Bug 665706 - Remove historic special handling for semicolons in URIs. r=rjesup
authorJulian Reschke <julian.reschke@gmx.de>
Mon, 22 Aug 2011 16:51:52 +0200
changeset 76701 1a09781a5480a2c8ae6f193a326cbf34167bc52b
parent 76694 ac57320746e8e9ec9dd3091740f84899d1d688b2
child 76702 f56c271dec3505cdeb356fb46266f7dde34855f0
push idunknown
push userunknown
push dateunknown
reviewersrjesup
bugs665706
milestone9.0a1
Bug 665706 - Remove historic special handling for semicolons in URIs. r=rjesup
modules/libjar/nsJARURI.cpp
netwerk/base/public/nsIURL.idl
netwerk/base/public/nsIURLParser.idl
netwerk/base/src/nsStandardURL.cpp
netwerk/base/src/nsStandardURL.h
netwerk/base/src/nsURLHelper.cpp
netwerk/base/src/nsURLParsers.cpp
netwerk/test/TestURLParser.cpp
netwerk/test/unit/test_bug429347.js
netwerk/test/urltest.cpp
xpcom/io/nsEscape.cpp
--- a/modules/libjar/nsJARURI.cpp
+++ b/modules/libjar/nsJARURI.cpp
@@ -580,29 +580,16 @@ nsJARURI::GetFilePath(nsACString& filePa
 
 NS_IMETHODIMP
 nsJARURI::SetFilePath(const nsACString& filePath)
 {
     return mJAREntry->SetFilePath(filePath);
 }
 
 NS_IMETHODIMP
-nsJARURI::GetParam(nsACString& param)
-{
-    param.Truncate();
-    return NS_OK;
-}
-
-NS_IMETHODIMP
-nsJARURI::SetParam(const nsACString& param)
-{
-    return NS_ERROR_NOT_AVAILABLE;
-}
-
-NS_IMETHODIMP
 nsJARURI::GetQuery(nsACString& query)
 {
     return mJAREntry->GetQuery(query);
 }
 
 NS_IMETHODIMP
 nsJARURI::SetQuery(const nsACString& query)
 {
--- a/netwerk/base/public/nsIURL.idl
+++ b/netwerk/base/public/nsIURL.idl
@@ -40,17 +40,16 @@
 #include "nsIURI.idl"
 
 /**
  * The nsIURL interface provides convenience methods that further
  * break down the path portion of nsIURI:
  *
  * http://host/directory/fileBaseName.fileExtension?query
  * http://host/directory/fileBaseName.fileExtension#ref
- * http://host/directory/fileBaseName.fileExtension;param
  *            \          \                       /
  *             \          -----------------------
  *              \                   |          /
  *               \               fileName     /
  *                ----------------------------
  *                            |
  *                        filePath
  */
@@ -66,23 +65,16 @@ interface nsIURL : nsIURI
      * URL.  For example, the filePath of "http://host/foo/bar.html#baz"
      * is "/foo/bar.html".
      *
      * Some characters may be escaped.
      */
     attribute AUTF8String filePath;
 
     /**
-     * Returns the parameters specified after the ; in the URL. 
-     *
-     * Some characters may be escaped.
-     */
-    attribute AUTF8String param;
-
-    /**
      * Returns the query portion (the part after the "?") of the URL.
      * If there isn't one, an empty string is returned.
      *
      * Some characters may be escaped.
      */
     attribute AUTF8String query;
 
 
--- a/netwerk/base/public/nsIURLParser.idl
+++ b/netwerk/base/public/nsIURLParser.idl
@@ -90,24 +90,23 @@ interface nsIURLParser : nsISupports
     /**
      * serverinfo = <hostname>:<port>
      */
     void parseServerInfo (in string serverinfo,            in long serverinfoLen,
                           out unsigned long hostnamePos,  out long hostnameLen,
                           out long port);
 
     /**
-     * ParsePath breaks the path string up into its 4 major components: a file path,
-     * a param string, a query string, and a reference string.
+     * ParsePath breaks the path string up into its 3 major components: a file path,
+     * a query string, and a reference string.
      *
-     * path = <filepath>;<param>?<query>#<ref>
+     * path = <filepath>?<query>#<ref>
      */
     void parsePath       (in string path,                  in long pathLen,
                           out unsigned long filepathPos,  out long filepathLen,
-                          out unsigned long paramPos,     out long paramLen,
                           out unsigned long queryPos,     out long queryLen,
                           out unsigned long refPos,       out long refLen);
 
