Added an option to use charset code point as collation key controlled by a pref, bug 18338.

intl/locale/src/unix/nsCollationUnix.cpp

@@ -32,6 +32,7 @@
 #include "nsIPosixLocale.h"
 #include "nsCOMPtr.h"
 #include "nsFileSpec.h" /* for nsAutoString */
+#include "nsIPref.h"
 //#define DEBUG_UNIX_COLLATION
 
 static NS_DEFINE_IID(kICollationIID, NS_ICOLLATION_IID);
@@ -65,6 +66,7 @@ nsCollationUnix::nsCollationUnix()
 {
   NS_INIT_REFCNT(); 
   mCollation = NULL;
+  mKeyAsCodePoint = PR_FALSE;
 }
 
 nsCollationUnix::~nsCollationUnix() 
@@ -80,6 +82,11 @@ nsresult nsCollationUnix::Initialize(nsILocale* locale)
 
   nsresult res;
 
+  nsCOMPtr<nsIPref> prefs = do_GetService(NS_PREF_PROGID);
+  if (prefs) {
+    res = prefs->GetBoolPref("intl.collationKeyAsCodePoint", &mKeyAsCodePoint);
+  }
+
   mCollation = new nsCollation;
   if (mCollation == NULL) {
     NS_ASSERTION(0, "mCollation creation failed");
@@ -177,11 +184,16 @@ nsresult nsCollationUnix::GetSortKeyLen(const nsCollationStrength strength,
 
   res = mCollation->UnicodeToChar(stringNormalized, &str, mCharset);
   if (NS_SUCCEEDED(res) && str != NULL) {
-    DoSetLocale();
-    // call strxfrm to calculate a key length 
-    int len = strxfrm(NULL, str, 0) + 1;
-    DoRestoreLocale();
-    *outLen = (len == -1) ? 0 : (PRUint32)len;
+    if (mKeyAsCodePoint) {
+      *outLen = nsCRT::strlen(str);
+    }
+    else {
+      DoSetLocale();
+      // call strxfrm to calculate a key length 
+      int len = strxfrm(NULL, str, 0) + 1;
+      DoRestoreLocale();
+      *outLen = (len == -1) ? 0 : (PRUint32)len;
+    }
     PR_Free(str);
   }
 
@@ -202,11 +214,17 @@ nsresult nsCollationUnix::CreateRawSortKey(const nsCollationStrength strength,
 
   res = mCollation->UnicodeToChar(stringNormalized, &str, mCharset);
   if (NS_SUCCEEDED(res) && str != NULL) {
-    DoSetLocale();
-    // call strxfrm to generate a key 
-    int len = strxfrm((char *) key, str, strlen(str));
-    DoRestoreLocale();
-    *outLen = (len == -1) ? 0 : (PRUint32)len;
+    if (mKeyAsCodePoint) {
+      *outLen = nsCRT::strlen(str);
+      nsCRT::memcpy(key, str, *outLen);
+    }
+    else {
+      DoSetLocale();
+      // call strxfrm to generate a key 
+      int len = strxfrm((char *) key, str, strlen(str));
+      DoRestoreLocale();
+      *outLen = (len == -1) ? 0 : (PRUint32)len;
+    }
     PR_Free(str);
   }
 

intl/locale/src/unix/nsCollationUnix.h

@@ -36,6 +36,7 @@ protected:
   nsString      mLocale;
   nsString      mSavedLocale;
   nsString      mCharset;
+  PRBool        mKeyAsCodePoint;
 
   void DoSetLocale();
   void DoRestoreLocale();

modules/libpref/src/init/all.js

@@ -284,6 +284,7 @@ pref("imageblocker.enabled",                false);
 pref("messages.new_window",                 true); // ML obsolete; use mailnews.message_in_thread_window
 pref("intl.accept_languages",               "en");
 pref("intl.accept_charsets",                "iso-8859-1,*,utf-8");
+pref("intl.collationKeyAsCodePoint",        false);
 
 pref("intl.charsetmenu.browser.static",    "iso-8859-1");
 pref("intl.charsetmenu.mailedit",          "iso-8859-1, iso-8859-2, iso-8859-3, iso-8859-4, iso-8859-9, iso-8859-10, iso-8859-14, iso-8859-15, iso-2022-jp, big5, gb2312, euc-kr, utf-8, koi8-r, windows-1251, koi8-u, iso-8859-7, viscii, tis-620, armscii-8");