From 2ba6c76fca21397959145e18c5ef376201209020 Mon Sep 17 00:00:00 2001
From: Sebastian Pipping <sebastian@pipping.org>
Date: Sun, 27 Feb 2022 16:58:08 +0100
Subject: [PATCH] lib: Relax fix to CVE-2022-25236 with regard to RFC
 3986 URI characters

---
 lib/xmlparse.c | 139 ++++++++++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 131 insertions(+), 8 deletions(-)

diff --git a/lib/xmlparse.c b/lib/xmlparse.c
index 59da19c..6fe2cf1 100644
--- a/lib/xmlparse.c
+++ b/lib/xmlparse.c
@@ -3705,6 +3705,117 @@ storeAtts(XML_Parser parser, const ENCODING *enc, const char *attStr,
   return XML_ERROR_NONE;
 }
 
+static XML_Bool
+is_rfc3986_uri_char(XML_Char candidate) {
+  // For the RFC 3986 ANBF grammar see
+  // https://datatracker.ietf.org/doc/html/rfc3986#appendix-A
+
+  switch (candidate) {
+  // From rule "ALPHA" (uppercase half)
+  case 'A':
+  case 'B':
+  case 'C':
+  case 'D':
+  case 'E':
+  case 'F':
+  case 'G':
+  case 'H':
+  case 'I':
+  case 'J':
+  case 'K':
+  case 'L':
+  case 'M':
+  case 'N':
+  case 'O':
+  case 'P':
+  case 'Q':
+  case 'R':
+  case 'S':
+  case 'T':
+  case 'U':
+  case 'V':
+  case 'W':
+  case 'X':
+  case 'Y':
+  case 'Z':
+
+  // From rule "ALPHA" (lowercase half)
+  case 'a':
+  case 'b':
+  case 'c':
+  case 'd':
+  case 'e':
+  case 'f':
+  case 'g':
+  case 'h':
+  case 'i':
+  case 'j':
+  case 'k':
+  case 'l':
+  case 'm':
+  case 'n':
+  case 'o':
+  case 'p':
+  case 'q':
+  case 'r':
+  case 's':
+  case 't':
+  case 'u':
+  case 'v':
+  case 'w':
+  case 'x':
+  case 'y':
+  case 'z':
+
+  // From rule "DIGIT"
+  case '0':
+  case '1':
+  case '2':
+  case '3':
+  case '4':
+  case '5':
+  case '6':
+  case '7':
+  case '8':
+  case '9':
+
+  // From rule "pct-encoded"
+  case '%':
+
+  // From rule "unreserved"
+  case '-':
+  case '.':
+  case '_':
+  case '~':
+
+  // From rule "gen-delims"
+  case ':':
+  case '/':
+  case '?':
+  case '#':
+  case '[':
+  case ']':
+  case '@':
+
+  // From rule "sub-delims"
+  case '!':
+  case '$':
+  case '&':
+  case '\'':
+  case '(':
+  case ')':
+  case '*':
+  case '+':
+  case ',':
+  case ';':
+  case '=':
+    return XML_TRUE;
+
+  default:
+    return XML_FALSE;
+  }
+}
+
 /* addBinding() overwrites the value of prefix->binding without checking.
    Therefore one must keep track of the old value outside of addBinding().
 */
@@ -3763,14 +3874,26 @@ addBinding(XML_Parser parser, PREFIX *prefix, const ATTRIBUTE_ID *attId,
         && (len > xmlnsLen || uri[len] != xmlnsNamespace[len]))
       isXMLNS = XML_FALSE;
 
-    // NOTE: While Expat does not validate namespace URIs against RFC 3986,
-    //       we have to at least make sure that the XML processor on top of
-    //       Expat (that is splitting tag names by namespace separator into
-    //       2- or 3-tuples (uri-local or uri-local-prefix)) cannot be confused
-    //       by an attacker putting additional namespace separator characters
-    //       into namespace declarations.  That would be ambiguous and not to
-    //       be expected.
-    if (parser->m_ns && (uri[len] == parser->m_namespaceSeparator)) {
+    // NOTE: While Expat does not validate namespace URIs against RFC 3986
+    //       today (and is not REQUIRED to do so with regard to the XML 1.0
+    //       namespaces specification) we have to at least make sure, that
+    //       the application on top of Expat (that is likely splitting expanded
+    //       element names ("qualified names") of form
+    //       "[uri sep] local [sep prefix] '\0'" back into 1, 2 or 3 pieces
+    //       in its element handler code) cannot be confused by an attacker
+    //       putting additional namespace separator characters into namespace
+    //       declarations.  That would be ambiguous and not to be expected.
+    //
+    //       While the HTML API docs of function XML_ParserCreateNS have been
+    //       advising against use of a namespace separator character that can
+    //       appear in a URI for >20 years now, some widespread applications
+    //       are using URI characters (':' (colon) in particular) for a
+    //       namespace separator, in practice.  To keep these applications
+    //       functional, we only reject namespaces URIs containing the
+    //       application-chosen namespace separator if the chosen separator
+    //       is a non-URI character with regard to RFC 3986.
+    if (parser->m_ns && (uri[len] == parser->m_namespaceSeparator)
+        && ! is_rfc3986_uri_char(uri[len])) {
       return XML_ERROR_SYNTAX;
     }
   }
-- 
1.8.3.1