     /**
      * ParseFilePath breaks the file path string up into: the directory portion,
      * file base name, and file extension.
      *
      * filepath = <directory><basename>.<extension>
--- a/netwerk/base/src/nsStandardURL.cpp
+++ b/netwerk/base/src/nsStandardURL.cpp
@@ -395,17 +395,16 @@ nsStandardURL::Clear()
     mHostEncoding = eEncoding_ASCII;
 
     mPath.Reset();
     mFilepath.Reset();
     mDirectory.Reset();
     mBasename.Reset();
 
     mExtension.Reset();
-    mParam.Reset();
     mQuery.Reset();
     mRef.Reset();
 
     InvalidateCache();
 }
 
 void
 nsStandardURL::InvalidateCache(PRBool invalidateCachedFile)
@@ -506,26 +505,26 @@ nsresult
 nsStandardURL::BuildNormalizedSpec(const char *spec)
 {
     // Assumptions: all member URLSegments must be relative the |spec| argument
     // passed to this function.
 
     // buffers for holding escaped url segments (these will remain empty unless
     // escaping is required).
     nsCAutoString encUsername, encPassword, encHost, encDirectory,
-      encBasename, encExtension, encParam, encQuery, encRef;
+      encBasename, encExtension, encQuery, encRef;
     PRBool useEncUsername, useEncPassword, useEncHost, useEncDirectory,
-      useEncBasename, useEncExtension, useEncParam, useEncQuery, useEncRef;
+      useEncBasename, useEncExtension, useEncQuery, useEncRef;
     nsCAutoString portbuf;
 
     //
     // escape each URL segment, if necessary, and calculate approximate normalized
     // spec length.
     //
-    // [scheme://][username[:password]@]host[:port]/path[;param][?query_string][#ref]
+    // [scheme://][username[:password]@]host[:port]/path[?query_string][#ref]
 
     PRUint32 approxLen = 0;
 
     // the scheme is already ASCII
     if (mScheme.mLen > 0)
         approxLen += mScheme.mLen + 3; // includes room for "://";
 
     // encode URL segments; convert UTF-8 to origin charset and possibly escape.
@@ -551,19 +550,16 @@ nsStandardURL::BuildNormalizedSpec(const
         approxLen += 1; // reserve space for possible leading '/' - may not be needed
         // Should just use mPath?  These are pessimistic, and thus waste space
         approxLen += encoder.EncodeSegmentCount(spec, mDirectory, esc_Directory,     encDirectory, useEncDirectory, 1);
         approxLen += encoder.EncodeSegmentCount(spec, mBasename,  esc_FileBaseName,  encBasename,  useEncBasename);
         approxLen += encoder.EncodeSegmentCount(spec, mExtension, esc_FileExtension, encExtension, useEncExtension, 1);
 
         // These next ones *always* add their leading character even if length is 0
         // Handles items like "http://#"
-        // ;param
-        if (mParam.mLen >= 0)
-            approxLen += 1 + encoder.EncodeSegmentCount(spec, mParam,     esc_Param,         encParam,     useEncParam);
         // ?query
         if (mQuery.mLen >= 0)
             approxLen += 1 + queryEncoder.EncodeSegmentCount(spec, mQuery, esc_Query,        encQuery,     useEncQuery);
         // #ref
         if (mRef.mLen >= 0)
             approxLen += 1 + encoder.EncodeSegmentCount(spec, mRef,       esc_Ref,           encRef,       useEncRef);
     }
 
@@ -673,20 +669,16 @@ nsStandardURL::BuildNormalizedSpec(const
 
         if (mExtension.mLen >= 0) {
             buf[i++] = '.';
             i = AppendSegmentToBuf(buf, i, spec, mExtension, &encExtension, useEncExtension);
         }
         // calculate corrected filepath length
         mFilepath.mLen = i - mFilepath.mPos;
 
-        if (mParam.mLen >= 0) {
-            buf[i++] = ';';
-            i = AppendSegmentToBuf(buf, i, spec, mParam, &encParam, useEncParam);
-        }
         if (mQuery.mLen >= 0) {
             buf[i++] = '?';
             i = AppendSegmentToBuf(buf, i, spec, mQuery, &encQuery, useEncQuery);
         }
         if (mRef.mLen >= 0) {
             buf[i++] = '#';
             i = AppendSegmentToBuf(buf, i, spec, mRef, &encRef, useEncRef);
         }
@@ -833,23 +825,21 @@ nsStandardURL::ParseURL(const char *spec
 
 nsresult
 nsStandardURL::ParsePath(const char *spec, PRUint32 pathPos, PRInt32 pathLen)
 {
     LOG(("ParsePath: %s pathpos %d len %d\n",spec,pathPos,pathLen));
 
