Patch: FYI: Possessive quantifiers in gnu.regexp (PR libgcj/20435)
Ziga Mahkovec
ziga.mahkovec@klika.si
Wed Jun 1 22:53:00 GMT 2005
I checked this in on the trunk.
This patch adds possessive quantifier support to gnu.regexp. Possessive
quantifiers (?+, *+, ++, {n,m}+) were introduced in Java 1.4. The
patch adds the capability to JAVA_1_4 syntax and also replaces PERL5
syntax as the default java.util.regex.Pattern syntax (note that
possessive quantifiers are currently the only difference between the two
syntaxes).
--
Ziga
2005-06-01 Ziga Mahkovec <ziga.mahkovec@klika.si>
PR libgcj/20435:
* gnu/regexp/RESyntax.java (RE_POSSESSIVE_OPS): New field.
(static): Add possessive matching to JAVA_1_4 syntax.
* gnu/regexp/RETokenRepeated.java (possessive): New field.
(makePossessive, isPossessive): New methods.
(match): Don't back off during possessive matching.
* gnu/regexp/RE.java (initalize): Accept possessive quantifier.
* java/util/regex/Pattern.java (constructor): Switch syntax from PERL5
to JAVA_1_4.
Index: gnu/regexp/RE.java
===================================================================
RCS file: /cvsroot/classpath/classpath/gnu/regexp/RE.java,v
retrieving revision 1.5
diff -u -p -r1.5 RE.java
--- gnu/regexp/RE.java 24 May 2005 08:24:35 -0000 1.5
+++ gnu/regexp/RE.java 28 May 2005 12:28:20 -0000
@@ -629,20 +629,29 @@ public class RE extends REToken {
currentToken = setRepeated(currentToken,0,Integer.MAX_VALUE,index);
}
- // ONE-OR-MORE REPEAT OPERATOR
+ // ONE-OR-MORE REPEAT OPERATOR / POSSESSIVE MATCHING OPERATOR
// + | \+ depending on RE_BK_PLUS_QM
// not available if RE_LIMITED_OPS is set
else if ((unit.ch == '+') && !syntax.get(RESyntax.RE_LIMITED_OPS) && (!syntax.get(RESyntax.RE_BK_PLUS_QM) ^ (unit.bk || quot))) {
if (currentToken == null)
throw new REException(getLocalizedMessage("repeat.no.token"),REException.REG_BADRPT,index);
- if (currentToken instanceof RETokenRepeated)
- throw new REException(getLocalizedMessage("repeat.chained"),REException.REG_BADRPT,index);
- if (currentToken instanceof RETokenWordBoundary || currentToken instanceof RETokenWordBoundary)
+
+ // Check for possessive matching on RETokenRepeated
+ if (currentToken instanceof RETokenRepeated) {
+ RETokenRepeated tokenRep = (RETokenRepeated)currentToken;
+ if (syntax.get(RESyntax.RE_POSSESSIVE_OPS) && !tokenRep.isPossessive() && !tokenRep.isStingy())
+ tokenRep.makePossessive();
+ else
+ throw new REException(getLocalizedMessage("repeat.chained"),REException.REG_BADRPT,index);
+
+ }
+ else if (currentToken instanceof RETokenWordBoundary || currentToken instanceof RETokenWordBoundary)
throw new REException(getLocalizedMessage("repeat.assertion"),REException.REG_BADRPT,index);
- if (currentToken.getMinimumLength() == 0)
+ else if (currentToken.getMinimumLength() == 0)
throw new REException(getLocalizedMessage("repeat.empty.token"),REException.REG_BADRPT,index);
- currentToken = setRepeated(currentToken,1,Integer.MAX_VALUE,index);
+ else
+ currentToken = setRepeated(currentToken,1,Integer.MAX_VALUE,index);
}
// ZERO-OR-ONE REPEAT OPERATOR / STINGY MATCHING OPERATOR
@@ -655,13 +664,14 @@ public class RE extends REToken {
// Check for stingy matching on RETokenRepeated
if (currentToken instanceof RETokenRepeated) {
- if (syntax.get(RESyntax.RE_STINGY_OPS) && !((RETokenRepeated)currentToken).isStingy())
- ((RETokenRepeated)currentToken).makeStingy();
- else
- throw new REException(getLocalizedMessage("repeat.chained"),REException.REG_BADRPT,index);
- }
- else if (currentToken instanceof RETokenWordBoundary || currentToken instanceof RETokenWordBoundary)
