some work on string

git-svn-id: http://svn.macosforge.org/repository/ruby/MacRuby/branches/icu@3581 23306eb0-4c56-4727-a40e-e92c0eb68959

encoding.c

@@ -9,9 +9,12 @@
  * Copyright (C) 2000 Information-technology Promotion Agency, Japan
  */
 
-#include "encoding.h"
 #include <string.h>
 
+#include "ruby.h"
+#include "ruby/encoding.h"
+#include "encoding.h"
+
 VALUE rb_cEncoding;
 
 static rb_encoding_t *default_internal = NULL;
@@ -241,9 +244,6 @@ Init_PreEncoding(void)
     default_internal = rb_encodings[ENCODING_UTF8];
 }
 
-VALUE
-mr_enc_s_is_compatible(VALUE klass, SEL sel, VALUE str1, VALUE str2);
-
 void
 Init_Encoding(void)
 {
@@ -266,10 +266,8 @@ Init_Encoding(void)
     rb_objc_define_method(CLASS_OF(rb_cEncoding), "aliases",
 	    mr_enc_s_aliases, 0);
     //rb_define_singleton_method(rb_cEncoding, "find", enc_find, 1);
-    // it's defined on Encoding, but it requires String's internals so it's
-    // defined with String
     rb_objc_define_method(CLASS_OF(rb_cEncoding), "compatible?",
-	    mr_enc_s_is_compatible, 2);
+	    mr_enc_s_is_compatible, 2); // in string.c
 
     //rb_define_method(rb_cEncoding, "_dump", enc_dump, -1);
     //rb_define_singleton_method(rb_cEncoding, "_load", enc_load, 1);

encoding.h

@@ -16,8 +16,6 @@
 extern "C" {
 #endif
 
-#include "ruby.h"
-
 #if defined(__cplusplus)
 # include "unicode/unistr.h"
 #else
@@ -84,6 +82,18 @@ rb_klass_is_rstr(VALUE klass)
 
 #define IS_RSTR(x) (rb_klass_is_rstr(*(VALUE *)x))
 
+static inline void
+rstr_modify(VALUE str)
+{
+    const long mask = RBASIC(str)->flags;
+    if ((mask & FL_FREEZE) == FL_FREEZE) {
+        rb_raise(rb_eRuntimeError, "can't modify frozen/immutable string");
+    }
+    if ((mask & FL_TAINT) == FL_TAINT && rb_safe_level() >= 4) {
+        rb_raise(rb_eSecurityError, "Insecure: can't modify string");
+    }
+}
+
 typedef struct {
     long start_offset_in_bytes;
     long end_offset_in_bytes;
@@ -276,6 +286,8 @@ void rb_str_get_uchars(VALUE str, UChar **chars_p, long *chars_len_p,
 	bool *need_free_p);
 long rb_str_chars_len(VALUE str);
 
+VALUE mr_enc_s_is_compatible(VALUE klass, SEL sel, VALUE str1, VALUE str2);
+
 // Return a string object appropriate for bstr_ calls. This does nothing for
 // data/binary RubyStrings.
 VALUE rb_str_bstr(VALUE str);

parse.y

@@ -5067,36 +5067,34 @@ yycompile(struct parser_params *parser, const char *f, int line)
 }
 #endif /* !RIPPER */
 
+struct lex_get_str_context {
+    VALUE str;
+    UChar *chars;
+    long chars_len;
+};
+
 static VALUE
 lex_get_str(struct parser_params *parser, VALUE s)
 {
-    long beg = 0, len; 
-    const long n = CFStringGetLength((CFStringRef)s);
+    struct lex_get_str_context *ctx = (struct lex_get_str_context *)s;
+
+    long beg = 0;
     if (lex_gets_ptr > 0) {
-	if (n == lex_gets_ptr) {
+	if (ctx->chars_len == lex_gets_ptr) {
 	    return Qnil;
 	}
 	beg += lex_gets_ptr;
     }
 
-    CFRange search_range;  
-    if (CFStringFindCharacterFromSet((CFStringRef)s, 
-		CFCharacterSetGetPredefined(kCFCharacterSetNewline),
-		CFRangeMake(beg, n - beg),
-		0,
-		&search_range)) {
-	lex_gets_ptr = search_range.location + 1;
-	len = search_range.location - beg;
-    }
-    else {
-	lex_gets_ptr = n;
-	len = lex_gets_ptr - beg;	
+    lex_gets_ptr = ctx->chars_len;
+    for (long i = beg; i < ctx->chars_len; i++) {
+	if (ctx->chars[i] == '\n') {
+	    lex_gets_ptr = i + 1;
+	    break;
+	}
     }
 
-    CFStringRef subs = CFStringCreateWithSubstring(NULL, (CFStringRef)s, 
-	    CFRangeMake(beg, lex_gets_ptr - beg));
-    CFMakeCollectable(subs);
-    return (VALUE)subs;
+    return rb_unicode_str_new(&ctx->chars[beg], lex_gets_ptr - beg);
 }
 
 static VALUE
@@ -5118,15 +5116,27 @@ rb_compile_string(const char *f, VALUE s, int line)
     return rb_parser_compile_string(rb_parser_new(), f, s, line);
 }
 
-NODE*
+NODE *
 rb_parser_compile_string(VALUE vparser, const char *f, VALUE s, int line)
 {
     struct parser_params *parser;
     Data_Get_Struct(vparser, struct parser_params, parser);
 
+    UChar *chars = NULL;
+    long chars_len = 0;
+    bool need_free = false;
+    rb_str_get_uchars(s, &chars, &chars_len, &need_free);
+    assert(!need_free);
+
+    struct lex_get_str_context *ctx = (struct lex_get_str_context *)
+	xmalloc(sizeof(struct lex_get_str_context));
+    GC_WB(&ctx->str, s);
+    ctx->chars = chars;
+    ctx->chars_len = chars_len;
+
     lex_gets = lex_get_str;
     lex_gets_ptr = 0;
-    GC_WB(&lex_input, s);
+    GC_WB(&lex_input, ctx);
     lex_pbeg = lex_p = lex_pend = 0;
     compile_for_eval = rb_parse_in_eval();
 

ruby.c

@@ -553,30 +553,33 @@ proc_options(int argc, char **argv, struct cmdline_options *opt)
 
 	  case 'e':
 	    forbid_setid("-e");
-	    if (!*++s) {
+	    if (*++s == '\0') {
 		s = argv[1];
-		argc--, argv++;
+		argc--;
+		argv++;
 	    }
-	    if (!s) {
+	    if (s == NULL) {
 		rb_raise(rb_eRuntimeError, "no code specified for -e");
 	    }
-	    if (!opt->e_script) {
-		opt->e_script = rb_str_new(0, 0);
-		if (opt->script == 0)
+	    if (opt->e_script == 0) {
+		opt->e_script = rb_str_new(NULL, 0);
+		if (opt->script == NULL) {
 		    opt->script = "-e";
+		}
 	    }
 	    rb_str_cat2(opt->e_script, s);
 	    rb_str_cat2(opt->e_script, "\n");
 	    break;
 
 	  case 'r':
 	    forbid_setid("-r");
-	    if (*++s) {
+	    if (*++s != '\0') {
 		add_modules(s);
 	    }
 	    else if (argv[1]) {
 		add_modules(argv[1]);
-		argc--, argv++;
+		argc--;
+		argv++;
 	    }
 	    break;