     nsresult rv = mParser->ParsePath(spec + pathPos, pathLen,
                                      &mFilepath.mPos, &mFilepath.mLen,
-                                     &mParam.mPos, &mParam.mLen,
                                      &mQuery.mPos, &mQuery.mLen,
                                      &mRef.mPos, &mRef.mLen);
     if (NS_FAILED(rv)) return rv;
 
     mFilepath.mPos += pathPos;
-    mParam.mPos += pathPos;
     mQuery.mPos += pathPos;
     mRef.mPos += pathPos;
 
     if (mFilepath.mLen > 0) {
         rv = mParser->ParseFilePath(spec + mFilepath.mPos, mFilepath.mLen,
                                     &mDirectory.mPos, &mDirectory.mLen,
                                     &mBasename.mPos, &mBasename.mLen,
                                     &mExtension.mPos, &mExtension.mLen);
@@ -1203,17 +1193,16 @@ nsStandardURL::SetSpec(const nsACString 
         LOG((" username  = (%u,%d)\n", mUsername.mPos,  mUsername.mLen));
         LOG((" password  = (%u,%d)\n", mPassword.mPos,  mPassword.mLen));
         LOG((" hostname  = (%u,%d)\n", mHost.mPos,      mHost.mLen));
         LOG((" path      = (%u,%d)\n", mPath.mPos,      mPath.mLen));
         LOG((" filepath  = (%u,%d)\n", mFilepath.mPos,  mFilepath.mLen));
         LOG((" directory = (%u,%d)\n", mDirectory.mPos, mDirectory.mLen));
         LOG((" basename  = (%u,%d)\n", mBasename.mPos,  mBasename.mLen));
         LOG((" extension = (%u,%d)\n", mExtension.mPos, mExtension.mLen));
-        LOG((" param     = (%u,%d)\n", mParam.mPos,     mParam.mLen));
         LOG((" query     = (%u,%d)\n", mQuery.mPos,     mQuery.mLen));
         LOG((" ref       = (%u,%d)\n", mRef.mPos,       mRef.mLen));
     }
 #endif
     return rv;
 }
 
 NS_IMETHODIMP
@@ -1628,17 +1617,16 @@ nsStandardURL::SetPath(const nsACString 
         mSpec.Cut(mPath.mPos + 1, mPath.mLen - 1);
         // these contain only a '/'
         mPath.mLen = 1;
         mDirectory.mLen = 1;
         mFilepath.mLen = 1;
         // these are no longer defined
         mBasename.mLen = -1;
         mExtension.mLen = -1;
-        mParam.mLen = -1;
         mQuery.mLen = -1;
         mRef.mLen = -1;
     }
     return NS_OK;
 }
 
 NS_IMETHODIMP
 nsStandardURL::Equals(nsIURI *unknownOther, PRBool *result)
@@ -1679,18 +1667,17 @@ nsStandardURL::EqualsInternal(nsIURI *un
     // URIs different
     if (!SegmentIs(mScheme, other->mSpec.get(), other->mScheme) ||
         // Check for host manually, since conversion to file will
         // ignore the host!
         !SegmentIs(mHost, other->mSpec.get(), other->mHost) ||
         !SegmentIs(mQuery, other->mSpec.get(), other->mQuery) ||
         !SegmentIs(mUsername, other->mSpec.get(), other->mUsername) ||
         !SegmentIs(mPassword, other->mSpec.get(), other->mPassword) ||
-        Port() != other->Port() ||
-        !SegmentIs(mParam, other->mSpec.get(), other->mParam)) {
+        Port() != other->Port()) {
         // No need to compare files or other URI parts -- these are different
         // beasties
         *result = PR_FALSE;
         return NS_OK;
     }
 
     if (refHandlingMode == eHonorRef &&
         !SegmentIs(mRef, other->mSpec.get(), other->mRef)) {
@@ -1790,17 +1777,16 @@ nsStandardURL::CloneInternal(nsStandardU
     clone->mUsername = mUsername;
     clone->mPassword = mPassword;
     clone->mHost = mHost;
     clone->mPath = mPath;
     clone->mFilepath = mFilepath;
     clone->mDirectory = mDirectory;
     clone->mBasename = mBasename;
     clone->mExtension = mExtension;
-    clone->mParam = mParam;
     clone->mQuery = mQuery;
     clone->mRef = mRef;
     clone->mOriginCharset = mOriginCharset;
     clone->mURLType = mURLType;
     clone->mParser = mParser;
     clone->mFile = mFile;
     clone->mHostA = mHostA ? nsCRT::strdup(mHostA) : nsnull;
     clone->mMutable = PR_TRUE;
@@ -2015,17 +2001,17 @@ nsStandardURL::GetCommonBaseSpec(nsIURI 
     while ((*thisIndex == *thatIndex) && *thisIndex)
     {
         thisIndex++;
         thatIndex++;
     }
 