- throw new REException(getLocalizedMessage("repeat.assertion"),REException.REG_BADRPT,index);
+ RETokenRepeated tokenRep = (RETokenRepeated)currentToken;
+ if (syntax.get(RESyntax.RE_STINGY_OPS) && !tokenRep.isStingy() && !tokenRep.isPossessive())
+ tokenRep.makeStingy();
+ else
+ throw new REException(getLocalizedMessage("repeat.chained"),REException.REG_BADRPT,index);
+ }
+ else if (currentToken instanceof RETokenWordBoundary || currentToken instanceof RETokenWordBoundary)
+ throw new REException(getLocalizedMessage("repeat.assertion"),REException.REG_BADRPT,index);
else
currentToken = setRepeated(currentToken,0,1,index);
}
Index: gnu/regexp/RESyntax.java
===================================================================
RCS file: /cvsroot/classpath/classpath/gnu/regexp/RESyntax.java,v
retrieving revision 1.1
diff -u -p -r1.1 RESyntax.java
--- gnu/regexp/RESyntax.java 7 Mar 2004 23:58:54 -0000 1.1
+++ gnu/regexp/RESyntax.java 28 May 2005 12:28:20 -0000
@@ -197,7 +197,12 @@ public final class RESyntax implements S
*/
public static final int RE_CHAR_CLASS_ESC_IN_LISTS = 24;
- private static final int BIT_TOTAL = 25;
+ /**
+ * Syntax bit. Possessive matching is allowed (++, *+, ?+, {x,y}+).
+ */
+ public static final int RE_POSSESSIVE_OPS = 25;
+
+ private static final int BIT_TOTAL = 26;
/**
* Predefined syntax.
@@ -425,6 +430,7 @@ public final class RESyntax implements S
RE_SYNTAX_JAVA_1_4 = new RESyntax(RE_SYNTAX_PERL5)
// XXX
+ .set(RE_POSSESSIVE_OPS) // *+,?+,++,{}+
.makeFinal();
}
Index: gnu/regexp/RETokenRepeated.java
===================================================================
RCS file: /cvsroot/classpath/classpath/gnu/regexp/RETokenRepeated.java,v
retrieving revision 1.1
diff -u -p -r1.1 RETokenRepeated.java
--- gnu/regexp/RETokenRepeated.java 7 Mar 2004 23:58:54 -0000 1.1
+++ gnu/regexp/RETokenRepeated.java 28 May 2005 12:28:20 -0000
@@ -44,6 +44,7 @@ final class RETokenRepeated extends RETo
private REToken token;
private int min,max;
private boolean stingy;
+ private boolean possessive;
RETokenRepeated(int subIndex, REToken token, int min, int max) {
super(subIndex);
@@ -61,6 +62,16 @@ final class RETokenRepeated extends RETo
boolean isStingy() {
return stingy;
}
+
+ /** Sets possessive matching mode to true. */
+ void makePossessive() {
+ possessive = true;
+ }
+
+ /** Queries if this token has possessive matching enabled. */
+ boolean isPossessive() {
+ return possessive;
+ }
/**
* The minimum length of a repeated token is the minimum length
@@ -172,6 +183,8 @@ final class RETokenRepeated extends RETo
}
}
// else did not match rest of the tokens, try again on smaller sample
+ // or break out when performing possessive matching
+ if (possessive) break;
}
if (allResults != null) {
mymatch.assignFrom(allResults); // does this get all?
Index: java/util/regex/Pattern.java
===================================================================
RCS file: /cvsroot/classpath/classpath/java/util/regex/Pattern.java,v
retrieving revision 1.11
diff -u -p -r1.11 Pattern.java
--- java/util/regex/Pattern.java 24 May 2005 08:24:35 -0000 1.11
+++ java/util/regex/Pattern.java 28 May 2005 12:28:21 -0000
@@ -84,8 +84,7 @@ public final class Pattern implements Se
// if ((flags & UNICODE_CASE) != 0) gnuFlags =
// if ((flags & CANON_EQ) != 0) gnuFlags =
- // Eventually there will be such a thing as JDK 1_4 syntax
- RESyntax syntax = RESyntax.RE_SYNTAX_PERL5;
+ RESyntax syntax = RESyntax.RE_SYNTAX_JAVA_1_4;
if ((flags & UNIX_LINES) != 0)
{
// Use a syntax set with \n for linefeeds?
More information about the Java-patches
mailing list