END_TEST
/* Regression test for SF bug #616863. */
static int XMLCALL
external_entity_handler(XML_Parser parser,
const XML_Char *context,
const XML_Char *base,
const XML_Char *systemId,
const XML_Char *publicId)
{
int callno = 1 + (int)XML_GetUserData(parser);
char *text;
XML_Parser p2;
if (callno == 1)
text = ("<!ELEMENT doc (e+)>\n"
"<!ATTLIST doc xmlns CDATA #IMPLIED>\n"
"<!ELEMENT e EMPTY>\n");
else
text = ("<?xml version='1.0' encoding='us-ascii'?>"
"<e/>");
XML_SetUserData(parser, (void *) callno);
p2 = XML_ExternalEntityParserCreate(parser, context, NULL);
if (XML_Parse(p2, text, strlen(text), XML_TRUE) == XML_STATUS_ERROR) {
xml_failure(p2);
return 0;
}
XML_ParserFree(p2);
return 1;
}
END_TEST
/* Regression test for SF bug #477667.
This test assures that any 8-bit character followed by a 7-bit
character will not be mistakenly interpreted as a valid UTF-8
sequence.
*/
START_TEST(test_illegal_utf8)
{
char text[100];
int i;
for (i = 128; i <= 255; ++i) {
sprintf(text, "<e>%ccd</e>", i);
if (XML_Parse(parser, text, strlen(text), XML_TRUE) == XML_STATUS_OK) {
sprintf(text,
"expected token error for '%c' (ordinal %d) in UTF-8 text",
i, i);
fail(text);
}
else if (XML_GetErrorCode(parser) != XML_ERROR_INVALID_TOKEN)
xml_failure(parser);
/* Reset the parser since we use the same parser repeatedly. */
XML_ParserReset(parser, NULL);
}
}
END_TEST
/* Regression test for SF bug #478332. */
START_TEST(test_really_long_lines)
{
/* This parses an input line longer than INIT_DATA_BUF_SIZE
characters long (defined to be 1024 in xmlparse.c). We take a
really cheesy approach to building the input buffer, because
this avoids writing bugs in buffer-filling code.
*/
char *text =
"<e>"
/* 64 chars */
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
/* until we have at least 1024 characters on the line: */
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-+"
"</e>";
if (XML_Parse(parser, text, strlen(text), XML_TRUE) == XML_STATUS_ERROR)
xml_failure(parser);
}
END_TEST
/* Regression test for SF bug #620106. */
static int XMLCALL
external_entity_loader_set_encoding(XML_Parser parser,
const XML_Char *context,
const XML_Char *base,
const XML_Char *systemId,
const XML_Char *publicId)
{
/* This text says it's an unsupported encoding, but it's really
UTF-8, which we tell Expat using XML_SetEncoding().
*/
char *text =
"<?xml encoding='iso-8859-3'?>"
"\xC3\xA9";
XML_Parser extparser;
extparser = XML_ExternalEntityParserCreate(parser, context, NULL);
if (extparser == NULL)
fail("Could not create external entity parser.");
if (!XML_SetEncoding(extparser, "utf-8"))
fail("XML_SetEncoding() ignored for external entity");
if ( XML_Parse(extparser, text, strlen(text), XML_TRUE)
== XML_STATUS_ERROR) {
xml_failure(parser);
return 0;
}
return 1;
}
END_TEST
START_TEST(test_bom_utf16_be)
{
char text[] = "\376\377\0<\0e\0/\0>";
if (XML_Parse(parser, text, sizeof(text) - 1, 1) == XML_STATUS_ERROR)
xml_failure(parser);
}
END_TEST
START_TEST(test_bom_utf16_le)
{
char text[] = "\377\376<\0e\0/\0>\0";
if (XML_Parse(parser, text, sizeof(text)-1, XML_TRUE) == XML_STATUS_ERROR)
xml_failure(parser);
}
END_TEST
START_TEST(test_bom_utf8)
{
/* This test is really just making sure we don't core on a UTF-8 BOM. */
char *text = "\357\273\277<e/>";
if (XML_Parse(parser, text, strlen(text), XML_TRUE) == XML_STATUS_ERROR)
xml_failure(parser);
}
END_TEST
START_TEST(test_ns_default_with_empty_uri)
{
char *text =
"<doc xmlns='http://xml.libexpat.org/'>\n"
" <e xmlns=''/>\n"
"</doc>";
if (XML_Parse(parser, text, strlen(text), XML_TRUE) == XML_STATUS_ERROR)
xml_failure(parser);
}
static void
run_ns_tagname_overwrite_test(char *text, char *result)
{
CharData storage;
CharData_Init(&storage);
XML_SetUserData(parser, &storage);
XML_SetElementHandler(parser,
overwrite_start_checker, overwrite_end_checker);
if (XML_Parse(parser, text, strlen(text), XML_TRUE) == XML_STATUS_ERROR)
xml_failure(parser);
CharData_CheckString(&storage, result);
}
END_TEST
/* Regression test for SF bug #1515266: missing check of stopped
parser in doContext() 'for' loop. */
START_TEST(test_suspend_parser_between_char_data_calls)
{
/* The sample data must be big enough that there are two calls to
the character data handler from within the inner "for" loop of
the XML_TOK_DATA_CHARS case in doContent(), and the character
handler must stop the parser and clear the character data
handler.