     // backup to just after previous slash so we grab an appropriate path
     // segment such as a directory (not partial segments)
-    // todo:  also check for file matches which include '?', '#', and ';'
+    // todo:  also check for file matches which include '?' and '#'
     while ((thisIndex != startCharPos) && (*(thisIndex-1) != '/'))
         thisIndex--;
 
     // grab spec from beginning to thisIndex
     aResult = Substring(mSpec, mScheme.mPos, thisIndex - mSpec.get());
 
     NS_RELEASE(stdurl2);
     return rv;
@@ -2096,17 +2082,17 @@ nsStandardURL::GetRelativeSpec(nsIURI *u
     while ((*thisIndex == *thatIndex) && *thisIndex)
     {
         thisIndex++;
         thatIndex++;
     }
 
     // backup to just after previous slash so we grab an appropriate path
     // segment such as a directory (not partial segments)
-    // todo:  also check for file matches with '#', '?' and ';'
+    // todo:  also check for file matches with '#' and '?'
     while ((*(thatIndex-1) != '/') && (thatIndex != startCharPos))
         thatIndex--;
 
     const char *limit = mSpec.get() + mFilepath.mPos + mFilepath.mLen;
 
     // need to account for slashes and add corresponding "../"
     for (; thisIndex <= limit && *thisIndex; ++thisIndex)
     {
@@ -2132,24 +2118,16 @@ NS_IMETHODIMP
 nsStandardURL::GetFilePath(nsACString &result)
 {
     result = Filepath();
     return NS_OK;
 }
 
 // result may contain unescaped UTF-8 characters
 NS_IMETHODIMP
-nsStandardURL::GetParam(nsACString &result)
-{
-    result = Param();
-    return NS_OK;
-}
-
-// result may contain unescaped UTF-8 characters
-NS_IMETHODIMP
 nsStandardURL::GetQuery(nsACString &result)
 {
     result = Query();
     return NS_OK;
 }
 
 // result may contain unescaped UTF-8 characters
 NS_IMETHODIMP
@@ -2258,37 +2236,30 @@ nsStandardURL::SetFilePath(const nsACStr
             if (mSpec.Length() > end)
                 spec.Append(mSpec.get() + end, mSpec.Length() - end);
         }
 
         return SetSpec(spec);
     }
     else if (mPath.mLen > 1) {
         mSpec.Cut(mPath.mPos + 1, mFilepath.mLen - 1);
-        // left shift param, query, and ref
-        ShiftFromParam(1 - mFilepath.mLen);
+        // left shift query, and ref
+        ShiftFromQuery(1 - mFilepath.mLen);
         // these contain only a '/'
         mPath.mLen = 1;
         mDirectory.mLen = 1;
         mFilepath.mLen = 1;
         // these are no longer defined
         mBasename.mLen = -1;
         mExtension.mLen = -1;
     }
     return NS_OK;
 }
 
 NS_IMETHODIMP
-nsStandardURL::SetParam(const nsACString &input)
-{
-    NS_NOTYETIMPLEMENTED("");
-    return NS_ERROR_NOT_IMPLEMENTED;
-}
-
-NS_IMETHODIMP
 nsStandardURL::SetQuery(const nsACString &input)
 {
     ENSURE_MUTABLE();
 
     const nsPromiseFlatCString &flat = PromiseFlatCString(input);
     const char *query = flat.get();
 
     LOG(("nsStandardURL::SetQuery [query=%s]\n", query));
@@ -2498,17 +2469,17 @@ nsStandardURL::SetFileName(const nsACStr
             
             mBasename.mLen = basename.mLen;
             mExtension.mLen = extension.mLen;
             if (mExtension.mLen >= 0)
                 mExtension.mPos = mBasename.mPos + mBasename.mLen + 1;
         }
     }
     if (shift) {
-        ShiftFromParam(shift);
+        ShiftFromQuery(shift);
         mFilepath.mLen += shift;
         mPath.mLen += shift;
     }
     return NS_OK;
 }
 
