closes bpo-38174: Update vendored expat library to 2.2.8. (GH-16346)

Fixes CVE-2019-15903. See full changelog at https://github.com/libexpat/libexpat/blob/R_2_2_8/expat/Changes.

1 files changed, 91 insertions, 121 deletions

diff --git a/Modules/expat/xmltok.h b/Modules/expat/xmltok.h
index 50926f3..2adbf53 100644
--- a/Modules/expat/xmltok.h
+++ b/Modules/expat/xmltok.h
@@ -38,16 +38,18 @@ extern "C" {
 #endif
 
 /* The following token may be returned by XmlContentTok */
-#define XML_TOK_TRAILING_RSQB -5 /* ] or ]] at the end of the scan; might be
-                                    start of illegal ]]> sequence */
+#define XML_TOK_TRAILING_RSQB                                                  \
+  -5 /* ] or ]] at the end of the scan; might be                               \
+        start of illegal ]]> sequence */
 /* The following tokens may be returned by both XmlPrologTok and
    XmlContentTok.
 */
-#define XML_TOK_NONE -4          /* The string to be scanned is empty */
-#define XML_TOK_TRAILING_CR -3   /* A CR at the end of the scan;
-                                    might be part of CRLF sequence */
-#define XML_TOK_PARTIAL_CHAR -2  /* only part of a multibyte sequence */
-#define XML_TOK_PARTIAL -1       /* only part of a token */
+#define XML_TOK_NONE -4 /* The string to be scanned is empty */
+#define XML_TOK_TRAILING_CR                                                    \
+  -3                            /* A CR at the end of the scan;                \
+                                   might be part of CRLF sequence */
+#define XML_TOK_PARTIAL_CHAR -2 /* only part of a multibyte sequence */
+#define XML_TOK_PARTIAL -1      /* only part of a token */
 #define XML_TOK_INVALID 0
 
 /* The following tokens are returned by XmlContentTok; some are also
@@ -62,24 +64,24 @@ extern "C" {
 #define XML_TOK_DATA_NEWLINE 7
 #define XML_TOK_CDATA_SECT_OPEN 8
 #define XML_TOK_ENTITY_REF 9
-#define XML_TOK_CHAR_REF 10               /* numeric character reference */
+#define XML_TOK_CHAR_REF 10 /* numeric character reference */
 
 /* The following tokens may be returned by both XmlPrologTok and
    XmlContentTok.
 */
-#define XML_TOK_PI 11                     /* processing instruction */
-#define XML_TOK_XML_DECL 12               /* XML decl or text decl */
+#define XML_TOK_PI 11       /* processing instruction */
+#define XML_TOK_XML_DECL 12 /* XML decl or text decl */
 #define XML_TOK_COMMENT 13
-#define XML_TOK_BOM 14                    /* Byte order mark */
+#define XML_TOK_BOM 14 /* Byte order mark */
 
 /* The following tokens are returned only by XmlPrologTok */
 #define XML_TOK_PROLOG_S 15
-#define XML_TOK_DECL_OPEN 16              /* <!foo */
-#define XML_TOK_DECL_CLOSE 17             /* > */
+#define XML_TOK_DECL_OPEN 16  /* <!foo */
+#define XML_TOK_DECL_CLOSE 17 /* > */
 #define XML_TOK_NAME 18
 #define XML_TOK_NMTOKEN 19
-#define XML_TOK_POUND_NAME 20             /* #name */
-#define XML_TOK_OR 21                     /* | */
+#define XML_TOK_POUND_NAME 20 /* #name */
+#define XML_TOK_OR 21         /* | */
 #define XML_TOK_PERCENT 22
 #define XML_TOK_OPEN_PAREN 23
 #define XML_TOK_CLOSE_PAREN 24
@@ -90,14 +92,14 @@ extern "C" {
 #define XML_TOK_INSTANCE_START 29
 
