r
- {\r
- i++;\r
- if (setBit)\r
- {\r
- bs.set(i);\r
- }\r
- }\r
- }\r
- pt.cbits = lastbs = bs;\r
- lasts = s;\r
- }\r
-\r
- // Wanted user to over-ride this in alpha version,\r
- // but it wasn't really necessary because of this trick:\r
- Regex newRegex()\r
- {\r
- try\r
- {\r
- return (Regex) getClass().newInstance();\r
- }\r
- catch (InstantiationException ie)\r
- {\r
- return null;\r
- }\r
- catch (IllegalAccessException iae)\r
- {\r
- return null;\r
- }\r
- }\r
-\r
- /** Only needed for creating your own extensions of\r
- Regex. This method adds the next Pattern in the chain\r
- of patterns or sets the Pattern if it is the first call. */\r
- protected void add(Pattern p2)\r
- {\r
- if (p == null)\r
- {\r
- p = p2;\r
- }\r
- else\r
- {\r
- p.add(p2);\r
- p2 = p;\r
- }\r
- }\r
-\r
- /** You only need to use this method if you are creating\r
- your own extentions to Regex.\r
- compile1 compiles one Pattern element, it can be\r
- over-ridden to allow the Regex compiler to understand\r
- new syntax. See deriv.java for an example. This routine\r
- is the heart of class Regex. Rthings has one integer\r
- member called intValue, it is used to keep track of the number\r
- of ()'s in the Pattern.\r
- @exception com.stevesoft.pat.RegSyntax is thrown when a nonsensensical\r
- pattern is supplied. For example, a pattern beginning\r
- with *. */\r
- protected void compile1(StrPos sp, Rthings mk)\r
- throws RegSyntax\r
- {\r
- if (sp.match('['))\r
- {\r
- sp.inc();\r
- add(matchBracket(sp));\r
- }\r
- else if (sp.match('|'))\r
- {\r
- if (or == null)\r
- {\r
- or = new Or();\r
- }\r
- if (p == null)\r
- {\r
- p = new NullPattern();\r
- }\r
- or.addOr(p);\r
- p = null;\r
- }\r
- else if (sp.incMatch("(?<"))\r
- {\r
- patInt i = sp.getPatInt();\r
- if (i == null)\r
- {\r
- RegSyntaxError.endItAll("No int after (?<");\r
- }\r
- add(new Backup(i.intValue()));\r
- if (!sp.match(')'))\r
- {\r
- RegSyntaxError.endItAll("No ) after (?<");\r
- }\r
- }\r
- else if (sp.incMatch("(?>"))\r
- {\r
- patInt i = sp.getPatInt();\r
- if (i == null)\r
- {\r
- RegSyntaxError.endItAll("No int after (?>");\r
- }\r
- add(new Backup( -i.intValue()));\r
- if (!sp.match(')'))\r
- {\r
- RegSyntaxError.endItAll("No ) after (?<");\r
- }\r
- }\r
- else if (sp.incMatch("(?@"))\r
- {\r
- char op = sp.c;\r
- sp.inc();\r
- char cl = sp.c;\r
- sp.inc();\r
- if (!sp.match(')'))\r
- {\r
- RegSyntaxError.endItAll(\r
- "(?@ does not have closing paren");\r
- }\r
- add(new Group(op, cl));\r
- }\r
- else if (sp.incMatch("(?#"))\r
- {\r
- while (!sp.match(')'))\r
- {\r
- sp.inc();\r
- }\r
- }\r
- else if (sp.dontMatch && sp.c == 'w')\r
- {\r
- //Regex r = new Regex();\r
- //r._compile("[a-zA-Z0-9_]",mk);\r
- //add(new Goop("\\w",r.thePattern));\r
- Bracket b = new Bracket(false);\r
- b.addOr(new Range('a', 'z'));\r
- b.addOr(new Range('A', 'Z'));\r
- b.addOr(new Range('0', '9'));\r
- b.addOr(new oneChar('_'));\r
- add(b);\r
- }\r
- else if (sp.dontMatch && sp.c == 'G')\r
- {\r
- add(new BackG());\r
- }\r
- else if (sp.dontMatch && sp.c == 's')\r
- {\r
- //Regex r = new Regex();\r
- //r._compile("[ \t\n\r\b]",mk);\r
- //add(new Goop("\\s",r.thePattern));\r
- Bracket b = new Bracket(false);\r
- b.addOr(new oneChar( (char) 32));\r
- b.addOr(new Range( (char) 8, (char) 10));\r
- b.addOr(new oneChar( (char) 13));\r
- add(b);\r
- }\r
- else if (sp.dontMatch && sp.c == 'd')\r
- {\r
- //Regex r = new Regex();\r
- //r._compile("[0-9]",mk);\r
- //add(new Goop("\\d",r.thePattern));\r
- Range digit = new Range('0', '9');\r
- digit.printBrackets = true;\r
- add(digit);\r
- }\r
- else if (sp.dontMatch && sp.c == 'W')\r
- {\r
- //Regex r = new Regex();\r
- //r._compile("[^a-zA-Z0-9_]",mk);\r
- //add(new Goop("\\W",r.thePattern));\r
- Bracket b = new Bracket(true);\r
- b.addOr(new Range('a', 'z'));\r
- b.addOr(new Range('A', 'Z'));\r
- b.addOr(new Range('0', '9'));\r
- b.addOr(new oneChar('_'));\r
- add(b);\r
- }\r
- else if (sp.dontMatch && sp.c == 'S')\r
- {\r
- //Regex r = new Regex();\r
- //r._compile("[^ \t\n\r\b]",mk);\r
- //add(new Goop("\\S",r.thePattern));\r
- Bracket b = new Bracket(true);\r
- b.addOr(new oneChar( (char) 32));\r
- b.addOr(new Range( (char) 8, (char) 10));\r
- b.addOr(new oneChar( (char) 13));\r
- add(b);\r
- }\r
- else if (sp.dontMatch && sp.c == 'D')\r
- {\r
- //Regex r = new Regex();\r
- //r._compile("[^0-9]",mk);\r
- //add(new Goop("\\D",r.thePattern));\r
- Bracket b = new Bracket(true);\r
- b.addOr(new Range('0', '9'));\r
- add(b);\r
- }\r
- else if (sp.dontMatch && sp.c == 'B')\r
- {\r
- Regex r = new Regex();\r
- r._compile("(?!" + back_slash + "b)", mk);\r
- add(r.thePattern);\r
- }\r
- else if (isOctalString(sp))\r
- {\r
- int d = sp.c - '0';\r
- sp.inc();\r
- d = 8 * d + sp.c - '0';\r
- StrPos sp2 = new StrPos(sp);\r
- sp2.inc();\r
- if (isOctalDigit(sp2, false))\r
- {\r
- sp.inc();\r
- d = 8 * d + sp.c - '0';\r
- }\r
- add(new oneChar( (char) d));\r
- }\r
- else if (sp.dontMatch && sp.c >= '1' && sp.c <= '9')\r
- {\r
- int iv = sp.c - '0';\r
- StrPos s2 = new StrPos(sp);\r
- s2.inc();\r
- if (!s2.dontMatch && s2.c >= '0' && s2.c <= '9')\r
- {\r
- iv = 10 * iv + (s2.c - '0');\r
- sp.inc();\r
- }\r
- add(new BackMatch(iv));\r
- }\r
- else if (sp.dontMatch && sp.c == 'b')\r
- {\r
- add(new Boundary());\r
- }\r
- else if (sp.match('\b'))\r
- {\r
- add(new Boundary());\r
- }\r
- else if (sp.match('$'))\r
- {\r
- add(new End(true));\r
- }\r
- else if (sp.dontMatch && sp.c == 'Z')\r
- {\r
- add(new End(false));\r
- }\r
- else if (sp.match('.'))\r
- {\r
- add(new Any());\r
- }\r
- else if (sp.incMatch("(??"))\r
- {\r
- StringBuffer sb = new StringBuffer();\r
- StringBuffer sb2 = new StringBuffer();\r
- while (!sp.match(')') && !sp.match(':'))\r
- {\r
- sb.append(sp.c);\r
- sp.inc();\r
- }\r
- if (sp.incMatch(":"))\r
- {\r
- while (!sp.match(')'))\r
- {\r
- sb2.append(sp.c);\r
- sp.inc();\r
- }\r
- }\r
- String sbs = sb.toString();\r
- if (validators.get(sbs) instanceof String)\r
- {\r
- String pat = (String) validators.get(sbs);\r
- Regex r = newRegex();\r
- Rthings rth = new Rthings(this);\r
- rth.noBackRefs = true;\r
- r._compile(pat, rth);\r
- add(r.thePattern);\r
- }\r
- else\r
- {\r
- Custom cm = new Custom(sb.toString());\r
- if (cm.v != null)\r
- {\r
- Validator v2 = cm.v.arg(sb2.toString());\r
- if (v2 != null)\r
- {\r
- v2.argsave = sb2.toString();\r
- String p = cm.v.pattern;\r
- cm.v = v2;\r
- v2.pattern = p;\r
- }\r
- Regex r = newRegex();\r
- Rthings rth = new Rthings(this);\r
- rth.noBackRefs = true;\r
- r._compile(cm.v.pattern, rth);\r
- cm.sub = r.thePattern;\r
- cm.sub.add(new CustomEndpoint(cm));\r
- cm.sub.setParent(cm);\r
- add(cm);\r
- }\r
- }\r
- }\r
- else if (sp.match('('))\r
- {\r
- mk.parenLevel++;\r
- Regex r = newRegex();\r
- // r.or = new Or();\r
- sp.inc();\r
- if (sp.incMatch("?:"))\r
- {\r
- r.or = new Or();\r
- }\r
- else if (sp.incMatch("?="))\r
- {\r
- r.or = new lookAhead(false);\r
- }\r
- else if (sp.incMatch("?!"))\r
- {\r
- r.or = new lookAhead(true);\r
- }\r
- else if (sp.match('?'))\r
- {\r
- sp.inc();\r
- do\r
- {\r
- if (sp.c == 'i')\r
- {\r
- mk.ignoreCase = true;\r
- }\r
- if (sp.c == 'Q')\r
- {\r
- mk.dontMatchInQuotes = true;\r
- }\r
- if (sp.c == 'o')\r
- {\r
- mk.optimizeMe = true;\r
- }\r
- if (sp.c == 'g')\r
- {\r
- mk.gFlag = true;\r
- }\r
- if (sp.c == 's')\r
- {\r
- mk.sFlag = true;\r
- }\r
- if (sp.c == 'm')\r
- {\r
- mk.mFlag = true;\r
- }\r
- sp.inc();\r
- }\r
- while (!sp.match(')') && !sp.eos);\r
- r = null;\r
- mk.parenLevel--;\r
- if (sp.eos) //throw new RegSyntax\r
- {\r
- RegSyntaxError.endItAll("Unclosed ()");\r
- }\r
- }\r
- else\r
- { // just ordinary parenthesis\r
- r.or = mk.noBackRefs ? new Or() : new OrMark(mk.val++);\r
- }\r
- if (r != null)\r
- {\r
- add(r._compile(sp, mk));\r
- }\r
- }\r
- else if (sp.match('^'))\r
- {\r
- add(new Start(true));\r