 NS_IMETHODIMP
 nsStandardURL::SetFileBaseName(const nsACString &input)
@@ -2795,19 +2766,16 @@ nsStandardURL::Read(nsIObjectInputStream
     if (NS_FAILED(rv)) return rv;
 
     rv = ReadSegment(stream, mBasename);
     if (NS_FAILED(rv)) return rv;
 
     rv = ReadSegment(stream, mExtension);
     if (NS_FAILED(rv)) return rv;
 
-    rv = ReadSegment(stream, mParam);
-    if (NS_FAILED(rv)) return rv;
-
     rv = ReadSegment(stream, mQuery);
     if (NS_FAILED(rv)) return rv;
 
     rv = ReadSegment(stream, mRef);
     if (NS_FAILED(rv)) return rv;
 
     rv = NS_ReadOptionalCString(stream, mOriginCharset);
     if (NS_FAILED(rv)) return rv;
@@ -2884,19 +2852,16 @@ nsStandardURL::Write(nsIObjectOutputStre
     if (NS_FAILED(rv)) return rv;
 
     rv = WriteSegment(stream, mBasename);
     if (NS_FAILED(rv)) return rv;
 
     rv = WriteSegment(stream, mExtension);
     if (NS_FAILED(rv)) return rv;
 
-    rv = WriteSegment(stream, mParam);
-    if (NS_FAILED(rv)) return rv;
-
     rv = WriteSegment(stream, mQuery);
     if (NS_FAILED(rv)) return rv;
 
     rv = WriteSegment(stream, mRef);
     if (NS_FAILED(rv)) return rv;
 
     rv = NS_WriteOptionalStringZ(stream, mOriginCharset.get());
     if (NS_FAILED(rv)) return rv;
@@ -2959,17 +2924,16 @@ nsStandardURL::Read(const IPC::Message *
         !ReadSegment(aMsg, aIter, mUsername) ||
         !ReadSegment(aMsg, aIter, mPassword) ||
         !ReadSegment(aMsg, aIter, mHost) ||
         !ReadSegment(aMsg, aIter, mPath) ||
         !ReadSegment(aMsg, aIter, mFilepath) ||
         !ReadSegment(aMsg, aIter, mDirectory) ||
         !ReadSegment(aMsg, aIter, mBasename) ||
         !ReadSegment(aMsg, aIter, mExtension) ||
-        !ReadSegment(aMsg, aIter, mParam) ||
         !ReadSegment(aMsg, aIter, mQuery) ||
         !ReadSegment(aMsg, aIter, mRef) ||
         !ReadParam(aMsg, aIter, &mOriginCharset) ||
         !ReadParam(aMsg, aIter, &isMutable) ||
         !ReadParam(aMsg, aIter, &supportsFileURL) ||
         !ReadParam(aMsg, aIter, &hostEncoding))
         return PR_FALSE;
 
@@ -3000,17 +2964,16 @@ nsStandardURL::Write(IPC::Message *aMsg)
     WriteSegment(aMsg, mUsername);
     WriteSegment(aMsg, mPassword);
     WriteSegment(aMsg, mHost);
     WriteSegment(aMsg, mPath);
     WriteSegment(aMsg, mFilepath);
     WriteSegment(aMsg, mDirectory);
     WriteSegment(aMsg, mBasename);
     WriteSegment(aMsg, mExtension);
-    WriteSegment(aMsg, mParam);
     WriteSegment(aMsg, mQuery);
     WriteSegment(aMsg, mRef);
     WriteParam(aMsg, mOriginCharset);
     WriteParam(aMsg, bool(mMutable));
     WriteParam(aMsg, bool(mSupportsFileURL));
     WriteParam(aMsg, mHostEncoding);
     // mSpecEncoding and mHostA are just caches that can be recovered as needed.
 }
--- a/netwerk/base/src/nsStandardURL.h
+++ b/netwerk/base/src/nsStandardURL.h
@@ -218,31 +218,29 @@ private:
     const nsDependentCSubstring Hostport(); // see below
     const nsDependentCSubstring Host();     // see below
     const nsDependentCSubstring Path()      { return Segment(mPath); }
     const nsDependentCSubstring Filepath()  { return Segment(mFilepath); }
     const nsDependentCSubstring Directory() { return Segment(mDirectory); }
     const nsDependentCSubstring Filename(); // see below
     const nsDependentCSubstring Basename()  { return Segment(mBasename); }
     const nsDependentCSubstring Extension() { return Segment(mExtension); }
-    const nsDependentCSubstring Param()     { return Segment(mParam); }
     const nsDependentCSubstring Query()     { return Segment(mQuery); }
     const nsDependentCSubstring Ref()       { return Segment(mRef); }
 