 /* The following occur only in element type declarations */
-#define XML_TOK_NAME_QUESTION 30          /* name? */
-#define XML_TOK_NAME_ASTERISK 31          /* name* */
-#define XML_TOK_NAME_PLUS 32              /* name+ */
-#define XML_TOK_COND_SECT_OPEN 33         /* <![ */
-#define XML_TOK_COND_SECT_CLOSE 34        /* ]]> */
-#define XML_TOK_CLOSE_PAREN_QUESTION 35   /* )? */
-#define XML_TOK_CLOSE_PAREN_ASTERISK 36   /* )* */
-#define XML_TOK_CLOSE_PAREN_PLUS 37       /* )+ */
+#define XML_TOK_NAME_QUESTION 30        /* name? */
+#define XML_TOK_NAME_ASTERISK 31        /* name* */
+#define XML_TOK_NAME_PLUS 32            /* name+ */
+#define XML_TOK_COND_SECT_OPEN 33       /* <![ */
+#define XML_TOK_COND_SECT_CLOSE 34      /* ]]> */
+#define XML_TOK_CLOSE_PAREN_QUESTION 35 /* )? */
+#define XML_TOK_CLOSE_PAREN_ASTERISK 36 /* )* */
+#define XML_TOK_CLOSE_PAREN_PLUS 37     /* )+ */
 #define XML_TOK_COMMA 38
 
 /* The following token is returned only by XmlAttributeValueTok */
@@ -112,20 +114,20 @@ extern "C" {
 #define XML_TOK_PREFIXED_NAME 41
 
 #ifdef XML_DTD
-#define XML_TOK_IGNORE_SECT 42
+#  define XML_TOK_IGNORE_SECT 42
 #endif /* XML_DTD */
 
 #ifdef XML_DTD
-#define XML_N_STATES 4
+#  define XML_N_STATES 4
 #else /* not XML_DTD */
-#define XML_N_STATES 3
+#  define XML_N_STATES 3
 #endif /* not XML_DTD */
 
 #define XML_PROLOG_STATE 0
 #define XML_CONTENT_STATE 1
 #define XML_CDATA_SECTION_STATE 2
 #ifdef XML_DTD
-#define XML_IGNORE_SECTION_STATE 3
+#  define XML_IGNORE_SECTION_STATE 3
 #endif /* XML_DTD */
 
 #define XML_N_LITERAL_TYPES 2
@@ -153,52 +155,41 @@ typedef struct {
 struct encoding;
 typedef struct encoding ENCODING;
 
-typedef int (PTRCALL *SCANNER)(const ENCODING *,
-                               const char *,
-                               const char *,
-                               const char **);
+typedef int(PTRCALL *SCANNER)(const ENCODING *, const char *, const char *,
+                              const char **);
 
 enum XML_Convert_Result {
   XML_CONVERT_COMPLETED = 0,
   XML_CONVERT_INPUT_INCOMPLETE = 1,
-  XML_CONVERT_OUTPUT_EXHAUSTED = 2  /* and therefore potentially input remaining as well */
+  XML_CONVERT_OUTPUT_EXHAUSTED
+  = 2 /* and therefore potentially input remaining as well */
 };
 
