bug 153, 1066: Convert properties of SMJS bookmark to/from UTF-8.

SpiderMonkey uses UTF-16 and the strings in struct bookmark are in
UTF-8.  Previously, the conversions behaved as if the strings had been
in ISO-8859-1.

SpiderMonkey also supports JS_SetCStringsAreUTF8(), which would make
the existing functions convert between UTF-16 and UTF-8, but that
effect is global so I dare not enable it yet.  Besides, I don't know
if that function works in all the SpiderMonkey versions that ELinks
claims to work with.

src/bookmarks/bookmarks.c

@@ -525,8 +525,16 @@ update_bookmark(struct bookmark *bm, int codepage,
 	return 1;
 }
 
-/* Search for a bookmark with the given title. Search in the given folder
- * or in the root if folder is NULL. */
+/** Search for a bookmark with the given title.  The search does not
+ * recurse into subfolders.
+ *
+ * @param folder
+ *   Search in this folder.  NULL means search in the root.
+ *
+ * @param title
+ *   Search for this title.  Must be in UTF-8 and not NULL.
+ *
+ * @return The bookmark, or NULL if not found.  */
 struct bookmark *
 get_bookmark_by_name(struct bookmark *folder, unsigned char *title)
 {
@@ -535,7 +543,6 @@ get_bookmark_by_name(struct bookmark *folder, unsigned char *title)
 
 	lh = folder ? &folder->child : &bookmarks;
 
-	/** @todo Bug 153: bookmark->title should be UTF-8 */
 	foreach (bookmark, *lh)
 		if (!strcmp(bookmark->title, title)) return bookmark;
 

src/scripting/smjs/bookmarks.c

@@ -8,6 +8,7 @@
 
 #include "bookmarks/bookmarks.h"
 #include "ecmascript/spidermonkey-shared.h"
+#include "intl/charsets.h"
 #include "main/event.h"
 #include "scripting/smjs/core.h"
 #include "scripting/smjs/elinks_object.h"
@@ -78,6 +79,60 @@ static const JSPropertySpec bookmark_props[] = {
 
 static JSObject *smjs_get_bookmark_folder_object(struct bookmark *bookmark);
 
+/** Convert a string retrieved from struct bookmark to a jsval.
+ *
+ * @return JS_TRUE if successful.  On error, report the error and
+ * return JS_FALSE.  */
+static JSBool
+bookmark_string_to_jsval(JSContext *ctx, const unsigned char *str, jsval *vp)
+{
+	JSString *jsstr = utf8_to_jsstring(ctx, str, -1);
+
+	if (jsstr == NULL)
+		return JS_FALSE;
+	*vp = STRING_TO_JSVAL(jsstr);
+	return JS_TRUE;
+}
+
+/** Convert a jsval to a string and store it in struct bookmark.
+ *
+ * @param ctx
+ *   Context for memory allocations and error reports.
+ * @param val
+ *   The @c jsval that should be converted.
+ * @param[in,out] result
+ *   A string allocated with mem_alloc().
+ *   On success, this function frees the original string, if any.
+ *
+ * @return JS_TRUE if successful.  On error, report the error to
+ * SpiderMonkey and return JS_FALSE.  */
+static JSBool
+jsval_to_bookmark_string(JSContext *ctx, jsval val, unsigned char **result)
+{
+	JSString *jsstr = NULL;
+	unsigned char *str;
+
+	/* jsstring_to_utf8() might GC; protect the string to come.  */
+	if (!JS_AddNamedRoot(ctx, &jsstr, "jsval_to_bookmark_string"))
+		return JS_FALSE;
+
+	jsstr = JS_ValueToString(ctx, val);
+	if (jsstr == NULL) {
+		JS_RemoveRoot(ctx, &jsstr);
+		return JS_FALSE;
+	}
+
+	str = jsstring_to_utf8(ctx, jsstr, NULL);
+	if (str == NULL) {
+		JS_RemoveRoot(ctx, &jsstr);
+		return JS_FALSE;
+	}
+
+	JS_RemoveRoot(ctx, &jsstr);
+	mem_free_set(result, str);
+	return JS_TRUE;
+}
+
 /* @bookmark_class.getProperty */
 static JSBool
 bookmark_get_property(JSContext *ctx, JSObject *obj, jsval id, jsval *vp)
@@ -102,17 +157,9 @@ bookmark_get_property(JSContext *ctx, JSObject *obj, jsval id, jsval *vp)
 
 	switch (JSVAL_TO_INT(id)) {
 	case BOOKMARK_TITLE:
-		/** @todo Bug 153: bookmark->title should be UTF-8 */
-		*vp = STRING_TO_JSVAL(JS_NewStringCopyZ(smjs_ctx,
-	                                                bookmark->title));
-
-		return JS_TRUE;
+		return bookmark_string_to_jsval(ctx, bookmark->title, vp);
 	case BOOKMARK_URL:
-		/** @todo Bug 1066: bookmark->url should be UTF-8 */
-		*vp = STRING_TO_JSVAL(JS_NewStringCopyZ(smjs_ctx,
-	                                                bookmark->url));
-
-		return JS_TRUE;
+		return bookmark_string_to_jsval(ctx, bookmark->url, vp);
 	case BOOKMARK_CHILDREN:
 		*vp = OBJECT_TO_JSVAL(smjs_get_bookmark_folder_object(bookmark));
 
