Bug 655658 - NetUtil.readInputStreamToString should have aCharset argument as optional. r=sdwilsh, sr=bz

js/xpconnect/src/XPCException.cpp

@@ -45,6 +45,7 @@
 #include "nsNetError.h"
 #include "mozStorage.h"
 #include "nsPluginError.h"
+#include "nsIUnicodeDecoder.h"
 
 /***************************************************************************/
 /* Quick and dirty mapping of well known result codes to strings. We only

js/xpconnect/src/xpc.msg

@@ -222,3 +222,5 @@ XPC_MSG_DEF(NS_ERROR_STORAGE_CONSTRAINT             , "SQLite database operation
 /* plugin related codes (from nsPluginError.h) */
 XPC_MSG_DEF(NS_ERROR_PLUGIN_TIME_RANGE_NOT_SUPPORTED, "Clearing site data by time range not supported by plugin")
 
+/* character converter related codes (from nsIUnicodeDecoder.h) */
+XPC_MSG_DEF(NS_ERROR_ILLEGAL_INPUT                  , "The input characters have illegal sequences")

netwerk/base/src/NetUtil.jsm

@@ -274,6 +274,12 @@ const NetUtil = {
      *        The input stream to read from.
      * @param aCount
      *        The number of bytes to read from the stream.
+     * @param aOptions [optional]
+     *        charset
+     *          The character encoding of stream data.
+     *        replacement
+     *          The character to replace unknown byte sequences.
+     *          If unset, it causes an exceptions to be thrown.
      *
      * @return the bytes from the input stream in string form.
      *
@@ -282,9 +288,11 @@ const NetUtil = {
      *         block the calling thread (non-blocking mode only).
      * @throws NS_ERROR_FAILURE if there are not enough bytes available to read
      *         aCount amount of data.
+     * @throws NS_ERROR_ILLEGAL_INPUT if aInputStream has invalid sequences
      */
     readInputStreamToString: function NetUtil_readInputStreamToString(aInputStream,
-                                                                      aCount)
+                                                                      aCount,
+                                                                      aOptions)
     {
         if (!(aInputStream instanceof Ci.nsIInputStream)) {
             let exception = new Components.Exception(
@@ -304,6 +312,33 @@ const NetUtil = {
             throw exception;
         }
 
+        if (aOptions && "charset" in aOptions) {
+          let cis = Cc["@mozilla.org/intl/converter-input-stream;1"].
+                    createInstance(Ci.nsIConverterInputStream);
+          try {
+            // When replacement is set, the character that is unknown sequence 
+            // replaces with aOptions.replacement character.
+            if (!("replacement" in aOptions)) {
+              // aOptions.replacement isn't set.
+              // If input stream has unknown sequences for aOptions.charset,
+              // throw NS_ERROR_ILLEGAL_INPUT.
+              aOptions.replacement = 0;
+            }
+
+            cis.init(aInputStream, aOptions.charset, aCount,
+                     aOptions.replacement);
+            let str = {};
+            cis.readString(-1, str);
+            cis.close();
+            return str.value;
+          }
+          catch (e) {
+            // Adjust the stack so it throws at the caller's location.
+            throw new Components.Exception(e.message, e.result,
+                                           Components.stack.caller, e.data);
+          }
+        }
+
         let sis = Cc["@mozilla.org/scriptableinputstream;1"].
                   createInstance(Ci.nsIScriptableInputStream);
         sis.init(aInputStream);

netwerk/test/unit/test_NetUtil.js

@@ -536,6 +536,59 @@ function test_readInputStreamToString_too_many_bytes()
   run_next_test();
 }
 
+function test_readInputStreamToString_with_charset()
+{
+  const TEST_DATA = "\uff10\uff11\uff12\uff13";
+  const TEST_DATA_UTF8 = "\xef\xbc\x90\xef\xbc\x91\xef\xbc\x92\xef\xbc\x93";
+  const TEST_DATA_SJIS = "\x82\x4f\x82\x50\x82\x51\x82\x52";
+
+  let istream = Cc["@mozilla.org/io/string-input-stream;1"].
+                createInstance(Ci.nsIStringInputStream);
+
+  istream.setData(TEST_DATA_UTF8, TEST_DATA_UTF8.length);
+  do_check_eq(NetUtil.readInputStreamToString(istream,
+                                              TEST_DATA_UTF8.length,
+                                              { charset: "UTF-8"}),
+              TEST_DATA);
+
+  istream.setData(TEST_DATA_SJIS, TEST_DATA_SJIS.length);
+  do_check_eq(NetUtil.readInputStreamToString(istream,
+                                              TEST_DATA_SJIS.length,
+                                              { charset: "Shift_JIS"}),
+              TEST_DATA);
+
+  run_next_test();
+}
+
+function test_readInputStreamToString_invalid_sequence()
+{
+  const TEST_DATA = "\ufffd\ufffd\ufffd\ufffd";
+  const TEST_DATA_UTF8 = "\xaa\xaa\xaa\xaa";
+
+  let istream = Cc["@mozilla.org/io/string-input-stream;1"].
+                createInstance(Ci.nsIStringInputStream);
+
+  istream.setData(TEST_DATA_UTF8, TEST_DATA_UTF8.length);
+  try {
+    NetUtil.readInputStreamToString(istream,
+                                    TEST_DATA_UTF8.length,
+                                    { charset: "UTF-8" });
+    do_throw("should throw!");
+  } catch (e) {
+    do_check_eq(e.result, Cr.NS_ERROR_ILLEGAL_INPUT);
+  }
+
+  istream.setData(TEST_DATA_UTF8, TEST_DATA_UTF8.length);
+  do_check_eq(NetUtil.readInputStreamToString(istream,
+                                              TEST_DATA_UTF8.length, {
+                                                charset: "UTF-8",
+                                                replacement: Ci.nsIConverterInputStream.DEFAULT_REPLACEMENT_CHARACTER}),
+              TEST_DATA);
+
+  run_next_test();
+}
+
+
 ////////////////////////////////////////////////////////////////////////////////
 //// Test Runner
 
@@ -565,6 +618,8 @@ function test_readInputStreamToString_too_many_bytes()
   test_readInputStreamToString_no_bytes_arg,
   test_readInputStreamToString_blocking_stream,
   test_readInputStreamToString_too_many_bytes,
+  test_readInputStreamToString_with_charset,
+  test_readInputStreamToString_invalid_sequence,
 ].forEach(add_test);
 let index = 0;