From: Jakub Skoczen <jakub@indexdata.dk>
Date: Mon, 17 Mar 2014 13:08:14 +0000 (+0100)
Subject: Allow a list of quoted terms
X-Git-Tag: v1.12~5
X-Git-Url: http://sru.miketaylor.org.uk/cgi-bin?a=commitdiff_plain;h=b76b718c1994b348b66a994dcf15c819c4e94617;p=cql-java-moved-to-github.git

Allow a list of quoted terms

Also mixed with  a list of words. To align with C CQL parser.
---

diff --git a/src/main/java/org/z3950/zing/cql/CQLLexer.java b/src/main/java/org/z3950/zing/cql/CQLLexer.java
index d7adeac..669dde7 100644
--- a/src/main/java/org/z3950/zing/cql/CQLLexer.java
+++ b/src/main/java/org/z3950/zing/cql/CQLLexer.java
@@ -70,7 +70,7 @@ public class CQLLexer implements CQLTokenizer {
       }
     //quoted string
     } else if (strchr("\"", c)) { //no single-quotes
-      what = '"';
+      what = TT_STRING;
       //remember quote char
       char mark = c;
       qi++;
diff --git a/src/main/java/org/z3950/zing/cql/CQLParser.java b/src/main/java/org/z3950/zing/cql/CQLParser.java
index a4f2fc8..b435aa3 100644
--- a/src/main/java/org/z3950/zing/cql/CQLParser.java
+++ b/src/main/java/org/z3950/zing/cql/CQLParser.java
@@ -235,13 +235,13 @@ public class CQLParser {
 
 	    debug("non-parenthesised term");
 	    word = matchSymbol("index or term");
-            while (lexer.what() == CQLTokenizer.TT_WORD && !isRelation()) {
+            while (isWordOrString() && !isRelation()) {
               word = word + " " + lexer.value();
-              match(CQLTokenizer.TT_WORD);
+              match(lexer.what());
             }
 
 	    if (!isRelation())
-		break;
+              break;
 
 	    index = word;
 	    String relstr = (lexer.what() == CQLTokenizer.TT_WORD ?
@@ -278,6 +278,11 @@ public class CQLParser {
 
 	return new CQLPrefixNode(name, identifier, node);
     }
+    
+    private boolean isWordOrString() {
+      return CQLTokenizer.TT_WORD == lexer.what() 
+        || CQLTokenizer.TT_STRING == lexer.what();
+    }
 
     private boolean isRelation() {
 	debug("isRelation: checking what()=" + lexer.what() +
@@ -326,12 +331,10 @@ public class CQLParser {
 
 	debug("in matchSymbol()");
 	if (lexer.what() == CQLTokenizer.TT_WORD ||
-	    lexer.what() == '"' ||
+	    lexer.what() == CQLTokenizer.TT_STRING ||
 	    // The following is a complete list of keywords.  Because
 	    // they're listed here, they can be used unquoted as
 	    // indexes, terms, prefix names and prefix identifiers.
-	    // ### Instead, we should ask the lexer whether what we
-	    // have is a keyword, and let the knowledge reside there.
             (allowKeywordTerms &&
 	    lexer.what() == CQLTokenizer.TT_AND ||
 	    lexer.what() == CQLTokenizer.TT_OR ||
diff --git a/src/main/java/org/z3950/zing/cql/CQLTokenizer.java b/src/main/java/org/z3950/zing/cql/CQLTokenizer.java
index 0425bb4..2d8b6b1 100644
--- a/src/main/java/org/z3950/zing/cql/CQLTokenizer.java
+++ b/src/main/java/org/z3950/zing/cql/CQLTokenizer.java
@@ -15,6 +15,7 @@ public interface CQLTokenizer {
   public static final int TT_WORD = -3;
   public static final int TT_NOTHING = -4;
   
+  public final static int TT_STRING = 999;      // quoted string
   public final static int TT_LE     = 1000;	// The "<=" relation
   public final static int TT_GE     = 1001;	// The ">=" relation
   public final static int TT_NE     = 1002;	// The "<>" relation