- }\r
- else if (sp.dontMatch && sp.c == 'A')\r
- {\r
- add(new Start(false));\r
- }\r
- else if (sp.match('*'))\r
- {\r
- addMulti(new patInt(0), new patInf());\r
- }\r
- else if (sp.match('+'))\r
- {\r
- addMulti(new patInt(1), new patInf());\r
- }\r
- else if (sp.match('?'))\r
- {\r
- addMulti(new patInt(0), new patInt(1));\r
- }\r
- else if (sp.match('{'))\r
- {\r
- boolean bad = false;\r
- StrPos sp2 = new StrPos(sp);\r
- //StringBuffer sb = new StringBuffer();\r
- sp.inc();\r
- patInt i1 = sp.getPatInt();\r
- patInt i2 = null;\r
- if (sp.match('}'))\r
- {\r
- i2 = i1;\r
- }\r
- else\r
- {\r
- if (!sp.match(','))\r
- {\r
- /*\r
- RegSyntaxError.endItAll(\r
- "String \"{"+i2+\r
- "\" should be followed with , or }");\r
- */\r
- bad = true;\r
- }\r
- sp.inc();\r
- if (sp.match('}'))\r
- {\r
- i2 = new patInf();\r
- }\r
- else\r
- {\r
- i2 = sp.getPatInt();\r
- }\r
- }\r
- if (i1 == null || i2 == null)\r
- {\r
- /*\r
- throw new RegSyntax("Badly formatted Multi: "\r
- +"{"+i1+","+i2+"}"); */\r
- bad = true;\r
- }\r
- if (bad)\r
- {\r
- sp.dup(sp2);\r
- add(new oneChar(sp.c));\r
- }\r
- else\r
- {\r
- addMulti(i1, i2);\r
- }\r
- }\r
- else if (sp.escMatch('x') && next2Hex(sp))\r
- {\r
- sp.inc();\r
- int d = getHexDigit(sp);\r
- sp.inc();\r
- d = 16 * d + getHexDigit(sp);\r
- add(new oneChar( (char) d));\r
- }\r
- else if (sp.escMatch('c'))\r
- {\r
- sp.inc();\r
- if (sp.c < Ctrl.cmap.length)\r
- {\r
- add(new oneChar(Ctrl.cmap[sp.c]));\r
- }\r
- else\r
- {\r
- add(new oneChar(sp.c));\r
- }\r
- }\r
- else if (sp.escMatch('f'))\r
- {\r
- add(new oneChar( (char) 12));\r
- }\r
- else if (sp.escMatch('a'))\r
- {\r
- add(new oneChar( (char) 7));\r
- }\r
- else if (sp.escMatch('t'))\r
- {\r
- add(new oneChar('\t'));\r
- }\r
- else if (sp.escMatch('n'))\r
- {\r
- add(new oneChar('\n'));\r
- }\r
- else if (sp.escMatch('r'))\r
- {\r
- add(new oneChar('\r'));\r
- }\r
- else if (sp.escMatch('b'))\r
- {\r
- add(new oneChar('\b'));\r
- }\r
- else if (sp.escMatch('e'))\r
- {\r
- add(new oneChar( (char) 27));\r
- }\r
- else\r
- {\r
- add(new oneChar(sp.c));\r
- if (sp.match(')'))\r
- {\r
- RegSyntaxError.endItAll("Unmatched right paren in pattern");\r
- }\r
- }\r
- }\r
-\r
- // compiles all Pattern elements, internal method\r
- private Pattern _compile(String pat, Rthings mk)\r
- throws RegSyntax\r
- {\r
- minMatch = null;\r
- sFlag = mFlag = ignoreCase = gFlag = false;\r
- StrPos sp = new StrPos(pat, 0);\r
- thePattern = _compile(sp, mk);\r
- pt.marks = null;\r
- return thePattern;\r
- }\r
-\r
- Pattern p = null;\r
- Or or = null;\r
- Pattern _compile(StrPos sp, Rthings mk)\r
- throws RegSyntax\r
- {\r
- while (! (sp.eos || (or != null && sp.match(')'))))\r
- {\r
- compile1(sp, mk);\r
- sp.inc();\r
- }\r
- if (sp.match(')'))\r
- {\r
- mk.parenLevel--;\r
- }\r
- else if (sp.eos && mk.parenLevel != 0)\r
- {\r
- RegSyntaxError.endItAll("Unclosed Parenthesis! lvl=" + mk.parenLevel);\r
- }\r
- if (or != null)\r
- {\r
- if (p == null)\r
- {\r
- p = new NullPattern();\r
- }\r
- or.addOr(p);\r
- return or;\r
- }\r
- return p == null ? new NullPattern() : p;\r
- }\r
-\r
- // add a multi object to the end of the chain\r
- // which applies to the last object\r
- void addMulti(patInt i1, patInt i2)\r
- throws RegSyntax\r
- {\r
- Pattern last, last2;\r
- for (last = p; last != null && last.next != null; last = last.next)\r
- {\r
- ;\r
- }\r
- if (last == null || last == p)\r
- {\r
- last2 = null;\r
- }\r
- else\r
- {\r
- for (last2 = p; last2.next != last; last2 = last2.next)\r
- {\r
- ;\r
- }\r
- }\r
- if (last instanceof Multi && i1.intValue() == 0 &&\r
- i2.intValue() == 1)\r
- {\r
- ( (Multi) last).matchFewest = true;\r
- }\r
- else if (last instanceof FastMulti && i1.intValue() == 0 &&\r
- i2.intValue() == 1)\r
- {\r
- ( (FastMulti) last).matchFewest = true;\r
- }\r
- else if (last instanceof DotMulti && i1.intValue() == 0 &&\r
- i2.intValue() == 1)\r
- {\r
- ( (DotMulti) last).matchFewest = true;\r
- }\r
- else if (last instanceof Multi\r
- || last instanceof DotMulti\r
- || last instanceof FastMulti)\r
- {\r
- throw new RegSyntax("Syntax error.");\r
- }\r
- else if (last2 == null)\r
- {\r
- p = mkMulti(i1, i2, p);\r
- }\r
- else\r
- {\r
- last2.next = mkMulti(i1, i2, last);\r
- }\r
- }\r
-\r
- final static Pattern mkMulti(patInt lo, patInt hi, Pattern p)\r
- throws RegSyntax\r
- {\r
- if (p instanceof Any && p.next == null)\r
- {\r
- return (Pattern)new DotMulti(lo, hi);\r
- }\r
- return RegOpt.safe4fm(p) ? (Pattern)new FastMulti(lo, hi, p) :\r
- (Pattern)new Multi(lo, hi, p);\r
- }\r
-\r
- // process the bracket operator\r
- Pattern matchBracket(StrPos sp)\r
- throws RegSyntax\r
- {\r
- Bracket ret;\r
- if (sp.match('^'))\r
- {\r
- ret = new Bracket(true);\r
- sp.inc();\r
- }\r
- else\r
- {\r
- ret = new Bracket(false);\r
- }\r
- if (sp.match(']'))\r
- {\r
- //throw new RegSyntax\r
- RegSyntaxError.endItAll("Unmatched []");\r
- }\r
-\r
- while (!sp.eos && !sp.match(']'))\r
- {\r
- StrPos s1 = new StrPos(sp);\r
- s1.inc();\r
- StrPos s1_ = new StrPos(s1);\r
- s1_.inc();\r
- if (s1.match('-') && !s1_.match(']'))\r
- {\r
- StrPos s2 = new StrPos(s1);\r
- s2.inc();\r
- if (!s2.eos)\r
- {\r
- ret.addOr(new Range(sp.c, s2.c));\r
- }\r
- sp.inc();\r
- sp.inc();\r
- }\r
- else if (sp.escMatch('Q'))\r
- {\r
- sp.inc();\r
- while (!sp.escMatch('E'))\r
- {\r
- ret.addOr(new oneChar(sp.c));\r
- sp.inc();\r
- }\r
- }\r
- else if (sp.escMatch('d'))\r
- {\r
- ret.addOr(new Range('0', '9'));\r
- }\r
- else if (sp.escMatch('s'))\r
- {\r
- ret.addOr(new oneChar( (char) 32));\r
- ret.addOr(new Range( (char) 8, (char) 10));\r
- ret.addOr(new oneChar( (char) 13));\r
- }\r
- else if (sp.escMatch('w'))\r
- {\r
- ret.addOr(new Range('a', 'z'));\r
- ret.addOr(new Range('A', 'Z'));\r
- ret.addOr(new Range('0', '9'));\r
- ret.addOr(new oneChar('_'));\r
- }\r
- else if (sp.escMatch('D'))\r
- {\r
- ret.addOr(new Range( (char) 0, (char) 47));\r
- ret.addOr(new Range( (char) 58, (char) 65535));\r
- }\r
- else if (sp.escMatch('S'))\r
- {\r
- ret.addOr(new Range( (char) 0, (char) 7));\r
- ret.addOr(new Range( (char) 11, (char) 12));\r
- ret.addOr(new Range( (char) 14, (char) 31));\r
- ret.addOr(new Range( (char) 33, (char) 65535));\r
- }\r
- else if (sp.escMatch('W'))\r
- {\r
- ret.addOr(new Range( (char) 0, (char) 64));\r
- ret.addOr(new Range( (char) 91, (char) 94));\r
- ret.addOr(new oneChar( (char) 96));\r
- ret.addOr(new Range( (char) 123, (char) 65535));\r
- }\r
- else if (sp.escMatch('x') && next2Hex(sp))\r
- {\r
- sp.inc();\r
- int d = getHexDigit(sp);\r
- sp.inc();\r
- d = 16 * d + getHexDigit(sp);\r
- ret.addOr(new oneChar( (char) d));\r
- }\r
- else if (sp.escMatch('a'))\r
- {\r
- ret.addOr(new oneChar( (char) 7));\r
- }\r
- else if (sp.escMatch('f'))\r
- {\r
- ret.addOr(new oneChar( (char) 12));\r
- }\r
- else if (sp.escMatch('e'))\r
- {\r
- ret.addOr(new oneChar( (char) 27));\r
- }\r
- else if (sp.escMatch('n'))\r
- {\r
- ret.addOr(new oneChar('\n'));\r
- }\r
- else if (sp.escMatch('t'))\r
- {\r
- ret.addOr(new oneChar('\t'));\r
- }\r
- else if (sp.escMatch('r'))\r
- {\r
- ret.addOr(new oneChar('\r'));\r
- }\r
- else if (sp.escMatch('c'))\r
- {\r
- sp.inc();\r
- if (sp.c < Ctrl.cmap.length)\r
- {\r
- ret.addOr(new oneChar(Ctrl.cmap[sp.c]));\r
- }\r
- else\r
- {\r
- ret.addOr(new oneChar(sp.c));\r
- }\r
- }\r
- else if (isOctalString(sp))\r
- {\r
- int d = sp.c - '0';\r
- sp.inc();\r
- d = 8 * d + sp.c - '0';\r
- StrPos sp2 = new StrPos(sp);\r
- sp2.inc();\r
- if (isOctalDigit(sp2, false))\r
- {\r
- sp.inc();\r
- d = 8 * d + sp.c - '0';\r
- }\r
- ret.addOr(new oneChar( (char) d));\r
- }\r
- else\r
- {\r
- ret.addOr(new oneChar(sp.c));\r
- }\r
- sp.inc();\r
- }\r
- return ret;\r
- }\r
-\r
- /** Converts the stored Pattern to a String -- this is a\r
- decompile. Note that \t and \n will really print out here,\r