*/
char *text = long_character_data_text;
XML_SetCharacterDataHandler(parser, clearing_aborting_character_handler);
resumable = XML_TRUE;
if (XML_Parse(parser, text, strlen(text), XML_TRUE) != XML_STATUS_SUSPENDED)
xml_failure(parser);
if (XML_GetErrorCode(parser) != XML_ERROR_NONE)
xml_failure(parser);
}
static void
run_attribute_check(XML_Char *text, XML_Char *expected)
{
CharData storage;
CharData_Init(&storage);
XML_SetUserData(parser, &storage);
XML_SetStartElementHandler(parser, accumulate_attribute);
if (XML_Parse(parser, text, strlen(text), 1) == XML_STATUS_ERROR)
xml_failure(parser);
CharData_CheckXMLChars(&storage, expected);
}
END_TEST
/* Test that no error is reported for unknown entities if we don't
read an external subset. This was fixed in Expat 1.95.5.
*/
START_TEST(test_wfc_undeclared_entity_unread_external_subset) {
char *text =
"<!DOCTYPE doc SYSTEM 'foo'>\n"
"<doc>&entity;</doc>";
if (XML_Parse(parser, text, strlen(text), XML_TRUE) == XML_STATUS_ERROR)
xml_failure(parser);
}
END_TEST
START_TEST(test_utf16_le_epilog_newline)
{
int first_chunk_bytes = 17;
char text[] =
"\xFF\xFE" /* BOM */
"<\000e\000/\000>\000" /* document element */
"\r\000\n\000\r\000\n\000"; /* epilog */
if (first_chunk_bytes >= sizeof(text) - 1)
fail("bad value of first_chunk_bytes");
if (XML_Parse(parser, text, first_chunk_bytes, 0) == XML_STATUS_ERROR)
xml_failure(parser);
else {
enum XML_Status rc;
rc = XML_Parse(parser, text + first_chunk_bytes,
sizeof(text) - first_chunk_bytes - 1, 1);
if (rc == XML_STATUS_ERROR)
xml_failure(parser);
}
}
END_TEST
/* Regression test #2 for SF bug #653180. */
START_TEST(test_column_number_after_parse)
{
char *text = "<tag></tag>";
int colno;
if (XML_Parse(parser, text, strlen(text), XML_FALSE) == XML_STATUS_ERROR)
xml_failure(parser);
colno = XML_GetCurrentColumnNumber(parser);
if (colno != 11) {
char buffer[100];
sprintf(buffer, "expected 11 columns, saw %d", colno);
fail(buffer);
}
}
END_TEST
/* Regression test for SF bug #824420.
Checks that an xmlns:prefix attribute set in an attribute's default
value isn't misinterpreted.
*/
START_TEST(test_ns_in_attribute_default_without_namespaces)
{
char *text =
"<!DOCTYPE e:element [\n"
" <!ATTLIST e:element\n"
" xmlns:e CDATA 'http://example.com/'>\n"
" ]>\n"
"<e:element/>";
if (XML_Parse(parser, text, strlen(text), XML_TRUE) == XML_STATUS_ERROR)
xml_failure(parser);
}
END_TEST
/* See related SF bug #673791.
When namespace processing is enabled, setting the namespace URI for
a prefix is not allowed; this test ensures that it *is* allowed
when namespace processing is not enabled.
(See Namespaces in XML, section 2.)
