Changeset 13170 in webkit

Timestamp:

Mar 6, 2006 11:59:09 AM (18 years ago)

Author:

ap

Message:

Patch by David Carson, reviewed by Darin.

• Fixed ​http://bugzilla.opendarwin.org/show_bug.cgi?id=7582 c_utility.cpp contains CFString OS X platform-dependent code; should use ICU

Tested with test case from:
​http://bugzilla.opendarwin.org/show_bug.cgi?id=5163

• bindings/c_utility.cpp (convertUTF8ToUTF16): Changed to using Unicode converter from ICU, and manual Latin-1 conversion.
• icu/unicode/ucnv.h: Copied from WebCore.
• icu/unicode/ucnv_err.h: Ditto.
• icu/unicode/uenum.h: Ditto.

Location:

trunk/JavaScriptCore

Files:

• ChangeLog (modified) (1 diff)
• bindings/c/c_utility.cpp (modified) (2 diffs)
• icu/unicode/ucnv.h (copied) (copied from trunk/WebCore/icu/unicode/ucnv.h)
• icu/unicode/ucnv_err.h (copied) (copied from trunk/WebCore/icu/unicode/ucnv_err.h)
• icu/unicode/uenum.h (copied) (copied from trunk/WebCore/icu/unicode/uenum.h)

trunk/JavaScriptCore/ChangeLog

@@ +1 @@
+2006-03-06  David Carson <dacarson@gmail.com>
+
+        Reviewed by Darin, landed by ap.
+
+        - Fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=7582
+        c_utility.cpp contains CFString OS X platform-dependent code; should use ICU
+
+        Tested with test case from:
+        http://bugzilla.opendarwin.org/show_bug.cgi?id=5163
+
+        * bindings/c_utility.cpp
+        (convertUTF8ToUTF16): Changed to using Unicode converter from ICU, and manual Latin-1 conversion.
+        * icu/unicode/ucnv.h: Copied from WebCore.
+        * icu/unicode/ucnv_err.h: Ditto.
+        * icu/unicode/uenum.h: Ditto.
+
 2006-03-05  Darin Adler  <darin@apple.com>
 

trunk/JavaScriptCore/bindings/c/c_utility.cpp

@@ -27 +27 @@
 #include "config.h"
 #include "c_utility.h"
+#include <unicode/ucnv.h>
 
 #include "JSType.h"
@@ -46 +47 @@
 {
     assert(UTF8Chars || UTF8Length == 0);
+    assert(UTF16Chars);
     
     if (UTF8Length == -1)
         UTF8Length = strlen(UTF8Chars);
-
-    CFStringRef stringRef = CFStringCreateWithBytes(NULL, (const UInt8*)UTF8Chars, (CFIndex)UTF8Length, kCFStringEncodingUTF8, false);
-    
+        
+    // UTF16Length maximum length is the length of the UTF8 string, plus one to include terminator
+    // Without the plus one, it will convert ok, but a warning is generated from the converter as
+    // there is not enough room for a terminating character.
+    *UTF16Length = UTF8Length + 1; 
+        
+    *UTF16Chars = 0;
+    UErrorCode status = U_ZERO_ERROR;
+    UConverter* conv = ucnv_open("utf8", &status);
+    if (U_SUCCESS(status)) { 
+        *UTF16Chars = (NPUTF16 *)malloc(sizeof(NPUTF16) * (*UTF16Length));
+        ucnv_setToUCallBack(conv, UCNV_TO_U_CALLBACK_STOP, 0, 0, 0, &status);
+        ucnv_toUChars(conv, *UTF16Chars, *UTF16Length, UTF8Chars, UTF8Length, &status); 
+        ucnv_close(conv);
+    } 
+    
+    // Check to see if the conversion was successful
     // Some plugins return invalid UTF-8 in NPVariantType_String, see <http://bugzilla.opendarwin.org/show_bug.cgi?id=5163>
-    if (!stringRef)
-        stringRef = CFStringCreateWithBytes(NULL, (const UInt8*)UTF8Chars, (CFIndex)UTF8Length, kCFStringEncodingISOLatin1, false);
-
-    // There is no "bad data" for kCFStringEncodingISOLatin1. It is unlikely that the plugin was really sending text in this encoding,
+    // There is no "bad data" for latin1. It is unlikely that the plugin was really sending text in this encoding,
     // but it should have used UTF-8, and now we are simply avoiding a crash.
-    assert(stringRef);
-
-    *UTF16Length = (unsigned int)CFStringGetLength(stringRef);
-    *UTF16Chars = (NPUTF16 *)malloc(sizeof(NPUTF16) * (*UTF16Length));
-
-    // Convert the string to UTF16.
-    CFRange range = { 0, *UTF16Length };
-    CFStringGetCharacters(stringRef, range, (UniChar *)*UTF16Chars);
-    CFRelease(stringRef);
+    if (!U_SUCCESS(status)) {
+        if (!*UTF16Chars)   // If the memory wasn't allocated, allocate it.
+            *UTF16Chars = (NPUTF16 *)malloc(sizeof(NPUTF16) * (*UTF16Length));
+ 
+        for (unsigned i = 0; i < *UTF16Length; i++)
+            (*UTF16Chars)[i] = UTF8Chars[i] & 0xFF;
+
+    }
 }