- Not just the two character representations.\r
- Also be prepared to see some strange output if your characters\r
- are not printable. */\r
- public String toString()\r
- {\r
- if (false && thePattern == null)\r
- {\r
- return "";\r
- }\r
- else\r
- {\r
- StringBuffer sb = new StringBuffer();\r
- if (esc != Pattern.ESC)\r
- {\r
- sb.append("(?e=");\r
- sb.append(esc);\r
- sb.append(")");\r
- }\r
- if (gFlag\r
- || mFlag\r
- || !dotDoesntMatchCR\r
- || sFlag\r
- || ignoreCase\r
- || dontMatchInQuotes\r
- || optimized())\r
- {\r
- sb.append("(?");\r
- if (ignoreCase)\r
- {\r
- sb.append("i");\r
- }\r
- if (mFlag)\r
- {\r
- sb.append("m");\r
- }\r
- if (sFlag || !dotDoesntMatchCR)\r
- {\r
- sb.append("s");\r
- }\r
- if (dontMatchInQuotes)\r
- {\r
- sb.append("Q");\r
- }\r
- if (optimized())\r
- {\r
- sb.append("o");\r
- }\r
- if (gFlag)\r
- {\r
- sb.append("g");\r
- }\r
- sb.append(")");\r
- }\r
- String patstr = thePattern.toString();\r
- if (esc != Pattern.ESC)\r
- {\r
- patstr = reEscape(patstr, Pattern.ESC, esc);\r
- }\r
- sb.append(patstr);\r
- return sb.toString();\r
- }\r
- }\r
-\r
- // Re-escape Pattern, allows us to use a different escape\r
- // character.\r
- static String reEscape(String s, char oldEsc, char newEsc)\r
- {\r
- if (oldEsc == newEsc)\r
- {\r
- return s;\r
- }\r
- int i;\r
- StringBuffer sb = new StringBuffer();\r
- for (i = 0; i < s.length(); i++)\r
- {\r
- if (s.charAt(i) == oldEsc && i + 1 < s.length())\r
- {\r
- if (s.charAt(i + 1) == oldEsc)\r
- {\r
- sb.append(oldEsc);\r
- }\r
- else\r
- {\r
- sb.append(newEsc);\r
- sb.append(s.charAt(i + 1));\r
- }\r
- i++;\r
- }\r
- else if (s.charAt(i) == newEsc)\r
- {\r
- sb.append(newEsc);\r
- sb.append(newEsc);\r
- }\r
- else\r
- {\r
- sb.append(s.charAt(i));\r
- }\r
- }\r
- return sb.toString();\r
- }\r
-\r
- /** This method implements FilenameFilter, allowing one\r
- to use a Regex to search through a directory using File.list.\r
- There is a FileRegex now that does this better.\r
- @see com.stevesoft.pat.FileRegex\r
- */\r
- public boolean accept(File dir, String s)\r
- {\r
- return search(s);\r
- }\r
-\r
- /** The version of this package */\r
- final static public String version()\r
- {\r
- return "lgpl release 1.5.3";\r
- }\r
-\r
- /** Once this method is called, the state of variables\r
- ignoreCase and dontMatchInQuotes should not be changed as the\r
- results will be unpredictable. However,\r
- search and matchAt will run more quickly. Note that you\r
- can check to see if the pattern has been optimized by calling\r
- the optimized() method.<p>This method will attempt to rewrite\r
- your pattern in a way that makes it faster (not all patterns\r
- execute at the same speed). In general, "(?: ... )" will be\r
- faster than "( ... )" so if you don't need the backreference,\r
- you should group using the former pattern.<p>It will also\r
- introduce new pattern elements that you can't get to otherwise,\r
- for example if you have a large table of strings, i.e. the\r
- months of the year "(January|February|...)" optimize() will make\r
- a Hashtable that takes it to the next appropriate pattern\r
- element -- eliminating the need for a linear search.\r
- @see com.stevesoft.pat.Regex#optimized\r
- @see com.stevesoft.pat.Regex#ignoreCase\r
- @see com.stevesoft.pat.Regex#dontMatchInQuotes\r
- @see com.stevesoft.pat.Regex#matchAt\r
- @see com.stevesoft.pat.Regex#search\r
- */\r
- public void optimize()\r
- {\r
- if (optimized() || thePattern == null)\r
- {\r
- return;\r
- }\r
- minMatch = new patInt(0); //thePattern.countMinChars();\r
- thePattern = RegOpt.opt(thePattern, ignoreCase,\r
- dontMatchInQuotes);\r
- skipper = Skip.findSkip(this);\r
- //RegOpt.setParents(this);\r
- return;\r
- }\r
-\r
- Skip skipper;\r
- /** This function returns true if the optimize method has\r
- been called. */\r
- public boolean optimized()\r
- {\r
- return minMatch != null;\r
- }\r
-\r
- /** A bit of syntactic surgar for those who want to make\r
- their code look more perl-like. To use this initialize\r
- your Regex object by saying:\r
- <pre>\r
- Regex r1 = Regex.perlCode("s/hello/goodbye/");\r
- Regex r2 = Regex.perlCode("s'fish'frog'i");\r
- Regex r3 = Regex.perlCode("m'hello');\r
- </pre>\r
- The i for ignoreCase is supported in\r
- this syntax, as well as m, s, and x. The g flat\r
- is a bit of a special case.<p>\r
- If you wish to replace all occurences of a pattern, you\r
- do not put a 'g' in the perlCode, but call Regex's\r
- replaceAll method.<p>\r
- If you wish to simply\r
- and only do a search for r2's pattern, you can do this\r
- by calling the searchFrom method method repeatedly, or\r
- by calling search repeatedly if the g flag is set.\r
- <p>\r
- Note: Currently perlCode does <em>not</em>\r
- support the (?e=#) syntax for\r
- changing the escape character.\r
- */\r
-\r
- public static Regex perlCode(String s)\r
- {\r
- // this file is big enough, see parsePerl.java\r
- // for this function.\r
- return parsePerl.parse(s);\r
- }\r
-\r
- static final char back_slash = '\\';\r
-\r
- /** Checks to see if there are only literal and no special\r
- pattern elements in this Regex. */\r
- public boolean isLiteral()\r
- {\r
- Pattern x = thePattern;\r
- while (x != null)\r
- {\r
- if (x instanceof oneChar)\r
- {\r
- ;\r
- }\r
- else if (x instanceof Skipped)\r
- {\r
- ;\r
- }\r
- else\r
- {\r
- return false;\r
- }\r
- x = x.next;\r
- }\r
- return true;\r
- }\r
-\r
- /** You only need to know about this if you are inventing\r
- your own pattern elements. */\r
- public patInt countMinChars()\r
- {\r
- return thePattern.countMinChars();\r
- }\r
-\r
- /** You only need to know about this if you are inventing\r
- your own pattern elements. */\r
- public patInt countMaxChars()\r
- {\r
- return thePattern.countMaxChars();\r
- }\r
-\r
- boolean isHexDigit(StrPos sp)\r
- {\r
- boolean r =\r
- !sp.eos && !sp.dontMatch\r
- && ( (sp.c >= '0' && sp.c <= '9')\r
- || (sp.c >= 'a' && sp.c <= 'f')\r
- || (sp.c >= 'A' && sp.c <= 'F'));\r
- return r;\r
- }\r
-\r
- boolean isOctalDigit(StrPos sp, boolean first)\r
- {\r
- boolean r =\r
- !sp.eos && ! (first ^ sp.dontMatch)\r
- && sp.c >= '0' && sp.c <= '7';\r
- return r;\r
- }\r
-\r
- int getHexDigit(StrPos sp)\r
- {\r
- if (sp.c >= '0' && sp.c <= '9')\r
- {\r
- return sp.c - '0';\r
- }\r
- if (sp.c >= 'a' && sp.c <= 'f')\r
- {\r
- return sp.c - 'a' + 10;\r
- }\r
- return sp.c - 'A' + 10;\r
- }\r
-\r
- boolean next2Hex(StrPos sp)\r
- {\r
- StrPos sp2 = new StrPos(sp);\r
- sp2.inc();\r
- if (!isHexDigit(sp2))\r
- {\r
- return false;\r
- }\r
- sp2.inc();\r
- if (!isHexDigit(sp2))\r
- {\r
- return false;\r
- }\r
- return true;\r
- }\r
-\r
- boolean isOctalString(StrPos sp)\r
- {\r
- if (!isOctalDigit(sp, true))\r
- {\r
- return false;\r
- }\r
- StrPos sp2 = new StrPos(sp);\r
- sp2.inc();\r
- if (!isOctalDigit(sp2, false))\r
- {\r
- return false;\r
- }\r
- return true;\r
- }\r
-}\r
+//
+// This software is now distributed according to
+// the Lesser Gnu Public License. Please see
+// http://www.gnu.org/copyleft/lesser.txt for
+// the details.
+// -- Happy Computing!
+//
+package com.stevesoft.pat;
+
+import java.io.*;
+import java.util.*;
+
+import com.stevesoft.pat.wrap.*;
+
+/** Matches a Unicode punctuation character. */
+class UnicodePunct extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && Prop.isPunct(s.charAt(from)) ? to : -1;
+ }
+}
+
+/** Matches a Unicode white space character. */
+class UnicodeWhite extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && Prop.isWhite(s.charAt(from)) ? to : -1;
+ }
+}
+
+/**
+ * Matches a character that is not a Unicode punctuation character.
+ */
+class NUnicodePunct extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && !Prop.isPunct(s.charAt(from)) ? to : -1;
+ }
+}
+
+/**
+ * Matches a character that is not a Unicode white space character.