     // shift the URLSegments to the right by diff
     void ShiftFromAuthority(PRInt32 diff) { mAuthority.mPos += diff; ShiftFromUsername(diff); }
     void ShiftFromUsername(PRInt32 diff)  { mUsername.mPos += diff; ShiftFromPassword(diff); }
     void ShiftFromPassword(PRInt32 diff)  { mPassword.mPos += diff; ShiftFromHost(diff); }
     void ShiftFromHost(PRInt32 diff)      { mHost.mPos += diff; ShiftFromPath(diff); }
     void ShiftFromPath(PRInt32 diff)      { mPath.mPos += diff; ShiftFromFilepath(diff); }
     void ShiftFromFilepath(PRInt32 diff)  { mFilepath.mPos += diff; ShiftFromDirectory(diff); }
     void ShiftFromDirectory(PRInt32 diff) { mDirectory.mPos += diff; ShiftFromBasename(diff); }
     void ShiftFromBasename(PRInt32 diff)  { mBasename.mPos += diff; ShiftFromExtension(diff); }
-    void ShiftFromExtension(PRInt32 diff) { mExtension.mPos += diff; ShiftFromParam(diff); }
-    void ShiftFromParam(PRInt32 diff)     { mParam.mPos += diff; ShiftFromQuery(diff); }
+    void ShiftFromExtension(PRInt32 diff) { mExtension.mPos += diff; ShiftFromQuery(diff); }
     void ShiftFromQuery(PRInt32 diff)     { mQuery.mPos += diff; ShiftFromRef(diff); }
     void ShiftFromRef(PRInt32 diff)       { mRef.mPos += diff; }
 
     // fastload helper functions
     nsresult ReadSegment(nsIBinaryInputStream *, URLSegment &);
     nsresult WriteSegment(nsIBinaryOutputStream *, const URLSegment &);
 
     // ipc helper functions
@@ -262,17 +260,16 @@ private:
     URLSegment mUsername;
     URLSegment mPassword;
     URLSegment mHost;
     URLSegment mPath;
     URLSegment mFilepath;
     URLSegment mDirectory;
     URLSegment mBasename;
     URLSegment mExtension;
-    URLSegment mParam;
     URLSegment mQuery;
     URLSegment mRef;
 
     nsCString              mOriginCharset;
     nsCOMPtr<nsIURLParser> mParser;
 
     // mFile is protected so subclasses can access it directly
 protected:
--- a/netwerk/base/src/nsURLHelper.cpp
+++ b/netwerk/base/src/nsURLHelper.cpp
@@ -216,17 +216,16 @@ net_ParseFileURL(const nsACString &inURL
                           nsnull, nsnull, // don't care about scheme
                           nsnull, nsnull, // don't care about authority
                           &pathPos, &pathLen);
     if (NS_FAILED(rv)) return rv;
 
     // invoke the parser to extract filepath from the path
     rv = parser->ParsePath(url + pathPos, pathLen,
                            &filepathPos, &filepathLen,
-                           nsnull, nsnull,  // don't care about param
                            nsnull, nsnull,  // don't care about query
                            nsnull, nsnull); // don't care about ref
     if (NS_FAILED(rv)) return rv;
 
     filepathPos += pathPos;
 
     // invoke the parser to extract the directory and filename from filepath
     rv = parser->ParseFilePath(url + filepathPos, filepathLen,
@@ -446,17 +445,16 @@ net_ResolveRelativePath(const nsACString
     PRBool stop = PR_FALSE;
     char c;
     for (; !stop; ++beg) {
         c = (beg == end) ? '\0' : *beg;
         //printf("%c [name=%s] [path=%s]\n", c, name.get(), path.get());
         switch (c) {
           case '\0':
           case '#':
-          case ';':
           case '?':
             stop = PR_TRUE;
             // fall through...
           case '/':
             // delimiter found
             if (name.EqualsLiteral("..")) {
                 // pop path
                 // If we already have the delim at end, then
--- a/netwerk/base/src/nsURLParsers.cpp
+++ b/netwerk/base/src/nsURLParsers.cpp
@@ -112,17 +112,16 @@ nsBaseURLParser::ParseURL(const char *sp
         switch (*p) {
             case ':':
                 if (!colon)
                     colon = p;
                 break;
             case '/': // start of filepath
             case '?': // start of query
             case '#': // start of ref
-            case ';': // start of param
                 if (!slash)
                     slash = p;
                 break;
             case '@': // username@hostname
             case '[': // start of IPv6 address literal
                 if (!stop)
                     stop = p;
                 break;
@@ -235,26 +234,25 @@ nsBaseURLParser::ParseServerInfo(const c
     if (port)
        *port = -1;
     return NS_OK;
 }
 