*/
START_TEST(test_empty_ns_without_namespaces)
{
char *text =
"<doc xmlns:prefix='http://www.example.com/'>\n"
" <e xmlns:prefix=''/>\n"
"</doc>";
if (XML_Parse(parser, text, strlen(text), XML_TRUE) == XML_STATUS_ERROR)
xml_failure(parser);
}
END_TEST
START_TEST(test_utf16)
{
/* <?xml version="1.0" encoding="UTF-16"?>
<doc a='123'>some text</doc>
*/
char text[] =
"\000<\000?\000x\000m\000\154\000 \000v\000e\000r\000s\000i\000o"
"\000n\000=\000'\0001\000.\000\060\000'\000 \000e\000n\000c\000o"
"\000d\000i\000n\000g\000=\000'\000U\000T\000F\000-\0001\000\066"
"\000'\000?\000>\000\n"
"\000<\000d\000o\000c\000 \000a\000=\000'\0001\0002\0003\000'"
"\000>\000s\000o\000m\000e\000 \000t\000e\000x\000t\000<\000/"
"\000d\000o\000c\000>";
if (XML_Parse(parser, text, sizeof(text)-1, XML_TRUE) == XML_STATUS_ERROR)
xml_failure(parser);
}
END_TEST
/* Test that no error is reported for unknown entities if we have read
an external subset, and standalone is false.
*/
START_TEST(test_wfc_undeclared_entity_with_external_subset) {
char *text =
"<?xml version='1.0' encoding='us-ascii'?>\n"
"<!DOCTYPE doc SYSTEM 'foo'>\n"
"<doc>&entity;</doc>";
char *foo_text =
"<!ELEMENT doc (#PCDATA)*>";
XML_SetParamEntityParsing(parser, XML_PARAM_ENTITY_PARSING_ALWAYS);
XML_SetUserData(parser, foo_text);
XML_SetExternalEntityRefHandler(parser, external_entity_loader);
if (XML_Parse(parser, text, strlen(text), XML_TRUE) == XML_STATUS_ERROR)
xml_failure(parser);
}
END_TEST
/* Regression test #1 for SF bug #653180. */
START_TEST(test_line_number_after_parse)
{
char *text =
"<tag>\n"
"\n"
"\n</tag>";
int lineno;
if (XML_Parse(parser, text, strlen(text), XML_FALSE) == XML_STATUS_ERROR)
xml_failure(parser);
lineno = XML_GetCurrentLineNumber(parser);
if (lineno != 4) {
char buffer[100];
sprintf(buffer, "expected 4 lines, saw %d", lineno);
fail(buffer);
}
}
END_TEST
static int
external_entity_loader(XML_Parser parser,
const XML_Char *context,
const XML_Char *base,
const XML_Char *systemId,
const XML_Char *publicId)
{
char *text = (char *)XML_GetUserData(parser);
XML_Parser extparser;
extparser = XML_ExternalEntityParserCreate(parser, context, NULL);
if (extparser == NULL)
fail("Could not create external entity parser.");
if (XML_Parse(extparser, text, strlen(text), 1) == XML_STATUS_ERROR) {
xml_failure(parser);
return 0;
}
return 1;
}
END_TEST
/* Regression test for SF bug #422239 (maybe).
It's not clear that this reproduces enough of the context
of the reported bug.
*/
START_TEST(test_line_count)
{
char *text =
"<e>\n"
" <e/>\n"
"</e>";
int lineno;
if (XML_Parse(parser, text, strlen(text), 1) == XML_STATUS_ERROR)
xml_failure(parser);
lineno = XML_GetCurrentLineNumber(parser);
if (lineno != 3) {
char buffer[100];
sprintf(buffer, "expected 3 lines, saw %d", lineno);
fail(buffer);
}
}
END_TEST
/* Regression test #4 for SF bug #673791. */
START_TEST(test_ns_prefix_with_empty_uri_4)
{
char *text =
"<!DOCTYPE doc [\n"
" <!ELEMENT prefix:doc EMPTY>\n"
" <!ATTLIST prefix:doc\n"
" xmlns:prefix CDATA 'http://xml.libexpat.org/'>\n"
"]>\n"
"<prefix:doc/>";
/* Packaged info expected by the end element handler;
the weird structuring lets us re-use the triplet_end_checker()
function also used for another test. */
char *elemstr[] = {
"http://xml.libexpat.org/ doc prefix"
};
XML_SetReturnNSTriplet(parser, XML_TRUE);
XML_SetUserData(parser, elemstr);
XML_SetEndElementHandler(parser, triplet_end_checker);
if (XML_Parse(parser, text, strlen(text), XML_TRUE) == XML_STATUS_ERROR)
xml_failure(parser);
}
