Changeset 520

Timestamp:

01/31/09 15:15:37 (3 years ago)

Author:

mooneer

Message:

Regular expressions can be specified with "r/.../[i]" syntax (ticket #79)

Location:

interpreter/trunk

Files:

• ChangeLog (modified) (1 diff)
• docs/kite-language.texi (modified) (1 diff)
• modules/System/regex.kt (modified) (1 diff)
• tests/Makefile.am (modified) (1 diff)
• tests/objs/loader/loader_notfound.kt (modified) (1 diff)
• tests/objs/regex/regex_syntax.kt (added)
• tests/objs/regex/regex_syntax.kt.out (added)
• vm/kite_lexer.l (modified) (3 diffs)
• vm/kite_parser.y (modified) (2 diffs)

interpreter/trunk/ChangeLog

@@ +1 @@
+1/31/2009:
+    * Regular expressions can be specified with "r/.../[i]" syntax (ticket #79)
+
+1/30/2009:
+    * Compile errors/warnings fixed when compiling with GCC 4.2. (ticket #82)
+    * Fixed infinite loop during string|split with an empty-string delimiter. (ticket #80)
+    * Fixed capture issue in the regular expression engine when using '?' next to parenthesis (ticket #81)
+    
 1/29/2009:
     * System.digest merged into main tree (thanks to Michael J. Edgar).
     * System.list|sort method added (credit: Michael J. Edgar).
-    
+
 1/1/2009:
     * 1.0.0 released.

interpreter/trunk/docs/kite-language.texi

@@ -88 +88 @@
 System.math.range object (see kdoc documentation) that represents all the
 numbers from one integer to another (possibly skipping numbers along the way).
+
+Regular expressions are also built-into the language syntax as of 1.0.1. Regular
+expressions begin with "r/" and end with either "/" (case-sensitive) or "/i"
+(case-insensitive). This is equivalent to creating a new System.regex object
+with the appropriate parameters.
 
 @node Declarations

interpreter/trunk/modules/System/regex.kt

@@ -42 +42 @@
 ];
 
+construct(
+    regex /[Regular expression to use.]/,
+    c_insensitive /[Case insensitivity on/off]/
+) /[Construct new regular expression object.]/
+[
+    this.internal_regex_obj = make __internal._regex(regex);
+    this.case_insensitive = c_insensitive;
+];
+
 method match(
     str /[String to search.]/

interpreter/trunk/tests/Makefile.am

@@ -256 +256 @@
 ./objs/regex/regex_boundary.kt \
 ./objs/regex/regex_boundary.kt.out \
+./objs/regex/regex_syntax.kt \
+./objs/regex/regex_syntax.kt.out \
 ./objs/base64/base64_encode.kt \
 ./objs/base64/base64_encode.kt.out \

interpreter/trunk/tests/objs/loader/loader_notfound.kt

@@ -1 +1 @@
 import "System.vm.loader";
 run [
-    System.vm.loader|loadClass("System.exceptions.TypeMismatch2");
+    System.vm.loader|loadClass("Kite.ThisClassDoesNotExist");
 ] catch [
     __exc|str()|print();

interpreter/trunk/vm/kite_lexer.l

@@ -28 +28 @@
 %}
 
-%x operator str sym docstr
+%x operator str sym docstr regex
 %option noyywrap
 %option bison-bridge bison-locations reentrant
@@ -105 +105 @@
 false { fillBoth(yyscanner); yylval->intValue = 0; return BOOL_VALUE; }
 
+"r/" {
+    kite_compiler_t *compiler = (kite_compiler_t*)yyextra;
+        fillFirst(yyscanner);
+        BEGIN(regex);
+        compiler->curStr = strdup("");
+}
+
 \" {
         kite_compiler_t *compiler = (kite_compiler_t*)yyextra;
@@ -117 +124 @@
         BEGIN(docstr);
         compiler->curStr = strdup("");
+}
+
+<regex>[^/]+ {
+    kite_compiler_t *compiler = (kite_compiler_t*)yyextra;
+        APPEND_STRING(yytext);
+        compiler->currentCol += strlen(yytext);
+}
+
+<regex>"\\/" {
+    kite_compiler_t *compiler = (kite_compiler_t*)yyextra;
+        APPEND_STRING("/");
+        compiler->currentCol += strlen(yytext);
+}
+
+<regex>"/i" {
+    kite_compiler_t *compiler = (kite_compiler_t*)yyextra;
+        fillLast(yyscanner);
+        BEGIN(INITIAL);
+        yylval->stringValue = compiler->curStr;
+        return REGEX_VALUE_CI;
+}
+
+<regex>"/" {
+    kite_compiler_t *compiler = (kite_compiler_t*)yyextra;
+        fillLast(yyscanner);
+        BEGIN(INITIAL);
+        yylval->stringValue = compiler->curStr;
+        return REGEX_VALUE;
 }
 

interpreter/trunk/vm/kite_parser.y

@@ -154 +154 @@
 %token RETURN
 %token VERSION_KEYWORD
+%token REGEX_VALUE
+%token REGEX_VALUE_CI
 
 /* Needed since we can't quite lay out the associativity using the rules yet. */
@@ -1047 +1049 @@
         kite_dereference_object(range_class);
         kite_dereference_object(range_object);
+        $<opValue>$ = NULL;
+    }
+    | REGEX_VALUE {
+        kite_compiler_t *compiler = (kite_compiler_t*)yyget_extra(parm);
+        kite_thread_t *thd = compiler->thd;
+        kite_object_t *param_list = kite_new_list(thd);
+        kite_object_t *regex_obj = kite_dereference_and_load(thd, "System.regex");
+        
+        kite_append_list(thd, param_list, kite_new_string(thd, $<stringValue>1));
+        
+        COMPILE_INSTRUCTION(kite_compile_push(regex_obj), @1.first_line);
+        COMPILE_INSTRUCTION(kite_compile_push(param_list), @1.first_line);
+        COMPILE_INSTRUCTION(kite_compile_make(), @1.first_line);
+        
+        free($<stringValue>1);
+        $<opValue>$ = NULL;
+    }
+    | REGEX_VALUE_CI {
+        kite_compiler_t *compiler = (kite_compiler_t*)yyget_extra(parm);
+        kite_thread_t *thd = compiler->thd;
+        kite_object_t *param_list = kite_new_list(thd);
+        kite_object_t *regex_obj = kite_dereference_and_load(thd, "System.regex");
+        
+        kite_append_list(thd, param_list, kite_new_string(thd, $<stringValue>1));
+        kite_append_list(thd, param_list, kite_new_boolean(thd, TRUE));
+        
+        COMPILE_INSTRUCTION(kite_compile_push(regex_obj), @1.first_line);
+        COMPILE_INSTRUCTION(kite_compile_push(param_list), @1.first_line);
+        COMPILE_INSTRUCTION(kite_compile_make(), @1.first_line);
+        
+        free($<stringValue>1);
         $<opValue>$ = NULL;
     }