@@ -149,24 +196,10 @@ bookmark_set_property(JSContext *ctx, JSObject *obj, jsval id, jsval *vp)
 		return JS_FALSE;
 
 	switch (JSVAL_TO_INT(id)) {
-	case BOOKMARK_TITLE: {
-		JSString *jsstr = JS_ValueToString(smjs_ctx, *vp);
-		unsigned char *str = JS_GetStringBytes(jsstr);
-
-		/** @todo Bug 153: bookmark->title should be UTF-8 */
-		mem_free_set(&bookmark->title, stracpy(str));
-
-		return JS_TRUE;
-	}
-	case BOOKMARK_URL: {
-		JSString *jsstr = JS_ValueToString(smjs_ctx, *vp);
-		unsigned char *str = JS_GetStringBytes(jsstr);
-
-		/** @todo Bug 1066: bookmark->url should be UTF-8 */
-		mem_free_set(&bookmark->url, stracpy(str));
-
-		return JS_TRUE;
-	}
+	case BOOKMARK_TITLE:
+		return jsval_to_bookmark_string(ctx, *vp, &bookmark->title);
+	case BOOKMARK_URL:
+		return jsval_to_bookmark_string(ctx, *vp, &bookmark->url);
 	default:
 		/* Unrecognized integer property ID; someone is using
 		 * the object as an array.  SMJS builtin classes (e.g.
@@ -209,7 +242,7 @@ bookmark_folder_get_property(JSContext *ctx, JSObject *obj, jsval id, jsval *vp)
 {
 	struct bookmark *bookmark;
 	struct bookmark *folder;
-	unsigned char *title;
+	unsigned char *title = NULL;
 
 	/* This can be called if @obj if not itself an instance of the
 	 * appropriate class but has one in its prototype chain.  Fail
@@ -222,14 +255,15 @@ bookmark_folder_get_property(JSContext *ctx, JSObject *obj, jsval id, jsval *vp)
 
 	*vp = JSVAL_NULL;
 
-	title = JS_GetStringBytes(JS_ValueToString(ctx, id));
-	if (!title) return JS_TRUE;
+	if (!jsval_to_bookmark_string(ctx, id, &title))
+		return JS_FALSE;
 
 	bookmark = get_bookmark_by_name(folder, title);
 	if (bookmark) {
 		*vp = OBJECT_TO_JSVAL(smjs_get_bookmark_object(bookmark));
 	}
 