+ */
+class NUnicodeWhite extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && !Prop.isWhite(s.charAt(from)) ? to : -1;
+ }
+}
+
+/** Matches a Unicode word character: an alphanumeric or underscore. */
+class UnicodeW extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ if (from >= s.length())
+ {
+ return -1;
+ }
+ char c = s.charAt(from);
+ return (Prop.isAlphabetic(c) || Prop.isDecimalDigit(c) || c == '_') ? to
+ : -1;
+ }
+}
+
+/** Matches a character that is not a Unicode alphanumeric or underscore. */
+class NUnicodeW extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ if (from >= s.length())
+ {
+ return -1;
+ }
+ char c = s.charAt(from);
+ return !(Prop.isAlphabetic(c) || Prop.isDecimalDigit(c) || c == '_') ? to
+ : -1;
+ }
+}
+
+/** Matches a Unicode decimal digit. */
+class UnicodeDigit extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && Prop.isDecimalDigit(s.charAt(from)) ? to
+ : -1;
+ }
+}
+
+/** Matches a character that is not a Unicode digit. */
+class NUnicodeDigit extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && !Prop.isDecimalDigit(s.charAt(from)) ? to
+ : -1;
+ }
+}
+
+/** Matches a Unicode math character. */
+class UnicodeMath extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && Prop.isMath(s.charAt(from)) ? to : -1;
+ }
+}
+
+/** Matches a non-math Unicode character. */
+class NUnicodeMath extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && !Prop.isMath(s.charAt(from)) ? to : -1;
+ }
+}
+
+/** Matches a Unicode currency symbol. */
+class UnicodeCurrency extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && Prop.isCurrency(s.charAt(from)) ? to : -1;
+ }
+}
+
+/** Matches a non-currency symbol Unicode character. */
+class NUnicodeCurrency extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && !Prop.isCurrency(s.charAt(from)) ? to : -1;
+ }
+}
+
+/** Matches a Unicode alphabetic character. */
+class UnicodeAlpha extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && Prop.isAlphabetic(s.charAt(from)) ? to : -1;
+ }
+}
+
+/** Matches a non-alphabetic Unicode character. */
+class NUnicodeAlpha extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && !Prop.isAlphabetic(s.charAt(from)) ? to
+ : -1;
+ }
+}
+
+/** Matches an upper case Unicode character. */
+class UnicodeUpper extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && isUpper(s.charAt(from)) ? to : -1;
+ }
+
+ final boolean isUpper(char c)
+ {
+ return c == CaseMgr.toUpperCase(c) && c != CaseMgr.toLowerCase(c);
+ }
+}
+
+/** Matches an upper case Unicode character. */
+class UnicodeLower extends UniValidator
+{
+ public int validate(StringLike s, int from, int to)
+ {
+ return from < s.length() && isLower(s.charAt(from)) ? to : -1;
+ }
+
+ final boolean isLower(char c)
+ {
+ return c != CaseMgr.toUpperCase(c) && c == CaseMgr.toLowerCase(c);
+ }
+}
+
+/**
+ * Regex provides the parser which constructs the linked list of Pattern classes
+ * from a String.
+ * <p>
+ * For the purpose of this documentation, the fact that java interprets the
+ * backslash will be ignored. In practice, however, you will need a double
+ * backslash to obtain a string that contains a single backslash character.
+ * Thus, the example pattern "\b" should really be typed as "\\b" inside java
+ * code.
+ * <p>
+ * Note that Regex is part of package "com.stevesoft.pat". To use it, simply
+ * import com.stevesoft.pat.Regex at the top of your file.
+ * <p>
+ * Regex is made with a constructor that takes a String that defines the regular
+ * expression. Thus, for example
+ *
+ * <pre>
+ * Regex r = new Regex("[a-c]*");
+ * </pre>
+ *
+ * matches any number of characters so long as the are 'a', 'b', or 'c').
+ * <p>
+ * To attempt to match the Pattern to a given string, you can use either the
+ * search(String) member function, or the matchAt(String,int position) member
+ * function. These functions return a boolean which tells you whether or not the
+ * thing worked, and sets the methods "charsMatched()" and "matchedFrom()" in
+ * the Regex object appropriately.
+ * <p>
+ * The portion of the string before the match can be obtained by the left()
+ * member, and the portion after the match can be obtained by the right()
+ * member.
+ * <p>
+ * Essentially, this package implements a syntax that is very much like the perl
+ * 5 regular expression syntax.
+ *
+ * Longer example:
+ *
+ * <pre>
+ * Regex r = new Regex("x(a|b)y");
+ * r.matchAt("xay", 0);
+ * System.out.println("sub = " + r.stringMatched(1));
+ * </pre>
+ *
+ * The above would print "sub = a".
+ *
+ * <pre>
+ * r.left() // would return "x"
+ * r.right() // would return "y"
+ * </pre>
+ *
+ * <p>
+ * Differences between this package and perl5:<br>
+ * The extended Pattern for setting flags, is now supported, but the flags are
+ * different. "(?i)" tells the pattern to ignore case, "(?Q)" sets the
+ * "dontMatchInQuotes" flag, and "(?iQ)" sets them both. You can change the
+ * escape character. The pattern
+ *
+ * <pre>
+ * (?e=#)#d+
+ * </pre>
+ *
+ * is the same as
+ *
+ * <pre>
+ * \d+
+ * </pre>, but note that the sequence
+ *
+ * <pre>
+ * (?e=#)
+ * </pre>
+ *
+ * <b>must</b> occur at the very beginning of the pattern. There may be other
+ * small differences as well. I will either make my package conform or note them
+ * as I become aware of them.
+ * <p>
+ * This package supports additional patterns not in perl5: <center> <table
+ * border=1>
+ * <tr>
+ * <td>(?@())</td>
+ * <td>Group</td>
+ * <td>This matches all characters between the '(' character and the balancing
+ * ')' character. Thus, it will match "()" as well as "(())". The balancing
+ * characters are arbitrary, thus (?@{}) matches on "{}" and "{{}}".</td>
+ * <tr>
+ * <td>(?<1)</td>
+ * <td>Backup</td>
+ * <td>Moves the pointer backwards within the text. This allows you to make a
+ * "look behind." It fails if it attempts to move to a position before the
+ * beginning of the string. "x(?<1)" is equivalent to "(?=x)". The number, 1
+ * in this example, is the number of characters to move backwards.</td>
+ * </table> </center>
+ * </dl>
+ *
+ * @author Steven R. Brandt
+ * @version package com.stevesoft.pat, release 1.5.3
+ * @see Pattern
+ */
+public class Regex extends RegRes implements FilenameFilter
+{
+ /**
+ * BackRefOffset gives the identity number of the first pattern. Version 1.0
+ * used zero, version 1.1 uses 1 to be more compatible with perl.
+ */
+ static int BackRefOffset = 1;
+
+ private static Pattern none = new NoPattern();
+
+ Pattern thePattern = none;
+
+ patInt minMatch = new patInt(0);
+
+ static Hashtable validators = new Hashtable();
+ static
+ {
+ define("p", "(?>1)", new UnicodePunct());
+ define("P", "(?>1)", new NUnicodePunct());
+ define("s", "(?>1)", new UnicodeWhite());
+ define("S", "(?>1)", new NUnicodeWhite());
+ define("w", "(?>1)", new UnicodeW());
+ define("W", "(?>1)", new NUnicodeW());
+ define("d", "(?>1)", new UnicodeDigit());
+ define("D", "(?>1)", new NUnicodeDigit());
+ define("m", "(?>1)", new UnicodeMath());
+ define("M", "(?>1)", new NUnicodeMath());
+ define("c", "(?>1)", new UnicodeCurrency());
+ define("C", "(?>1)", new NUnicodeCurrency());
+ define("a", "(?>1)", new UnicodeAlpha());
+ define("A", "(?>1)", new NUnicodeAlpha());
+ define("uc", "(?>1)", new UnicodeUpper());
+ define("lc", "(?>1)", new UnicodeLower());
+ }
+
+ /** Set the dontMatch in quotes flag. */
+ public void setDontMatchInQuotes(boolean b)
+ {
+ dontMatchInQuotes = b;
+ }
+
+ /** Find out if the dontMatchInQuotes flag is enabled. */
+ public boolean getDontMatchInQuotes()
+ {
+ return dontMatchInQuotes;
+ }
+
+ boolean dontMatchInQuotes = false;
+
+ /**
+ * Set the state of the ignoreCase flag. If set to true, then the pattern
+ * matcher will ignore case when searching for a match.
+ */
+ public void setIgnoreCase(boolean b)
+ {
+ ignoreCase = b;
+ }
+
+ /**
+ * Get the state of the ignoreCase flag. Returns true if we are ignoring the
+ * case of the pattern, false otherwise.
+ */
+ public boolean getIgnoreCase()
+ {
+ return ignoreCase;
+ }
+
+ boolean ignoreCase = false;
+
+ static boolean defaultMFlag = false;
+
+ /**
+ * Set the default value of the m flag. If it is set to true, then the MFlag
+ * will be on for any regex search executed.
+ */
+ public static void setDefaultMFlag(boolean mFlag)
+ {
+ defaultMFlag = mFlag;
+ }
+
+ /**
+ * Get the default value of the m flag. If it is set to true, then the MFlag
+ * will be on for any regex search executed.
+ */
+ public static boolean getDefaultMFlag()
+ {
+ return defaultMFlag;
+ }
+
+ /**
+ * Initializes the object without a Pattern. To supply a Pattern use
+ * compile(String s).
+ *
+ * @see com.stevesoft.pat.Regex#compile(java.lang.String)
+ */
+ public Regex()
+ {
+ }
+
+ /**
+ * Create and compile a Regex, but do not throw any exceptions. If you wish to
+ * have exceptions thrown for syntax errors, you must use the Regex(void)
+ * constructor to create the Regex object, and then call the compile method.
+ * Therefore, you should only call this method when you know your pattern is
+ * right. I will probably become more like
+ *
+ * @see com.stevesoft.pat.Regex#search(java.lang.String)
+ * @see com.stevesoft.pat.Regex#compile(java.lang.String)
+ */
+ public Regex(String s)
+ {
+ try
+ {
+ compile(s);
+ } catch (RegSyntax rs)
+ {
+ }
+ }
+
+ ReplaceRule rep = null;
+
+ /**
+ * Create and compile both a Regex and a ReplaceRule.
+ *
+ * @see com.stevesoft.pat.ReplaceRule
+ * @see com.stevesoft.pat.Regex#compile(java.lang.String)
+ */
+ public Regex(String s, String rp)
+ {
+ this(s);
+ rep = ReplaceRule.perlCode(rp);
+ }
+
+ /**
+ * Create and compile a Regex, but give it the ReplaceRule specified. This
+ * allows the user finer control of the Replacement process, if that is
+ * desired.
+ *
+ * @see com.stevesoft.pat.ReplaceRule
+ * @see com.stevesoft.pat.Regex#compile(java.lang.String)
+ */
+ public Regex(String s, ReplaceRule rp)
+ {
+ this(s);
+ rep = rp;
+ }
+
+ /**
+ * Change the ReplaceRule of this Regex by compiling a new one using String
+ * rp.
+ */
+ public void setReplaceRule(String rp)
+ {
+ rep = ReplaceRule.perlCode(rp);
+ repr = null; // Clear Replacer history
+ }
+
+ /** Change the ReplaceRule of this Regex to rp. */
+ public void setReplaceRule(ReplaceRule rp)
+ {
+ rep = rp;
+ }
+
+ /**
+ * Test to see if a custom defined rule exists.
+ *
+ * @see com.stevesoft.pat#define(java.lang.String,java.lang.String,Validator)
+ */
+ public static boolean isDefined(String nm)
+ {
+ return validators.get(nm) != null;
+ }
+
+ /**
+ * Removes a custom defined rule.
+ *
+ * @see com.stevesoft.pat#define(java.lang.String,java.lang.String,Validator)
+ */
+ public static void undefine(String nm)
+ {
+ validators.remove(nm);
+ }
+
+ /**
+ * Defines a method to create a new rule. See test/deriv2.java and
+ * test/deriv3.java for examples of how to use it.
+ */
+ public static void define(String nm, String pat, Validator v)
+ {
+ v.pattern = pat;
+ validators.put(nm, v);
+ }
+
+ /**
+ * Defines a shorthand for a pattern. The pattern will be invoked by a string
+ * that has the form "(??"+nm+")".
