File expat-CVE-2022-25236-relax-fix.patch of Package expat.23180
Index: expat-2.2.5/doc/reference.html
===================================================================
--- expat-2.2.5.orig/doc/reference.html
+++ expat-2.2.5/doc/reference.html
@@ -934,6 +934,14 @@ the local part will be concatenated with
to support RDF processors. It is a programming error to use the null separator
with <a href= "#XML_SetReturnNSTriplet">namespace triplets</a>.</div>
+<p><strong>Note:</strong>
+Expat does not validate namespace URIs (beyond encoding)
+against RFC 3986 today (and is not required to do so with regard to
+the XML 1.0 namespaces specification) but it may start doing that
+in future releases. Before that, an application using Expat must
+be ready to receive namespace URIs containing non-URI characters.
+</p>
+
<pre class="fcndec" id="XML_ParserCreate_MM">
XML_Parser XMLCALL
XML_ParserCreate_MM(const XML_Char *encoding,
Index: expat-2.2.5/lib/expat.h
===================================================================
--- expat-2.2.5.orig/lib/expat.h
+++ expat-2.2.5/lib/expat.h
@@ -244,6 +244,17 @@ XML_ParserCreate(const XML_Char *encodin
and the local part will be concatenated without any separator.
It is a programming error to use the separator '\0' with namespace
triplets (see XML_SetReturnNSTriplet).
+ If a namespace separator is chosen that can be part of a URI or
+ part of an XML name, splitting an expanded name back into its
+ 1, 2 or 3 original parts on application level in the element handler
+ may end up vulnerable, so these are advised against; sane choices for
+ a namespace separator are e.g. '\n' (line feed) and '|' (pipe).
+
+ Note that Expat does not validate namespace URIs (beyond encoding)
+ against RFC 3986 today (and is not required to do so with regard to
+ the XML 1.0 namespaces specification) but it may start doing that
+ in future releases. Before that, an application using Expat must
+ be ready to receive namespace URIs containing non-URI characters.
*/
XMLPARSEAPI(XML_Parser)
XML_ParserCreateNS(const XML_Char *encoding, XML_Char namespaceSeparator);
Index: expat-2.2.5/lib/xmlparse.c
===================================================================
--- expat-2.2.5.orig/lib/xmlparse.c
+++ expat-2.2.5/lib/xmlparse.c
@@ -3625,6 +3625,117 @@ storeAtts(XML_Parser parser, const ENCOD
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().
*/
@@ -3686,14 +3797,26 @@ addBinding(XML_Parser parser, PREFIX *pr
&& (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;
}
}
Index: expat-2.2.5/tests/runtests.c
===================================================================
--- expat-2.2.5.orig/tests/runtests.c
+++ expat-2.2.5/tests/runtests.c
@@ -8135,16 +8135,18 @@ START_TEST(test_ns_separator_in_uri) {
struct test_case {
enum XML_Status expectedStatus;
const char *doc;
+ XML_Char namesep;
};
struct test_case cases[] = {
- {XML_STATUS_OK, "<doc xmlns='one_two' />"},
- {XML_STATUS_ERROR, "<doc xmlns='one
two' />"},
+ {XML_STATUS_OK, "<doc xmlns='one_two' />", XCS('\n')},
+ {XML_STATUS_ERROR, "<doc xmlns='one
two' />", XCS('\n')},
+ {XML_STATUS_OK, "<doc xmlns='one:two' />", XCS(':')},
};
size_t i = 0;
size_t failCount = 0;
for (; i < sizeof(cases) / sizeof(cases[0]); i++) {
- XML_Parser parser = XML_ParserCreateNS(NULL, '\n');
+ XML_Parser parser = XML_ParserCreateNS(NULL, cases[i].namesep);
XML_SetElementHandler(parser, dummy_start_element, dummy_end_element);
if (XML_Parse(parser, cases[i].doc, (int)strlen(cases[i].doc),
/*isFinal*/ XML_TRUE)