+	mem_free(title);
 	return JS_TRUE;
 }
 

src/scripting/smjs/core.c

@@ -8,6 +8,7 @@
 
 #include "config/home.h"
 #include "ecmascript/spidermonkey-shared.h"
+#include "intl/charsets.h"
 #include "main/module.h"
 #include "osdep/osdep.h"
 #include "scripting/scripting.h"
@@ -163,3 +164,163 @@ cleanup_smjs(struct module *module)
 	JS_DestroyContext(smjs_ctx);
 	spidermonkey_runtime_release();
 }
+
+/** Convert a UTF-8 string to a JSString.
+ *
+ * @param ctx
+ *   Allocate the string in this JSContext.
+ * @param[in] str
+ *   The input string that should be converted.
+ * @param[in] length
+ *   Length of @a str in bytes, or -1 if it is null-terminated.
+ *
+ * @return the new string.  On error, report the error to SpiderMonkey
+ * and return NULL.  */
+JSString *
+utf8_to_jsstring(JSContext *ctx, const unsigned char *str, int length)
+{
+	size_t in_bytes;
+	const unsigned char *in_end;
+	size_t utf16_alloc;
+	jschar *utf16;
+	size_t utf16_used;
+	JSString *jsstr;
+
+	if (length == -1)
+		in_bytes = strlen(str);
+	else
+		in_bytes = length;
+
+	/* Each byte of input can become at most one UTF-16 unit.
+	 * Check whether the multiplication could overflow.  */
+	assert(!needs_utf16_surrogates(UCS_REPLACEMENT_CHARACTER));
+	if (in_bytes > ((size_t) -1) / sizeof(jschar)) {
+		JS_ReportAllocationOverflow(ctx);
+		return NULL;
+	}
+	utf16_alloc = in_bytes;
+	/* Don't use fmem_alloc here because long strings could
+	 * exhaust the stack.  */
+	utf16 = mem_alloc(utf16_alloc * sizeof(jschar));
+	if (utf16 == NULL) {
+		JS_ReportOutOfMemory(ctx);
+		return NULL;
+	}
+
+	in_end = str + in_bytes;
+
+	utf16_used = 0;
+	for (;;) {
+		unicode_val_T unicode;
+
+		unicode = utf8_to_unicode((unsigned char **) &str, in_end);
+		if (unicode == UCS_NO_CHAR)
+			break;
+
+		if (needs_utf16_surrogates(unicode)) {
+			assert(utf16_alloc - utf16_used >= 2);
+			if_assert_failed { mem_free(utf16); return NULL; }
+			utf16[utf16_used++] = get_utf16_high_surrogate(unicode);
+			utf16[utf16_used++] = get_utf16_low_surrogate(unicode);
+		} else {
+			assert(utf16_alloc - utf16_used >= 1);
+			if_assert_failed { mem_free(utf16); return NULL; }
+			utf16[utf16_used++] = unicode;
+		}
+	}
+
+	jsstr = JS_NewUCString(ctx, utf16, utf16_used);
+	mem_free(utf16);
+	return jsstr;
+}
+
+/** Convert a jschar array to UTF-8 and append it to struct string.
+ * Replace misused surrogate codepoints with UCS_REPLACEMENT_CHARACTER.
+ *
+ * @param[in,out] utf8
+ *   The function appends characters to this UTF-8 string.
+ *
+ * @param[in] utf16
+ *   Pointer to the first element in an array of jschars.
+ *
+ * @param[i] len
+ *   Number of jschars in the @a utf16 array.
+ *
+ * @return @a utf8 if successful, or NULL if not.  */
+static struct string *
+add_jschars_to_utf8_string(struct string *utf8,
+			   const jschar *utf16, size_t len)
+{
+	size_t pos;
+
+	for (pos = 0; pos < len; ) {
+		unicode_val_T unicode = utf16[pos++];
+
+		if (is_utf16_surrogate(unicode)) {
+			if (is_utf16_high_surrogate(unicode)
+			    && pos < len
+			    && is_utf16_low_surrogate(utf16[pos])) {
+				unicode = join_utf16_surrogates(unicode,
+								utf16[pos++]);
+			} else {
+				unicode = UCS_REPLACEMENT_CHARACTER;
+			}
+		}
+
+		if (unicode == 0) {
+			if (!add_char_to_string(utf8, '\0'))
+				return NULL;
+		} else {
+			if (!add_to_string(utf8, encode_utf8(unicode)))
+				return NULL;
+		}
+	}
+
+	return utf8;
+}
+
+/** Convert a JSString to a UTF-8 string.
+ *
+ * @param ctx
+ *   For reporting errors.
+ * @param[in] jsstr
+ *   The input string that should be converted.  Must not be NULL.
+ * @param[out] length
+ *   Optional.  The number of bytes in the returned string,
+ *   not counting the terminating null.
+ *
+ * @return the new string, which the caller must eventually free
+ * with mem_free().  On error, report the error to SpiderMonkey
+ * and return NULL; *@a length is then undefined.  */
+unsigned char *
+jsstring_to_utf8(JSContext *ctx, JSString *jsstr, int *length)
+{
+	size_t utf16_len;
+	const jschar *utf16;
+	struct string utf8;
+
+	utf16_len = JS_GetStringLength(jsstr);
+	utf16 = JS_GetStringChars(jsstr); /* stays owned by jsstr */
+	if (utf16 == NULL) {
+		/* JS_GetStringChars doesn't have a JSContext *
+		 * parameter so it can't report the error
+		 * (and can't collect garbage either).  */
+		JS_ReportOutOfMemory(ctx);
+		return NULL;
+	}
+
+	if (!init_string(&utf8)) {
+		JS_ReportOutOfMemory(ctx);
+		return NULL;
+	}
+
+	if (!add_jschars_to_utf8_string(&utf8, utf16, utf16_len)) {
+		done_string(&utf8);
+		JS_ReportOutOfMemory(ctx);
+		return NULL;
+	}
+
+	if (length)
+		*length = utf8.length;
+	return utf8.source;
+}

src/scripting/smjs/core.h

@@ -16,4 +16,9 @@ void alert_smjs_error(unsigned char *msg);
 void init_smjs(struct module *module);
 void cleanup_smjs(struct module *module);
 
+JSString *utf8_to_jsstring(JSContext *ctx, const unsigned char *str,
+			   int length);
+unsigned char *jsstring_to_utf8(JSContext *ctx, JSString *jsstr,
+				int *length);
+
 #endif