+ */
+ public static void define(String nm, String pat)
+ {
+ validators.put(nm, pat);
+ }
+
+ /** Get the current ReplaceRule. */
+ public ReplaceRule getReplaceRule()
+ {
+ return rep;
+ }
+
+ Replacer repr = null;
+
+ final Replacer _getReplacer()
+ {
+ return repr == null ? repr = new Replacer() : repr;
+ }
+
+ public Replacer getReplacer()
+ {
+ if (repr == null)
+ {
+ repr = new Replacer();
+ }
+ repr.rh.me = this;
+ repr.rh.prev = null;
+ return repr;
+ }
+
+ /**
+ * Replace the first occurence of this pattern in String s according to the
+ * ReplaceRule.
+ *
+ * @see com.stevesoft.pat.ReplaceRule
+ * @see com.stevesoft.pat.Regex#getReplaceRule()
+ */
+ public String replaceFirst(String s)
+ {
+ return _getReplacer().replaceFirstRegion(s, this, 0, s.length())
+ .toString();
+ }
+
+ /**
+ * Replace the first occurence of this pattern in String s beginning with
+ * position pos according to the ReplaceRule.
+ *
+ * @see com.stevesoft.pat.ReplaceRule
+ * @see com.stevesoft.pat.Regex#getReplaceRule()
+ */
+ public String replaceFirstFrom(String s, int pos)
+ {
+ return _getReplacer().replaceFirstRegion(s, this, pos, s.length())
+ .toString();
+ }
+
+ /**
+ * Replace the first occurence of this pattern in String s beginning with
+ * position start and ending with end according to the ReplaceRule.
+ *
+ * @see com.stevesoft.pat.ReplaceRule
+ * @see com.stevesoft.pat.Regex#getReplaceRule()
+ */
+ public String replaceFirstRegion(String s, int start, int end)
+ {
+ return _getReplacer().replaceFirstRegion(s, this, start, end)
+ .toString();
+ }
+
+ /**
+ * Replace all occurences of this pattern in String s according to the
+ * ReplaceRule.
+ *
+ * @see com.stevesoft.pat.ReplaceRule
+ * @see com.stevesoft.pat.Regex#getReplaceRule()
+ */
+ public String replaceAll(String s)
+ {
+ return _getReplacer().replaceAllRegion(s, this, 0, s.length())
+ .toString();
+ }
+
+ public StringLike replaceAll(StringLike s)
+ {
+ return _getReplacer().replaceAllRegion(s, this, 0, s.length());
+ }
+
+ /**
+ * Replace all occurences of this pattern in String s beginning with position
+ * pos according to the ReplaceRule.
+ *
+ * @see com.stevesoft.pat.ReplaceRule
+ * @see com.stevesoft.pat.Regex#getReplaceRule()
+ */
+ public String replaceAllFrom(String s, int pos)
+ {
+ return _getReplacer().replaceAllRegion(s, this, pos, s.length())
+ .toString();
+ }
+
+ /**
+ * Replace all occurences of this pattern in String s beginning with position
+ * start and ending with end according to the ReplaceRule.
+ *
+ * @see com.stevesoft.pat.ReplaceRule
+ * @see com.stevesoft.pat.Regex#getReplaceRule()
+ */
+ public String replaceAllRegion(String s, int start, int end)
+ {
+ return _getReplacer().replaceAllRegion(s, this, start, end).toString();
+ }
+
+ /** Essentially clones the Regex object */
+ public Regex(Regex r)
+ {
+ super((RegRes) r);
+ dontMatchInQuotes = r.dontMatchInQuotes;
+ esc = r.esc;
+ ignoreCase = r.ignoreCase;
+ gFlag = r.gFlag;
+ if (r.rep == null)
+ {
+ rep = null;
+ }
+ else
+ {
+ rep = (ReplaceRule) r.rep.clone();
+ }
+ /*
+ * try { compile(r.toString()); } catch(RegSyntax r_) {}
+ */
+ thePattern = r.thePattern.clone(new Hashtable());
+ minMatch = r.minMatch;
+ skipper = r.skipper;
+ }
+
+ /**
+ * By default, the escape character is the backslash, but you can make it
+ * anything you want by setting this variable.
+ */
+ public char esc = Pattern.ESC;
+
+ /**
+ * This method compiles a regular expression, making it possible to call the
+ * search or matchAt methods.
+ *
+ * @exception com.stevesoft.pat.RegSyntax
+ * is thrown if a syntax error is encountered in the
+ * pattern. For example, "x{3,1}" or "*a" are not valid
+ * patterns.
+ * @see com.stevesoft.pat.Regex#search
+ * @see com.stevesoft.pat.Regex#matchAt
+ */
+ public void compile(String prepat) throws RegSyntax
+ {
+ String postpat = parsePerl.codify(prepat, true);
+ String pat = postpat == null ? prepat : postpat;
+ minMatch = null;
+ ignoreCase = false;
+ dontMatchInQuotes = false;
+ Rthings mk = new Rthings(this);
+ int offset = mk.val;
+ String newpat = pat;
+ thePattern = none;
+ p = null;
+ or = null;
+ minMatch = new patInt(0);
+ StrPos sp = new StrPos(pat, 0);
+ if (sp.incMatch("(?e="))
+ {
+ char newEsc = sp.c;
+ sp.inc();
+ if (sp.match(')'))
+ {
+ newpat = reEscape(pat.substring(6), newEsc, Pattern.ESC);
+ }
+ }
+ else if (esc != Pattern.ESC)
+ {
+ newpat = reEscape(pat, esc, Pattern.ESC);
+ }
+ thePattern = _compile(newpat, mk);
+ numSubs_ = mk.val - offset;
+ mk.set(this);
+ }
+
+ /*
+ * If a Regex is compared against a Regex, a check is done to see that the
+ * patterns are equal as well as the most recent match. If a Regex is compare
+ * with a RegRes, only the result of the most recent match is compared.
+ */
+ public boolean equals(Object o)
+ {
+ if (o instanceof Regex)
+ {
+ if (toString().equals(o.toString()))
+ {
+ return super.equals(o);
+ }
+ else
+ {
+ return false;
+ }
+ }
+ else
+ {
+ return super.equals(o);
+ }
+ }
+
+ /** A clone by any other name would smell as sweet. */
+ public Object clone()
+ {
+ return new Regex(this);
+ }
+
+ /** Return a clone of the underlying RegRes object. */
+ public RegRes result()
+ {
+ return (RegRes) super.clone();
+ }
+
+ // prep sets global variables of class
+ // Pattern so that it can access them
+ // during an attempt at a match
+ Pthings pt = new Pthings();
+
+ final Pthings prep(StringLike s)
+ {
+ // if(gFlag)
+ pt.lastPos = matchedTo();
+ if (pt.lastPos < 0)
+ {
+ pt.lastPos = 0;
+ }
+ if ((s == null ? null : s.unwrap()) != (src == null ? null : s.unwrap()))
+ {
+ pt.lastPos = 0;
+ }
+ src = s;
+ pt.dotDoesntMatchCR = dotDoesntMatchCR && (!sFlag);
+ pt.mFlag = (mFlag | defaultMFlag);
+ pt.ignoreCase = ignoreCase;
+ pt.no_check = false;
+ if (pt.marks != null)
+ {
+ for (int i = 0; i < pt.marks.length; i++)
+ {
+ pt.marks[i] = -1;
+ }
+ }
+ pt.marks = null;
+ pt.nMarks = numSubs_;
+ pt.src = s;
+ if (dontMatchInQuotes)
+ {
+ setCbits(s, pt);
+ }
+ else
+ {
+ pt.cbits = null;
+ }
+ return pt;
+ }
+
+ /**
+ * Attempt to match a Pattern beginning at a specified location within the
+ * string.
+ *
+ * @see com.stevesoft.pat.Regex#search
+ */
+ public boolean matchAt(String s, int start_pos)
+ {
+ return _search(s, start_pos, start_pos);
+ }
+
+ /**
+ * Attempt to match a Pattern beginning at a specified location within the
+ * StringLike.
+ *
+ * @see com.stevesoft.pat.Regex#search
+ */
+ public boolean matchAt(StringLike s, int start_pos)
+ {
+ return _search(s, start_pos, start_pos);
+ }
+
+ /**
+ * Search through a String for the first occurrence of a match.
+ *
+ * @see com.stevesoft.pat.Regex#searchFrom
+ * @see com.stevesoft.pat.Regex#matchAt
+ */
+ public boolean search(String s)
+ {
+ if (s == null)
+ {
+ throw new NullPointerException("Null String Given to Regex.search");
+ }
+ return _search(s, 0, s.length());
+ }
+
+ public boolean search(StringLike sl)
+ {
+ if (sl == null)
+ {
+ throw new NullPointerException(
+ "Null StringLike Given to Regex.search");
+ }
+ return _search(sl, 0, sl.length());
+ }
+
+ public boolean reverseSearch(String s)
+ {
+ if (s == null)
+ {
+ throw new NullPointerException(
+ "Null String Given to Regex.reverseSearch");
+ }
+ return _reverseSearch(s, 0, s.length());
+ }
+
+ public boolean reverseSearch(StringLike sl)
+ {
+ if (sl == null)
+ {
+ throw new NullPointerException(
+ "Null StringLike Given to Regex.reverseSearch");
+ }
+ return _reverseSearch(sl, 0, sl.length());
+ }
+
+ /**
+ * Search through a String for the first occurence of a match, but start at
+ * position
+ *
+ * <pre>
+ * start
+ * </pre>
+ */
+ public boolean searchFrom(String s, int start)
+ {
+ if (s == null)
+ {
+ throw new NullPointerException(
+ "Null String Given to Regex.searchFrom");
+ }
+ return _search(s, start, s.length());
+ }
+
+ public boolean searchFrom(StringLike s, int start)
+ {
+ if (s == null)
+ {
+ throw new NullPointerException(
+ "Null String Given to Regex.searchFrom");
+ }
+ return _search(s, start, s.length());
+ }
+
+ /**
+ * Search through a region of a String for the first occurence of a match.
+ */
+ public boolean searchRegion(String s, int start, int end)
+ {
+ if (s == null)
+ {
+ throw new NullPointerException(
+ "Null String Given to Regex.searchRegion");
+ }
+ return _search(s, start, end);
+ }
+
+ /**
+ * Set this to change the default behavior of the "." pattern. By default it
+ * now matches perl's behavior and fails to match the '\n' character.
