02_utf8_flag.t, remove Char.h, and replace Char.xs with the following…

… code http://use.perl.org/comments.pl?sid=44457&cid=71604
jozef · Jan 29, 2010 · 0b52a4c · 0b52a4c
1 parent 50adeb9
commit 0b52a4c
Show file tree

Hide file tree

Showing 3 changed files with 43 additions and 53 deletions.
diff --git a/lib/XML/Char.h b/lib/XML/Char.h
diff --git a/lib/XML/Char.xs b/lib/XML/Char.xs
@@ -3,66 +3,45 @@
 #include "XSUB.h"
 
 #include "ppport.h"
-#include "Char.h"
 
+static UV
+octet_to_uvuni(const U8 *s, STRLEN *retlen)
+{
+    *retlen = 1;
+    return (UV) *s;
+}
 
-MODULE = XML::Char		PACKAGE = XML::Char		
+MODULE = XML::Char    PACKAGE = XML::Char
 
-int
+void
 _valid_xml_string(string)
     SV* string;
-    
+
     PREINIT:
         STRLEN len;
         U8 * bytes;
         int in_range;
         int range_index;
-        
+
         STRLEN ret_len;
         UV     uniuv;
-    CODE:
-        bytes  = (U8*)SvPV(string, len);
-        if (!is_utf8_string(bytes, len)) {
-            // warn("no utf8\n");
-
-            RETVAL = 0;
-        }
-        else {
-            // by default return true (ex. empty string)
-            RETVAL = 1;
-
-            // loop through all UTF-8 characters and make sure they are in allowed ranges
-            while (len > 0) {
-                // get unicode character value
-                uniuv = utf8_to_uvuni(bytes, &ret_len);
-                // warn("code: 0x%X len: %d\n", uniuv, ret_len);
-                bytes += ret_len;
-                len   -= ret_len;
-
-                // loop through allowed ranges and check if the character is in any of them
-                range_index = 0;
-                in_range = 0;
-                while (xml_ranges_from[range_index] != 0) {
-                    // rangers are sorted so if the unicode value is smaller than current range_from then it is not in any range
-                    if (uniuv < xml_ranges_from[range_index]) {
-                        break;
-                    }
-                    // if the unicode value fall in this range it's valid
-                    if ((uniuv >= xml_ranges_from[range_index]) && (uniuv <= xml_ranges_to[range_index])) {
-                        // in the range
-                        in_range = 1;
-                        break;
-                    }
-                    range_index++;
-                }
-
-                // if the current character is not in allowed ranges return false
-                if (!in_range) {
-                    RETVAL = 0;
-                    break;
-                }
-            }
+        UV     (*next_chr)(const U8 *s, STRLEN *retlen);
+
+    PPCODE:
+        bytes    = (U8*)SvPV(string, len);
+        next_chr = SvUTF8(string) ? &utf8_to_uvuni : &octet_to_uvuni;
+
+        while (len > 0) {
+            uniuv = (*next_chr)(bytes, &ret_len);
+            bytes += ret_len;
+            len   -= ret_len;
+
+            if (
+                (uniuv < 0x20) && (uniuv != 0x9) && (uniuv != 0xA) && (uniuv != 0xD)
+                || (uniuv >  0xD7FF) && (uniuv <  0xE000)
+                || (uniuv >  0xFFFD) && (uniuv < 0x10000)
+                || (uniuv > 0x1FFFF)
+            ) XSRETURN_NO;
         }
-
-    OUTPUT:
-        RETVAL
+
+        XSRETURN_YES;
diff --git a/t/02_utf8_flag.t b/t/02_utf8_flag.t
@@ -0,0 +1,14 @@
+#!/usr/bin/perl
+
+use strict;
+use warnings;
+use utf8 ();
+
+use Test::More tests => 2;
+use XML::Char;
+
+my $str = "\xC3";
+is( XML::Char->valid($str), !!1, "accept U+00C3 with UTF8 flag off" );
+
+utf8::upgrade($str);
+is( XML::Char->valid($str), !!1, "accept U+00C3 with UTF8 flag on" );