 NS_IMETHODIMP
 nsBaseURLParser::ParsePath(const char *path, PRInt32 pathLen,
                            PRUint32 *filepathPos, PRInt32 *filepathLen,
-                           PRUint32 *paramPos, PRInt32 *paramLen,
                            PRUint32 *queryPos, PRInt32 *queryLen,
                            PRUint32 *refPos, PRInt32 *refLen)
 {
     NS_PRECONDITION(path, "null pointer");
 
     if (pathLen < 0)
         pathLen = strlen(path);
 
-    // path = [/]<segment1>/<segment2>/<...>/<segmentN>;<param>?<query>#<ref>
+    // path = [/]<segment1>/<segment2>/<...>/<segmentN>?<query>#<ref>
 
     // XXX PL_strnpbrk would be nice, but it's buggy
 
     // search for first occurrence of either ? or #
     const char *query_beg = 0, *query_end = 0;
     const char *ref_beg = 0;
     const char *p = 0;
     for (p = path; p < path + pathLen; ++p) {
@@ -278,39 +276,23 @@ nsBaseURLParser::ParsePath(const char *p
     else
         SET_RESULT(query, 0, -1);
 
     if (ref_beg)
         SET_RESULT(ref, ref_beg - path, pathLen - (ref_beg - path));
     else
         SET_RESULT(ref, 0, -1);
 
-    // search backwards for param
-    const char *param_beg = 0;
     const char *end;
     if (query_beg)
         end = query_beg - 1;
     else if (ref_beg)
         end = ref_beg - 1;
     else
         end = path + pathLen;
-    for (p = end - 1; p >= path && *p != '/'; --p) {
-        if (*p == ';') {
-            // found param
-            param_beg = p + 1;
-        }
-    }
-
-    if (param_beg) {
-        // found <filepath>;<param>
-        SET_RESULT(param, param_beg - path, end - param_beg);
-        end = param_beg - 1;
-    }
-    else
-        SET_RESULT(param, 0, -1);
 
     // an empty file path is no file path
     if (end != path)
         SET_RESULT(filepath, 0, end - path);
     else
         SET_RESULT(filepath, 0, -1);
     return NS_OK;
 }
@@ -432,17 +414,17 @@ nsNoAuthURLParser::ParseAfterScheme(cons
                     nsCRT::IsAsciiAlpha(spec[2]) &&
                     ((specLen == 4) || (spec[4] == '/') || (spec[4] == '\\'))) {
                     pos = 1;
                     break;  
                 } 
 #endif
                 // Ignore apparent authority; path is everything after it
                 for (p = spec + 2; p < spec + specLen; ++p) {
-                    if (*p == '/' || *p == '?' || *p == '#' || *p == ';')
+                    if (*p == '/' || *p == '?' || *p == '#')
                         break;
                 }
             }
             SET_RESULT(auth, 0, -1);
             if (p && p != spec+specLen)
                 SET_RESULT(path, p - spec, specLen - (p - spec));
             else
                 SET_RESULT(path, 0, -1);
@@ -662,17 +644,17 @@ nsAuthURLParser::ParseAfterScheme(const 
     NS_PRECONDITION(specLen >= 0, "unexpected");
 
     PRUint32 nslash = CountConsecutiveSlashes(spec, specLen);
 
     // search for the end of the authority section
     const char *end = spec + specLen;
     const char *p;
     for (p = spec + nslash; p < end; ++p) {
-        if (*p == '/' || *p == '?' || *p == '#' || *p == ';')
+        if (*p == '/' || *p == '?' || *p == '#')
             break;
     }
     if (p < end) {
         // spec = [/]<auth><path>
         SET_RESULT(auth, nslash, p - (spec + nslash));
         SET_RESULT(path, p - spec, specLen - (p - spec));
     }
     else {
--- a/netwerk/test/TestURLParser.cpp
+++ b/netwerk/test/TestURLParser.cpp
@@ -65,28 +65,26 @@ parse_file_path(nsIURLParser *urlParser,
     PRINT_SUBFIELD(filepath, ext);
 }
 
 static void
 parse_path(nsIURLParser *urlParser, char *path, PRInt32 pathLen)
 {
     PRINT_FIELD(path);
 