+ */
+ public static boolean dotDoesntMatchCR = true;
+
+ StringLike gFlags;
+
+ int gFlagto = 0;
+
+ boolean gFlag = false;
+
+ /** Set the 'g' flag */
+ public void setGFlag(boolean b)
+ {
+ gFlag = b;
+ }
+
+ /** Get the state of the 'g' flag. */
+ public boolean getGFlag()
+ {
+ return gFlag;
+ }
+
+ boolean sFlag = false;
+
+ /** Get the state of the sFlag */
+ public boolean getSFlag()
+ {
+ return sFlag;
+ }
+
+ boolean mFlag = false;
+
+ /** Get the state of the sFlag */
+ public boolean getMFlag()
+ {
+ return mFlag;
+ }
+
+ final boolean _search(String s, int start, int end)
+ {
+ return _search(new StringWrap(s), start, end);
+ }
+
+ final boolean _search(StringLike s, int start, int end)
+ {
+ if (gFlag && gFlagto > 0 && gFlags != null
+ && s.unwrap() == gFlags.unwrap())
+ {
+ start = gFlagto;
+ }
+ gFlags = null;
+
+ Pthings pt = prep(s);
+
+ int up = (minMatch == null ? end : end - minMatch.i);
+
+ if (up < start && end >= start)
+ {
+ up = start;
+ }
+
+ if (skipper == null)
+ {
+ for (int i = start; i <= up; i++)
+ {
+ charsMatched_ = thePattern.matchAt(s, i, pt);
+ if (charsMatched_ >= 0)
+ {
+ matchFrom_ = thePattern.mfrom;
+ marks = pt.marks;
+ gFlagto = matchFrom_ + charsMatched_;
+ gFlags = s;
+ return didMatch_ = true;
+ }
+ }
+ }
+ else
+ {
+ pt.no_check = true;
+ for (int i = start; i <= up; i++)
+ {
+ i = skipper.find(src, i, up);
+ if (i < 0)
+ {
+ charsMatched_ = matchFrom_ = -1;
+ return didMatch_ = false;
+ }
+ charsMatched_ = thePattern.matchAt(s, i, pt);
+ if (charsMatched_ >= 0)
+ {
+ matchFrom_ = thePattern.mfrom;
+ marks = pt.marks;
+ gFlagto = matchFrom_ + charsMatched_;
+ gFlags = s;
+ return didMatch_ = true;
+ }
+ }
+ }
+ return didMatch_ = false;
+ }
+
+ /*
+ * final boolean _search(LongStringLike s,long start,long end) { if(gFlag &&
+ * gFlagto > 0 && s==gFlags) start = gFlagto; gFlags = null;
+ *
+ * Pthings pt=prep(s);
+ *
+ * int up = end;//(minMatch == null ? end : end-minMatch.i);
+ *
+ * if(up < start && end >= start) up = start;
+ *
+ * if(skipper == null) { for(long i=start;i<=up;i++) { charsMatched_ =
+ * thePattern.matchAt(s,i,pt); if(charsMatched_ >= 0) { matchFrom_ =
+ * thePattern.mfrom; marks = pt.marks; gFlagto = matchFrom_+charsMatched_;
+ * return didMatch_=true; } } } else { pt.no_check = true; for(long i=start;i<=up;i++) {
+ * i = skipper.find(src,i,up); if(i<0) { charsMatched_ = matchFrom_ = -1;
+ * return didMatch_ = false; } charsMatched_ = thePattern.matchAt(s,i,pt);
+ * if(charsMatched_ >= 0) { matchFrom_ = thePattern.mfrom; marks = pt.marks;
+ * gFlagto = matchFrom_+charsMatched_; gFlags = s; return didMatch_=true; }
+ * else { i = s.adjustIndex(i); up = s.adjustEnd(i); } } } return
+ * didMatch_=false; }
+ */
+
+ boolean _reverseSearch(String s, int start, int end)
+ {
+ return _reverseSearch(new StringWrap(s), start, end);
+ }
+
+ boolean _reverseSearch(StringLike s, int start, int end)
+ {
+ if (gFlag && gFlagto > 0 && s.unwrap() == gFlags.unwrap())
+ {
+ end = gFlagto;
+ }
+ gFlags = null;
+ Pthings pt = prep(s);
+ for (int i = end; i >= start; i--)
+ {
+ charsMatched_ = thePattern.matchAt(s, i, pt);
+ if (charsMatched_ >= 0)
+ {
+ matchFrom_ = thePattern.mfrom;
+ marks = pt.marks;
+ gFlagto = matchFrom_ - 1;
+ gFlags = s;
+ return didMatch_ = true;
+ }
+ }
+ return didMatch_ = false;
+ }
+
+ // This routine sets the cbits variable
+ // of class Pattern. Cbits is true for
+ // the bit corresponding to a character inside
+ // a set of quotes.
+ static StringLike lasts = null;
+
+ static BitSet lastbs = null;
+
+ static void setCbits(StringLike s, Pthings pt)
+ {
+ if (s == lasts)
+ {
+ pt.cbits = lastbs;
+ return;
+ }
+ BitSet bs = new BitSet(s.length());
+ char qc = ' ';
+ boolean setBit = false;
+ for (int i = 0; i < s.length(); i++)
+ {
+ if (setBit)
+ {
+ bs.set(i);
+ }
+ char c = s.charAt(i);
+ if (!setBit && c == '"')
+ {
+ qc = c;
+ setBit = true;
+ bs.set(i);
+ }
+ else if (!setBit && c == '\'')
+ {
+ qc = c;
+ setBit = true;
+ bs.set(i);
+ }
+ else if (setBit && c == qc)
+ {
+ setBit = false;
+ }
+ else if (setBit && c == '\\' && i + 1 < s.length())
+ {
+ i++;
+ if (setBit)
+ {
+ bs.set(i);
+ }
+ }
+ }
+ pt.cbits = lastbs = bs;
+ lasts = s;
+ }
+
+ // Wanted user to over-ride this in alpha version,
+ // but it wasn't really necessary because of this trick:
+ Regex newRegex()
+ {
+ try
+ {
+ return (Regex) getClass().newInstance();
+ } catch (InstantiationException ie)
+ {
+ return null;
+ } catch (IllegalAccessException iae)
+ {
+ return null;
+ }
+ }
+
+ /**
+ * Only needed for creating your own extensions of Regex. This method adds the
+ * next Pattern in the chain of patterns or sets the Pattern if it is the
+ * first call.
+ */
+ protected void add(Pattern p2)
+ {
+ if (p == null)
+ {
+ p = p2;
+ }
+ else
+ {
+ p.add(p2);
+ p2 = p;
+ }
+ }
+
+ /**
+ * You only need to use this method if you are creating your own extentions to
+ * Regex. compile1 compiles one Pattern element, it can be over-ridden to
+ * allow the Regex compiler to understand new syntax. See deriv.java for an
+ * example. This routine is the heart of class Regex. Rthings has one integer
+ * member called intValue, it is used to keep track of the number of ()'s in
+ * the Pattern.
+ *
+ * @exception com.stevesoft.pat.RegSyntax
+ * is thrown when a nonsensensical pattern is supplied. For
+ * example, a pattern beginning with *.
+ */
+ protected void compile1(StrPos sp, Rthings mk) throws RegSyntax
+ {
+ if (sp.match('['))
+ {
+ sp.inc();
+ add(matchBracket(sp));
+ }
+ else if (sp.match('|'))
+ {
+ if (or == null)
+ {
+ or = new Or();
+ }
+ if (p == null)
+ {
+ p = new NullPattern();
+ }
+ or.addOr(p);
+ p = null;
+ }
+ else if (sp.incMatch("(?<"))
+ {
+ patInt i = sp.getPatInt();
+ if (i == null)
+ {
+ RegSyntaxError.endItAll("No int after (?<");
+ }
+ add(new Backup(i.intValue()));
+ if (!sp.match(')'))
+ {
+ RegSyntaxError.endItAll("No ) after (?<");
+ }
+ }
+ else if (sp.incMatch("(?>"))
+ {
+ patInt i = sp.getPatInt();
+ if (i == null)
+ {
+ RegSyntaxError.endItAll("No int after (?>");
+ }
+ add(new Backup(-i.intValue()));
+ if (!sp.match(')'))
+ {
+ RegSyntaxError.endItAll("No ) after (?<");
+ }
+ }
+ else if (sp.incMatch("(?@"))
+ {
+ char op = sp.c;
+ sp.inc();
+ char cl = sp.c;
+ sp.inc();
+ if (!sp.match(')'))
+ {
+ RegSyntaxError.endItAll("(?@ does not have closing paren");
+ }
+ add(new Group(op, cl));
+ }
+ else if (sp.incMatch("(?#"))
+ {
+ while (!sp.match(')'))
+ {
+ sp.inc();
+ }
+ }
+ else if (sp.dontMatch && sp.c == 'w')
+ {
+ // Regex r = new Regex();
+ // r._compile("[a-zA-Z0-9_]",mk);
+ // add(new Goop("\\w",r.thePattern));
+ Bracket b = new Bracket(false);
+ b.addOr(new Range('a', 'z'));
+ b.addOr(new Range('A', 'Z'));
+ b.addOr(new Range('0', '9'));
+ b.addOr(new oneChar('_'));
+ add(b);
+ }
+ else if (sp.dontMatch && sp.c == 'G')
+ {
+ add(new BackG());
+ }
+ else if (sp.dontMatch && sp.c == 's')
+ {
+ // Regex r = new Regex();
+ // r._compile("[ \t\n\r\b]",mk);
+ // add(new Goop("\\s",r.thePattern));
+ Bracket b = new Bracket(false);
+ b.addOr(new oneChar((char) 32));
+ b.addOr(new Range((char) 8, (char) 10));
+ b.addOr(new oneChar((char) 13));
+ add(b);
+ }
+ else if (sp.dontMatch && sp.c == 'd')
+ {
+ // Regex r = new Regex();
+ // r._compile("[0-9]",mk);
+ // add(new Goop("\\d",r.thePattern));
+ Range digit = new Range('0', '9');
+ digit.printBrackets = true;
+ add(digit);
+ }
+ else if (sp.dontMatch && sp.c == 'W')
+ {
+ // Regex r = new Regex();
+ // r._compile("[^a-zA-Z0-9_]",mk);
+ // add(new Goop("\\W",r.thePattern));
+ Bracket b = new Bracket(true);
+ b.addOr(new Range('a', 'z'));
+ b.addOr(new Range('A', 'Z'));
+ b.addOr(new Range('0', '9'));
+ b.addOr(new oneChar('_'));
+ add(b);
+ }
+ else if (sp.dontMatch && sp.c == 'S')
+ {
+ // Regex r = new Regex();
+ // r._compile("[^ \t\n\r\b]",mk);
+ // add(new Goop("\\S",r.thePattern));
+ Bracket b = new Bracket(true);
+ b.addOr(new oneChar((char) 32));
+ b.addOr(new Range((char) 8, (char) 10));
+ b.addOr(new oneChar((char) 13));
+ add(b);
+ }
+ else if (sp.dontMatch && sp.c == 'D')
+ {
+ // Regex r = new Regex();
+ // r._compile("[^0-9]",mk);
+ // add(new Goop("\\D",r.thePattern));
+ Bracket b = new Bracket(true);
+ b.addOr(new Range('0', '9'));
+ add(b);
+ }
+ else if (sp.dontMatch && sp.c == 'B')
+ {
+ Regex r = new Regex();
+ r._compile("(?!" + back_slash + "b)", mk);
+ add(r.thePattern);
+ }
+ else if (isOctalString(sp))
+ {
+ int d = sp.c - '0';
+ sp.inc();
+ d = 8 * d + sp.c - '0';
+ StrPos sp2 = new StrPos(sp);
+ sp2.inc();
+ if (isOctalDigit(sp2, false))
+ {
+ sp.inc();
+ d = 8 * d + sp.c - '0';
+ }
+ add(new oneChar((char) d));
+ }
+ else if (sp.dontMatch && sp.c >= '1' && sp.c <= '9')
+ {
+ int iv = sp.c - '0';
+ StrPos s2 = new StrPos(sp);
+ s2.inc();
+ if (!s2.dontMatch && s2.c >= '0' && s2.c <= '9')
+ {
+ iv = 10 * iv + (s2.c - '0');
+ sp.inc();
+ }
+ add(new BackMatch(iv));
+ }
+ else if (sp.dontMatch && sp.c == 'b')
+ {
+ add(new Boundary());
+ }
+ else if (sp.match('\b'))
+ {
+ add(new Boundary());
+ }
+ else if (sp.match('$'))
+ {
+ add(new End(true));
+ }
+ else if (sp.dontMatch && sp.c == 'Z')
+ {
+ add(new End(false));
+ }
+ else if (sp.match('.'))
