Add mrb_utf8_from_locale, mrb_utf8_free, mrb_locale_from_utf8, mrb_lo…

…cale_free. Just works for windows.

include/mruby.h

@@ -259,6 +259,18 @@ mrb_value mrb_str_new_cstr(mrb_state*, const char*);
 mrb_value mrb_str_new_static(mrb_state *mrb, const char *p, size_t len);
 #define mrb_str_new_lit(mrb, lit) mrb_str_new_static(mrb, (lit), sizeof(lit) - 1)
 
+#ifdef _WIN32
+char* mrb_utf8_from_locale(const char *p, size_t len);
+char* mrb_locale_from_utf8(const char *p, size_t len);
+#define mrb_locale_free(p) free(p)
+#define mrb_utf8_free(p) free(p)
+#else
+char* mrb_utf8_from_locale(p, l) (p)
+char* mrb_locale_from_utf8(p, l) (p)
+#define mrb_locale_free(p)
+#define mrb_utf8_free(p)
+#endif
+
 mrb_state* mrb_open(void);
 mrb_state* mrb_open_allocf(mrb_allocf, void *ud);
 void mrb_close(mrb_state*);

mrbgems/mruby-bin-mirb/tools/mirb/mirb.c

@@ -297,6 +297,8 @@ main(int argc, char **argv)
 
 
   while (TRUE) {
+    char *utf8;
+
 #ifndef ENABLE_READLINE
     print_cmdline(code_block_open);
 
@@ -333,14 +335,19 @@ main(int argc, char **argv)
       strcpy(ruby_code, last_code_line);
     }
 
+	utf8 = mrb_utf8_from_locale(ruby_code, -1);
+	if (!utf8) abort();
+
     /* parse code */
     parser = mrb_parser_new(mrb);
-    parser->s = ruby_code;
-    parser->send = ruby_code + strlen(ruby_code);
+    parser->s = utf8;
+    parser->send = utf8 + strlen(utf8);
     parser->lineno = cxt->lineno;
     mrb_parser_parse(parser, cxt);
     code_block_open = is_code_block_open(parser);
 
+    mrb_utf8_free(utf8);
+
     if (code_block_open) {
       /* no evaluation of code */
     }

mrbgems/mruby-bin-mruby/tools/mruby/mruby.c

@@ -193,7 +193,11 @@ main(int argc, char **argv)
 
   ARGV = mrb_ary_new_capa(mrb, args.argc);
   for (i = 0; i < args.argc; i++) {
-    mrb_ary_push(mrb, ARGV, mrb_str_new(mrb, args.argv[i], strlen(args.argv[i])));
+    char* utf8 = mrb_utf8_from_locale(args.argv[i], -1);
+    if (utf8) {
+      mrb_ary_push(mrb, ARGV, mrb_str_new(mrb, utf8, strlen(utf8)));
+      mrb_utf8_free(utf8);
+    }
   }
   mrb_define_global_const(mrb, "ARGV", ARGV);
 
@@ -216,9 +220,12 @@ main(int argc, char **argv)
       v = mrb_load_file_cxt(mrb, args.rfp, c);
     }
     else {
+      char* utf8 = mrb_utf8_from_locale(args.cmdline, -1);
+      if (!utf8) abort();
       mrbc_filename(mrb, c, "-e");
       mrb_gv_set(mrb, zero_sym, mrb_str_new_lit(mrb, "-e"));
-      v = mrb_load_string_cxt(mrb, args.cmdline, c);
+      v = mrb_load_string_cxt(mrb, utf8, c);
+      mrb_utf8_free(utf8);
     }
   }
   mrbc_context_free(mrb, c);

mrbgems/mruby-print/src/print.c

@@ -1,17 +1,21 @@
 #include "mruby.h"
 #include "mruby/string.h"
 #include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
 
 static void
 printstr(mrb_state *mrb, mrb_value obj)
 {
-  char *s;
-  int len;
-
   if (mrb_string_p(obj)) {
-    s = RSTRING_PTR(obj);
-    len = RSTRING_LEN(obj);
-    fwrite(s, len, 1, stdout);
+    struct RString *str;
+    char* ptr;
+    str = mrb_str_ptr(obj);
+    ptr = mrb_locale_from_utf8(RSTRING_PTR(obj), RSTRING_LEN(obj));
+    if (ptr) {
+      fwrite(ptr, strlen(ptr), 1, stdout);
+      mrb_locale_free(ptr);
+    }
   }
 }
 

src/string.c

@@ -54,6 +54,69 @@ typedef struct mrb_shared_string {
 static mrb_value str_replace(mrb_state *mrb, struct RString *s1, struct RString *s2);
 static mrb_value mrb_str_subseq(mrb_state *mrb, mrb_value str, mrb_int beg, mrb_int len);
 
+#ifdef _WIN32
+#include <windows.h>
+
+char*
+mrb_utf8_from_locale(const char *str, size_t len)
+{
+  wchar_t* wcsp;
+  char* mbsp;
+  size_t mbssize, wcssize;
+
+  if (len == 0)
+    return strdup("");
+  if (len == -1)
+	len = strlen(str);
+  wcssize = MultiByteToWideChar(GetACP(), 0, str, len,  NULL, 0);
+  wcsp = (wchar_t*) malloc((wcssize + 1) * sizeof(wchar_t));
+  if (!wcsp)
+    return NULL;
+  wcssize = MultiByteToWideChar(GetACP(), 0, str, len, wcsp, wcssize + 1);
+  wcsp[wcssize] = 0;
+
+  mbssize = WideCharToMultiByte(CP_UTF8, 0, (LPCWSTR) wcsp, -1, NULL, 0, NULL, NULL);
+  mbsp = (char*) malloc((mbssize + 1));
+  if (!mbsp) {
+    free(wcsp);
+    return NULL;
+  }
+  mbssize = WideCharToMultiByte(CP_UTF8, 0, (LPCWSTR) wcsp, -1, mbsp, mbssize, NULL, NULL);
+  mbsp[mbssize] = 0;
+  free(wcsp);
+  return mbsp;
+}
+
+char*
+mrb_locale_from_utf8(const char *utf8, size_t len)
+{
+  wchar_t* wcsp;
+  char* mbsp;
+  size_t mbssize, wcssize;
+
+  if (len == 0)
+    return strdup("");
+  if (len == -1)
+	len = strlen(utf8);
+  wcssize = MultiByteToWideChar(CP_UTF8, 0, utf8, len,  NULL, 0);
+  wcsp = (wchar_t*) malloc((wcssize + 1) * sizeof(wchar_t));
+  if (!wcsp)
+    return NULL;
+  wcssize = MultiByteToWideChar(CP_UTF8, 0, utf8, len, wcsp, wcssize + 1);
+  wcsp[wcssize] = 0;
+  mbssize = WideCharToMultiByte(GetACP(), 0, (LPCWSTR) wcsp, -1, NULL, 0, NULL, NULL);
+  mbsp = (char*) malloc((mbssize + 1));
+  if (!mbsp) {
+    free(wcsp);
+    return NULL;
+  }
+  mbssize = WideCharToMultiByte(GetACP(), 0, (LPCWSTR) wcsp, -1, mbsp, mbssize, NULL, NULL);
+  mbsp[mbssize] = 0;
+  free(wcsp);
+  return mbsp;
+}
+#endif
+
 #define RESIZE_CAPA(s,capacity) do {\
   if (STR_EMBED_P(s)) {\
     if (RSTRING_EMBED_LEN_MAX < (capacity)) {\