 struct encoding {
   SCANNER scanners[XML_N_STATES];
   SCANNER literalScanners[XML_N_LITERAL_TYPES];
-  int (PTRCALL *nameMatchesAscii)(const ENCODING *,
-                                  const char *,
-                                  const char *,
-                                  const char *);
-  int (PTRFASTCALL *nameLength)(const ENCODING *, const char *);
+  int(PTRCALL *nameMatchesAscii)(const ENCODING *, const char *, const char *,
+                                 const char *);
+  int(PTRFASTCALL *nameLength)(const ENCODING *, const char *);
   const char *(PTRFASTCALL *skipS)(const ENCODING *, const char *);
-  int (PTRCALL *getAtts)(const ENCODING *enc,
-                         const char *ptr,
-                         int attsMax,
-                         ATTRIBUTE *atts);
-  int (PTRFASTCALL *charRefNumber)(const ENCODING *enc, const char *ptr);
-  int (PTRCALL *predefinedEntityName)(const ENCODING *,
-                                      const char *,
-                                      const char *);
-  void (PTRCALL *updatePosition)(const ENCODING *,
-                                 const char *ptr,
-                                 const char *end,
-                                 POSITION *);
-  int (PTRCALL *isPublicId)(const ENCODING *enc,
-                            const char *ptr,
-                            const char *end,
-                            const char **badPtr);
-  enum XML_Convert_Result (PTRCALL *utf8Convert)(const ENCODING *enc,
-                              const char **fromP,
-                              const char *fromLim,
-                              char **toP,
-                              const char *toLim);
-  enum XML_Convert_Result (PTRCALL *utf16Convert)(const ENCODING *enc,
-                               const char **fromP,
-                               const char *fromLim,
-                               unsigned short **toP,
-                               const unsigned short *toLim);
+  int(PTRCALL *getAtts)(const ENCODING *enc, const char *ptr, int attsMax,
+                        ATTRIBUTE *atts);
+  int(PTRFASTCALL *charRefNumber)(const ENCODING *enc, const char *ptr);
+  int(PTRCALL *predefinedEntityName)(const ENCODING *, const char *,
+                                     const char *);
+  void(PTRCALL *updatePosition)(const ENCODING *, const char *ptr,
+                                const char *end, POSITION *);
+  int(PTRCALL *isPublicId)(const ENCODING *enc, const char *ptr,
+                           const char *end, const char **badPtr);
+  enum XML_Convert_Result(PTRCALL *utf8Convert)(const ENCODING *enc,
+                                                const char **fromP,
+                                                const char *fromLim, char **toP,
+                                                const char *toLim);
+  enum XML_Convert_Result(PTRCALL *utf16Convert)(const ENCODING *enc,
+                                                 const char **fromP,
+                                                 const char *fromLim,
+                                                 unsigned short **toP,
+                                                 const unsigned short *toLim);
   int minBytesPerChar;
   char isUtf8;
   char isUtf16;
@@ -225,66 +216,62 @@ struct encoding {
    the prolog outside literals, comments and processing instructions.
 */
 
-
-#define XmlTok(enc, state, ptr, end, nextTokPtr) \
+#define XmlTok(enc, state, ptr, end, nextTokPtr)                               \
   (((enc)->scanners[state])(enc, ptr, end, nextTokPtr))
 
-#define XmlPrologTok(enc, ptr, end, nextTokPtr) \
-   XmlTok(enc, XML_PROLOG_STATE, ptr, end, nextTokPtr)
+#define XmlPrologTok(enc, ptr, end, nextTokPtr)                                \
+  XmlTok(enc, XML_PROLOG_STATE, ptr, end, nextTokPtr)
 
-#define XmlContentTok(enc, ptr, end, nextTokPtr) \
-   XmlTok(enc, XML_CONTENT_STATE, ptr, end, nextTokPtr)
+#define XmlContentTok(enc, ptr, end, nextTokPtr)                               \
+  XmlTok(enc, XML_CONTENT_STATE, ptr, end, nextTokPtr)
 
-#define XmlCdataSectionTok(enc, ptr, end, nextTokPtr) \
-   XmlTok(enc, XML_CDATA_SECTION_STATE, ptr, end, nextTokPtr)
+#define XmlCdataSectionTok(enc, ptr, end, nextTokPtr)                          \
+  XmlTok(enc, XML_CDATA_SECTION_STATE, ptr, end, nextTokPtr)
 
 #ifdef XML_DTD
 
-#define XmlIgnoreSectionTok(enc, ptr, end, nextTokPtr) \
-   XmlTok(enc, XML_IGNORE_SECTION_STATE, ptr, end, nextTokPtr)
+#  define XmlIgnoreSectionTok(enc, ptr, end, nextTokPtr)                       \
+    XmlTok(enc, XML_IGNORE_SECTION_STATE, ptr, end, nextTokPtr)
 
 #endif /* XML_DTD */
 
 /* This is used for performing a 2nd-level tokenization on the content
    of a literal that has already been returned by XmlTok.
 */
-#define XmlLiteralTok(enc, literalType, ptr, end, nextTokPtr) \
+#define XmlLiteralTok(enc, literalType, ptr, end, nextTokPtr)                  \
   (((enc)->literalScanners[literalType])(enc, ptr, end, nextTokPtr))
 
-#define XmlAttributeValueTok(enc, ptr, end, nextTokPtr) \
-   XmlLiteralTok(enc, XML_ATTRIBUTE_VALUE_LITERAL, ptr, end, nextTokPtr)
+#define XmlAttributeValueTok(enc, ptr, end, nextTokPtr)                        \
+  XmlLiteralTok(enc, XML_ATTRIBUTE_VALUE_LITERAL, ptr, end, nextTokPtr)
 