+ {
+ add(new Any());
+ }
+ else if (sp.incMatch("(??"))
+ {
+ StringBuffer sb = new StringBuffer();
+ StringBuffer sb2 = new StringBuffer();
+ while (!sp.match(')') && !sp.match(':'))
+ {
+ sb.append(sp.c);
+ sp.inc();
+ }
+ if (sp.incMatch(":"))
+ {
+ while (!sp.match(')'))
+ {
+ sb2.append(sp.c);
+ sp.inc();
+ }
+ }
+ String sbs = sb.toString();
+ if (validators.get(sbs) instanceof String)
+ {
+ String pat = (String) validators.get(sbs);
+ Regex r = newRegex();
+ Rthings rth = new Rthings(this);
+ rth.noBackRefs = true;
+ r._compile(pat, rth);
+ add(r.thePattern);
+ }
+ else
+ {
+ Custom cm = new Custom(sb.toString());
+ if (cm.v != null)
+ {
+ Validator v2 = cm.v.arg(sb2.toString());
+ if (v2 != null)
+ {
+ v2.argsave = sb2.toString();
+ String p = cm.v.pattern;
+ cm.v = v2;
+ v2.pattern = p;
+ }
+ Regex r = newRegex();
+ Rthings rth = new Rthings(this);
+ rth.noBackRefs = true;
+ r._compile(cm.v.pattern, rth);
+ cm.sub = r.thePattern;
+ cm.sub.add(new CustomEndpoint(cm));
+ cm.sub.setParent(cm);
+ add(cm);
+ }
+ }
+ }
+ else if (sp.match('('))
+ {
+ mk.parenLevel++;
+ Regex r = newRegex();
+ // r.or = new Or();
+ sp.inc();
+ if (sp.incMatch("?:"))
+ {
+ r.or = new Or();
+ }
+ else if (sp.incMatch("?="))
+ {
+ r.or = new lookAhead(false);
+ }
+ else if (sp.incMatch("?!"))
+ {
+ r.or = new lookAhead(true);
+ }
+ else if (sp.match('?'))
+ {
+ sp.inc();
+ do
+ {
+ if (sp.c == 'i')
+ {
+ mk.ignoreCase = true;
+ }
+ if (sp.c == 'Q')
+ {
+ mk.dontMatchInQuotes = true;
+ }
+ if (sp.c == 'o')
+ {
+ mk.optimizeMe = true;
+ }
+ if (sp.c == 'g')
+ {
+ mk.gFlag = true;
+ }
+ if (sp.c == 's')
+ {
+ mk.sFlag = true;
+ }
+ if (sp.c == 'm')
+ {
+ mk.mFlag = true;
+ }
+ sp.inc();
+ } while (!sp.match(')') && !sp.eos);
+ r = null;
+ mk.parenLevel--;
+ if (sp.eos) // throw new RegSyntax
+ {
+ RegSyntaxError.endItAll("Unclosed ()");
+ }
+ }
+ else
+ { // just ordinary parenthesis
+ r.or = mk.noBackRefs ? new Or() : new OrMark(mk.val++);
+ }
+ if (r != null)
+ {
+ add(r._compile(sp, mk));
+ }
+ }
+ else if (sp.match('^'))
+ {
+ add(new Start(true));
+ }
+ else if (sp.dontMatch && sp.c == 'A')
+ {
+ add(new Start(false));
+ }
+ else if (sp.match('*'))
+ {
+ addMulti(new patInt(0), new patInf());
+ }
+ else if (sp.match('+'))
+ {
+ addMulti(new patInt(1), new patInf());
+ }
+ else if (sp.match('?'))
+ {
+ addMulti(new patInt(0), new patInt(1));
+ }
+ else if (sp.match('{'))
+ {
+ boolean bad = false;
+ StrPos sp2 = new StrPos(sp);
+ // StringBuffer sb = new StringBuffer();
+ sp.inc();
+ patInt i1 = sp.getPatInt();
+ patInt i2 = null;
+ if (sp.match('}'))
+ {
+ i2 = i1;
+ }
+ else
+ {
+ if (!sp.match(','))
+ {
+ /*
+ * RegSyntaxError.endItAll( "String \"{"+i2+ "\" should be followed
+ * with , or }");
+ */
+ bad = true;
+ }
+ sp.inc();
+ if (sp.match('}'))
+ {
+ i2 = new patInf();
+ }
+ else
+ {
+ i2 = sp.getPatInt();
+ }
+ }
+ if (i1 == null || i2 == null)
+ {
+ /*
+ * throw new RegSyntax("Badly formatted Multi: " +"{"+i1+","+i2+"}");
+ */
+ bad = true;
+ }
+ if (bad)
+ {
+ sp.dup(sp2);
+ add(new oneChar(sp.c));
+ }
+ else
+ {
+ addMulti(i1, i2);
+ }
+ }
+ else if (sp.escMatch('x') && next2Hex(sp))
+ {
+ sp.inc();
+ int d = getHexDigit(sp);
+ sp.inc();
+ d = 16 * d + getHexDigit(sp);
+ add(new oneChar((char) d));
+ }
+ else if (sp.escMatch('c'))
+ {
+ sp.inc();
+ if (sp.c < Ctrl.cmap.length)
+ {
+ add(new oneChar(Ctrl.cmap[sp.c]));
+ }
+ else
+ {
+ add(new oneChar(sp.c));
+ }
+ }
+ else if (sp.escMatch('f'))
+ {
+ add(new oneChar((char) 12));
+ }
+ else if (sp.escMatch('a'))
+ {
+ add(new oneChar((char) 7));
+ }
+ else if (sp.escMatch('t'))
+ {
+ add(new oneChar('\t'));
+ }
+ else if (sp.escMatch('n'))
+ {
+ add(new oneChar('\n'));
+ }
+ else if (sp.escMatch('r'))
+ {
+ add(new oneChar('\r'));
+ }
+ else if (sp.escMatch('b'))
+ {
+ add(new oneChar('\b'));
+ }
+ else if (sp.escMatch('e'))
+ {
+ add(new oneChar((char) 27));
+ }
+ else
+ {
+ add(new oneChar(sp.c));
+ if (sp.match(')'))
+ {
+ RegSyntaxError.endItAll("Unmatched right paren in pattern");
+ }
+ }
+ }
+
+ // compiles all Pattern elements, internal method
+ private Pattern _compile(String pat, Rthings mk) throws RegSyntax
+ {
+ minMatch = null;
+ sFlag = mFlag = ignoreCase = gFlag = false;
+ StrPos sp = new StrPos(pat, 0);
+ thePattern = _compile(sp, mk);
+ pt.marks = null;
+ return thePattern;
+ }
+
+ Pattern p = null;
+
+ Or or = null;
+
+ Pattern _compile(StrPos sp, Rthings mk) throws RegSyntax
+ {
+ while (!(sp.eos || (or != null && sp.match(')'))))
+ {
+ compile1(sp, mk);
+ sp.inc();
+ }
+ if (sp.match(')'))
+ {
+ mk.parenLevel--;
+ }
+ else if (sp.eos && mk.parenLevel != 0)
+ {
+ RegSyntaxError.endItAll("Unclosed Parenthesis! lvl=" + mk.parenLevel);
+ }
+ if (or != null)
+ {
+ if (p == null)
+ {
+ p = new NullPattern();
+ }
+ or.addOr(p);
+ return or;
+ }
+ return p == null ? new NullPattern() : p;
+ }
+
+ // add a multi object to the end of the chain
+ // which applies to the last object
+ void addMulti(patInt i1, patInt i2) throws RegSyntax
+ {
+ Pattern last, last2;
+ for (last = p; last != null && last.next != null; last = last.next)
+ {
+ ;
+ }
+ if (last == null || last == p)
+ {
+ last2 = null;
+ }
+ else
+ {
+ for (last2 = p; last2.next != last; last2 = last2.next)
+ {
+ ;
+ }
+ }
+ if (last instanceof Multi && i1.intValue() == 0 && i2.intValue() == 1)
+ {
+ ((Multi) last).matchFewest = true;
+ }
+ else if (last instanceof FastMulti && i1.intValue() == 0
+ && i2.intValue() == 1)
+ {
+ ((FastMulti) last).matchFewest = true;
+ }
+ else if (last instanceof DotMulti && i1.intValue() == 0
+ && i2.intValue() == 1)
+ {
+ ((DotMulti) last).matchFewest = true;
+ }
+ else if (last instanceof Multi || last instanceof DotMulti
+ || last instanceof FastMulti)
+ {
+ throw new RegSyntax("Syntax error.");
+ }
+ else if (last2 == null)
+ {
+ p = mkMulti(i1, i2, p);
+ }
+ else
+ {
+ last2.next = mkMulti(i1, i2, last);
+ }
+ }
+
+ final static Pattern mkMulti(patInt lo, patInt hi, Pattern p)
+ throws RegSyntax
+ {
+ if (p instanceof Any && p.next == null)
+ {
+ return (Pattern) new DotMulti(lo, hi);
+ }
+ return RegOpt.safe4fm(p) ? (Pattern) new FastMulti(lo, hi, p)
+ : (Pattern) new Multi(lo, hi, p);
+ }
+
+ // process the bracket operator
+ Pattern matchBracket(StrPos sp) throws RegSyntax
+ {
+ Bracket ret;
+ if (sp.match('^'))
+ {
+ ret = new Bracket(true);
+ sp.inc();
+ }
+ else
+ {
+ ret = new Bracket(false);
+ }
+ if (sp.match(']'))
+ {
+ // throw new RegSyntax
+ RegSyntaxError.endItAll("Unmatched []");
+ }
+
+ while (!sp.eos && !sp.match(']'))
+ {
+ StrPos s1 = new StrPos(sp);
+ s1.inc();
+ StrPos s1_ = new StrPos(s1);
+ s1_.inc();
+ if (s1.match('-') && !s1_.match(']'))
+ {
+ StrPos s2 = new StrPos(s1);
+ s2.inc();
+ if (!s2.eos)
+ {
+ ret.addOr(new Range(sp.c, s2.c));
+ }
+ sp.inc();
+ sp.inc();
+ }
+ else if (sp.escMatch('Q'))
+ {
+ sp.inc();
+ while (!sp.escMatch('E'))
+ {
+ ret.addOr(new oneChar(sp.c));
+ sp.inc();
+ }
+ }
+ else if (sp.escMatch('d'))
+ {
+ ret.addOr(new Range('0', '9'));
+ }
+ else if (sp.escMatch('s'))
+ {
+ ret.addOr(new oneChar((char) 32));
+ ret.addOr(new Range((char) 8, (char) 10));
+ ret.addOr(new oneChar((char) 13));
+ }
+ else if (sp.escMatch('w'))
+ {
+ ret.addOr(new Range('a', 'z'));
+ ret.addOr(new Range('A', 'Z'));
+ ret.addOr(new Range('0', '9'));
+ ret.addOr(new oneChar('_'));
+ }
+ else if (sp.escMatch('D'))
+ {
+ ret.addOr(new Range((char) 0, (char) 47));
+ ret.addOr(new Range((char) 58, (char) 65535));
+ }
+ else if (sp.escMatch('S'))
+ {
+ ret.addOr(new Range((char) 0, (char) 7));
+ ret.addOr(new Range((char) 11, (char) 12));
+ ret.addOr(new Range((char) 14, (char) 31));
+ ret.addOr(new Range((char) 33, (char) 65535));
+ }
+ else if (sp.escMatch('W'))
+ {
+ ret.addOr(new Range((char) 0, (char) 64));
+ ret.addOr(new Range((char) 91, (char) 94));
+ ret.addOr(new oneChar((char) 96));
+ ret.addOr(new Range((char) 123, (char) 65535));
+ }
+ else if (sp.escMatch('x') && next2Hex(sp))
+ {
+ sp.inc();
+ int d = getHexDigit(sp);
+ sp.inc();
+ d = 16 * d + getHexDigit(sp);
+ ret.addOr(new oneChar((char) d));
+ }
+ else if (sp.escMatch('a'))
+ {
+ ret.addOr(new oneChar((char) 7));
+ }
+ else if (sp.escMatch('f'))
+ {
+ ret.addOr(new oneChar((char) 12));
+ }
+ else if (sp.escMatch('e'))
+ {
+ ret.addOr(new oneChar((char) 27));
+ }
+ else if (sp.escMatch('n'))
+ {
+ ret.addOr(new oneChar('\n'));
+ }
+ else if (sp.escMatch('t'))
+ {
+ ret.addOr(new oneChar('\t'));
+ }
+ else if (sp.escMatch('r'))
+ {
+ ret.addOr(new oneChar('\r'));
+ }
+ else if (sp.escMatch('c'))
+ {
+ sp.inc();
+ if (sp.c < Ctrl.cmap.length)
+ {
+ ret.addOr(new oneChar(Ctrl.cmap[sp.c]));
+ }
+ else
+ {
+ ret.addOr(new oneChar(sp.c));
+ }
+ }
+ else if (isOctalString(sp))
+ {
+ int d = sp.c - '0';
+ sp.inc();
+ d = 8 * d + sp.c - '0';
+ StrPos sp2 = new StrPos(sp);
+ sp2.inc();
+ if (isOctalDigit(sp2, false))
+ {
+ sp.inc();
+ d = 8 * d + sp.c - '0';
+ }
+ ret.addOr(new oneChar((char) d));
+ }
+ else
+ {
+ ret.addOr(new oneChar(sp.c));
+ }
+ sp.inc();
+ }
+ return ret;
+ }
+
+ /**
+ * Converts the stored Pattern to a String -- this is a decompile. Note that
+ * \t and \n will really print out here, Not just the two character
+ * representations. Also be prepared to see some strange output if your
+ * characters are not printable.