-    PRUint32 filePos, paramPos, queryPos, refPos;
-    PRInt32 fileLen, paramLen, queryLen, refLen;
+    PRUint32 filePos, queryPos, refPos;
+    PRInt32 fileLen, queryLen, refLen;
 
     urlParser->ParsePath(path, pathLen,
                          &filePos, &fileLen,
-                         &paramPos, &paramLen,
                          &queryPos, &queryLen,
                          &refPos, &refLen);
 
     if (fileLen != -1)
         parse_file_path(urlParser, path + filePos, fileLen);
-    PRINT_SUBFIELD(path, param);
     PRINT_SUBFIELD(path, query);
     PRINT_SUBFIELD(path, ref);
 }
 
 int
 main(int argc, char **argv)
 {
     if (test_common_init(&argc, &argv) != 0)
--- a/netwerk/test/unit/test_bug429347.js
+++ b/netwerk/test/unit/test_bug429347.js
@@ -9,28 +9,34 @@ function run_test() {
   var uri1 = ios.newURI("http://example.com#bar", null, null);
   var uri2 = ios.newURI("http://example.com/#bar", null, null);
   do_check_true(uri1.equals(uri2));
 
   uri1.spec = "http://example.com?bar";
   uri2.spec = "http://example.com/?bar";
   do_check_true(uri1.equals(uri2));
 
+  // see https://bugzilla.mozilla.org/show_bug.cgi?id=665706
+  // ";" is not parsed as special anymore and thus ends up
+  // in the authority component (see RFC 3986)
   uri1.spec = "http://example.com;bar";
   uri2.spec = "http://example.com/;bar";
-  do_check_true(uri1.equals(uri2));
+  do_check_false(uri1.equals(uri2));
 
   uri1.spec = "http://example.com#";
   uri2.spec = "http://example.com/#";
   do_check_true(uri1.equals(uri2));
 
   uri1.spec = "http://example.com?";
   uri2.spec = "http://example.com/?";
   do_check_true(uri1.equals(uri2));
 
+  // see https://bugzilla.mozilla.org/show_bug.cgi?id=665706
+  // ";" is not parsed as special anymore and thus ends up
+  // in the authority component (see RFC 3986)
   uri1.spec = "http://example.com;";
   uri2.spec = "http://example.com/;";
-  do_check_true(uri1.equals(uri2));
+  do_check_false(uri1.equals(uri2));
 
   uri1.spec = "http://example.com";
   uri2.spec = "http://example.com/";
   do_check_true(uri1.equals(uri2));
 }
--- a/netwerk/test/urltest.cpp
+++ b/netwerk/test/urltest.cpp
@@ -132,18 +132,19 @@ nsresult writeoutto(const char* i_pURL, 
         output += RESULT();
         output += ',';
         rv = tURL->GetFileBaseName(temp);
         output += RESULT();
         output += ',';
         rv = tURL->GetFileExtension(temp);
         output += RESULT();
         output += ',';
-        rv = tURL->GetParam(temp);
-        output += RESULT();
+        // removed with https://bugzilla.mozilla.org/show_bug.cgi?id=665706
+        // rv = tURL->GetParam(temp); 
+        // output += RESULT();
         output += ',';
         rv = tURL->GetQuery(temp);
         output += RESULT();
         output += ',';
         rv = tURL->GetRef(temp);
         output += RESULT();
         output += ',';
         rv = tURL->GetSpec(temp);
--- a/xpcom/io/nsEscape.cpp
+++ b/xpcom/io/nsEscape.cpp
@@ -348,17 +348,17 @@ nsEscapeHTML2(const PRUnichar *aSourceBu
 //----------------------------------------------------------------------------------------
 
 const int EscapeChars[256] =
 /*      0    1    2    3    4    5    6    7    8    9    A    B    C    D    E    F */
 {
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,       /* 0x */
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,  	    /* 1x */
         0,1023,   0, 512,1023,   0,1023,   0,1023,1023,1023,1023,1023,1023, 953, 784,       /* 2x   !"#$%&'()*+,-./	 */
-     1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1008, 912,   0,1008,   0, 768,       /* 3x  0123456789:;<=>?	 */
+     1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1008,1008,   0,1008,   0, 768,       /* 3x  0123456789:;<=>?	 */
      1008,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,       /* 4x  @ABCDEFGHIJKLMNO  */
      1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023, 896, 896, 896, 896,1023,       /* 5x  PQRSTUVWXYZ[\]^_	 */
         0,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,       /* 6x  `abcdefghijklmno	 */
      1023,1023,1023,1023,1023,1023,1023,1023,1023,1023,1023, 896,1012, 896,1023,   0,       /* 7x  pqrstuvwxyz{|}~	 */
         0    /* 8x  DEL               */
 };
 
 #define NO_NEED_ESC(C) (EscapeChars[((unsigned int) (C))] & (flags))