+ */
+ public String toString()
+ {
+ if (false && thePattern == null)
+ {
+ return "";
+ }
+ else
+ {
+ StringBuffer sb = new StringBuffer();
+ if (esc != Pattern.ESC)
+ {
+ sb.append("(?e=");
+ sb.append(esc);
+ sb.append(")");
+ }
+ if (gFlag || mFlag || !dotDoesntMatchCR || sFlag || ignoreCase
+ || dontMatchInQuotes || optimized())
+ {
+ sb.append("(?");
+ if (ignoreCase)
+ {
+ sb.append("i");
+ }
+ if (mFlag)
+ {
+ sb.append("m");
+ }
+ if (sFlag || !dotDoesntMatchCR)
+ {
+ sb.append("s");
+ }
+ if (dontMatchInQuotes)
+ {
+ sb.append("Q");
+ }
+ if (optimized())
+ {
+ sb.append("o");
+ }
+ if (gFlag)
+ {
+ sb.append("g");
+ }
+ sb.append(")");
+ }
+ String patstr = thePattern.toString();
+ if (esc != Pattern.ESC)
+ {
+ patstr = reEscape(patstr, Pattern.ESC, esc);
+ }
+ sb.append(patstr);
+ return sb.toString();
+ }
+ }
+
+ // Re-escape Pattern, allows us to use a different escape
+ // character.
+ static String reEscape(String s, char oldEsc, char newEsc)
+ {
+ if (oldEsc == newEsc)
+ {
+ return s;
+ }
+ int i;
+ StringBuffer sb = new StringBuffer();
+ for (i = 0; i < s.length(); i++)
+ {
+ if (s.charAt(i) == oldEsc && i + 1 < s.length())
+ {
+ if (s.charAt(i + 1) == oldEsc)
+ {
+ sb.append(oldEsc);
+ }
+ else
+ {
+ sb.append(newEsc);
+ sb.append(s.charAt(i + 1));
+ }
+ i++;
+ }
+ else if (s.charAt(i) == newEsc)
+ {
+ sb.append(newEsc);
+ sb.append(newEsc);
+ }
+ else
+ {
+ sb.append(s.charAt(i));
+ }
+ }
+ return sb.toString();
+ }
+
+ /**
+ * This method implements FilenameFilter, allowing one to use a Regex to
+ * search through a directory using File.list. There is a FileRegex now that
+ * does this better.
+ *
+ * @see com.stevesoft.pat.FileRegex
+ */
+ public boolean accept(File dir, String s)
+ {
+ return search(s);
+ }
+
+ /** The version of this package */
+ final static public String version()
+ {
+ return "lgpl release 1.5.3";
+ }
+
+ /**
+ * Once this method is called, the state of variables ignoreCase and
+ * dontMatchInQuotes should not be changed as the results will be
+ * unpredictable. However, search and matchAt will run more quickly. Note that
+ * you can check to see if the pattern has been optimized by calling the
+ * optimized() method.
+ * <p>
+ * This method will attempt to rewrite your pattern in a way that makes it
+ * faster (not all patterns execute at the same speed). In general, "(?: ... )"
+ * will be faster than "( ... )" so if you don't need the backreference, you
+ * should group using the former pattern.
+ * <p>
+ * It will also introduce new pattern elements that you can't get to
+ * otherwise, for example if you have a large table of strings, i.e. the
+ * months of the year "(January|February|...)" optimize() will make a
+ * Hashtable that takes it to the next appropriate pattern element --
+ * eliminating the need for a linear search.
+ *
+ * @see com.stevesoft.pat.Regex#optimized
+ * @see com.stevesoft.pat.Regex#ignoreCase
+ * @see com.stevesoft.pat.Regex#dontMatchInQuotes
+ * @see com.stevesoft.pat.Regex#matchAt
+ * @see com.stevesoft.pat.Regex#search
+ */
+ public void optimize()
+ {
+ if (optimized() || thePattern == null)
+ {
+ return;
+ }
+ minMatch = new patInt(0); // thePattern.countMinChars();
+ thePattern = RegOpt.opt(thePattern, ignoreCase, dontMatchInQuotes);
+ skipper = Skip.findSkip(this);
+ // RegOpt.setParents(this);
+ return;
+ }
+
+ Skip skipper;
+
+ /**
+ * This function returns true if the optimize method has been called.
+ */
+ public boolean optimized()
+ {
+ return minMatch != null;
+ }
+
+ /**
+ * A bit of syntactic surgar for those who want to make their code look more
+ * perl-like. To use this initialize your Regex object by saying:
+ *
+ * <pre>
+ * Regex r1 = Regex.perlCode("s/hello/goodbye/");
+ * Regex r2 = Regex.perlCode("s'fish'frog'i");
+ * Regex r3 = Regex.perlCode("m'hello');
+ * </pre>
+ *
+ * The i for ignoreCase is supported in this syntax, as well as m, s, and x.
+ * The g flat is a bit of a special case.
+ * <p>
+ * If you wish to replace all occurences of a pattern, you do not put a 'g' in
+ * the perlCode, but call Regex's replaceAll method.
+ * <p>
+ * If you wish to simply and only do a search for r2's pattern, you can do
+ * this by calling the searchFrom method method repeatedly, or by calling
+ * search repeatedly if the g flag is set.
+ * <p>
+ * Note: Currently perlCode does <em>not</em> support the (?e=#) syntax for
+ * changing the escape character.
+ */
+
+ public static Regex perlCode(String s)
+ {
+ // this file is big enough, see parsePerl.java
+ // for this function.
+ return parsePerl.parse(s);
+ }
+
+ static final char back_slash = '\\';
+
+ /**
+ * Checks to see if there are only literal and no special pattern elements in
+ * this Regex.
+ */
+ public boolean isLiteral()
+ {
+ Pattern x = thePattern;
+ while (x != null)
+ {
+ if (x instanceof oneChar)
+ {
+ ;
+ }
+ else if (x instanceof Skipped)
+ {
+ ;
+ }
+ else
+ {
+ return false;
+ }
+ x = x.next;
+ }
+ return true;
+ }
+
+ /**
+ * You only need to know about this if you are inventing your own pattern
+ * elements.
+ */
+ public patInt countMinChars()
+ {
+ return thePattern.countMinChars();
+ }
+
+ /**
+ * You only need to know about this if you are inventing your own pattern
+ * elements.
+ */
+ public patInt countMaxChars()
+ {
+ return thePattern.countMaxChars();
+ }
+
+ boolean isHexDigit(StrPos sp)
+ {
+ boolean r = !sp.eos
+ && !sp.dontMatch
+ && ((sp.c >= '0' && sp.c <= '9')
+ || (sp.c >= 'a' && sp.c <= 'f') || (sp.c >= 'A' && sp.c <= 'F'));
+ return r;
+ }
+
+ boolean isOctalDigit(StrPos sp, boolean first)
+ {
+ boolean r = !sp.eos && !(first ^ sp.dontMatch) && sp.c >= '0'
+ && sp.c <= '7';
+ return r;
+ }
+
+ int getHexDigit(StrPos sp)
+ {
+ if (sp.c >= '0' && sp.c <= '9')
+ {
+ return sp.c - '0';
+ }
+ if (sp.c >= 'a' && sp.c <= 'f')
+ {
+ return sp.c - 'a' + 10;
+ }
+ return sp.c - 'A' + 10;
+ }
+
+ boolean next2Hex(StrPos sp)
+ {
+ StrPos sp2 = new StrPos(sp);
+ sp2.inc();
+ if (!isHexDigit(sp2))
+ {
+ return false;
+ }
+ sp2.inc();
+ if (!isHexDigit(sp2))
+ {
+ return false;
+ }
+ return true;
+ }
+
+ boolean isOctalString(StrPos sp)
+ {
+ if (!isOctalDigit(sp, true))
+ {
+ return false;
+ }
+ StrPos sp2 = new StrPos(sp);
+ sp2.inc();
+ if (!isOctalDigit(sp2, false))
+ {
+ return false;
+ }
+ return true;
+ }
+}
git://source.jalview.org / jalview.git / blobdiff