-#define XmlEntityValueTok(enc, ptr, end, nextTokPtr) \
-   XmlLiteralTok(enc, XML_ENTITY_VALUE_LITERAL, ptr, end, nextTokPtr)
+#define XmlEntityValueTok(enc, ptr, end, nextTokPtr)                           \
+  XmlLiteralTok(enc, XML_ENTITY_VALUE_LITERAL, ptr, end, nextTokPtr)
 
-#define XmlNameMatchesAscii(enc, ptr1, end1, ptr2) \
+#define XmlNameMatchesAscii(enc, ptr1, end1, ptr2)                             \
   (((enc)->nameMatchesAscii)(enc, ptr1, end1, ptr2))
 
-#define XmlNameLength(enc, ptr) \
-  (((enc)->nameLength)(enc, ptr))
+#define XmlNameLength(enc, ptr) (((enc)->nameLength)(enc, ptr))
 
-#define XmlSkipS(enc, ptr) \
-  (((enc)->skipS)(enc, ptr))
+#define XmlSkipS(enc, ptr) (((enc)->skipS)(enc, ptr))
 
-#define XmlGetAttributes(enc, ptr, attsMax, atts) \
+#define XmlGetAttributes(enc, ptr, attsMax, atts)                              \
   (((enc)->getAtts)(enc, ptr, attsMax, atts))
 
-#define XmlCharRefNumber(enc, ptr) \
-  (((enc)->charRefNumber)(enc, ptr))
+#define XmlCharRefNumber(enc, ptr) (((enc)->charRefNumber)(enc, ptr))
 
-#define XmlPredefinedEntityName(enc, ptr, end) \
+#define XmlPredefinedEntityName(enc, ptr, end)                                 \
   (((enc)->predefinedEntityName)(enc, ptr, end))
 
-#define XmlUpdatePosition(enc, ptr, end, pos) \
+#define XmlUpdatePosition(enc, ptr, end, pos)                                  \
   (((enc)->updatePosition)(enc, ptr, end, pos))
 
-#define XmlIsPublicId(enc, ptr, end, badPtr) \
+#define XmlIsPublicId(enc, ptr, end, badPtr)                                   \
   (((enc)->isPublicId)(enc, ptr, end, badPtr))
 
-#define XmlUtf8Convert(enc, fromP, fromLim, toP, toLim) \
+#define XmlUtf8Convert(enc, fromP, fromLim, toP, toLim)                        \
   (((enc)->utf8Convert)(enc, fromP, fromLim, toP, toLim))
 
-#define XmlUtf16Convert(enc, fromP, fromLim, toP, toLim) \
+#define XmlUtf16Convert(enc, fromP, fromLim, toP, toLim)                       \
   (((enc)->utf16Convert)(enc, fromP, fromLim, toP, toLim))
 
 typedef struct {
@@ -292,16 +279,11 @@ typedef struct {
   const ENCODING **encPtr;
 } INIT_ENCODING;
 
-int XmlParseXmlDecl(int isGeneralTextEntity,
-                    const ENCODING *enc,
-                    const char *ptr,
-                    const char *end,
-                    const char **badPtr,
-                    const char **versionPtr,
-                    const char **versionEndPtr,
+int XmlParseXmlDecl(int isGeneralTextEntity, const ENCODING *enc,
+                    const char *ptr, const char *end, const char **badPtr,
+                    const char **versionPtr, const char **versionEndPtr,
                     const char **encodingNamePtr,
-                    const ENCODING **namedEncodingPtr,
-                    int *standalonePtr);
+                    const ENCODING **namedEncodingPtr, int *standalonePtr);
 
 int XmlInitEncoding(INIT_ENCODING *, const ENCODING **, const char *name);
 const ENCODING *XmlGetUtf8InternalEncoding(void);
@@ -310,34 +292,22 @@ int FASTCALL XmlUtf8Encode(int charNumber, char *buf);
 int FASTCALL XmlUtf16Encode(int charNumber, unsigned short *buf);
 int XmlSizeOfUnknownEncoding(void);
 
+typedef int(XMLCALL *CONVERTER)(void *userData, const char *p);
 
-typedef int (XMLCALL *CONVERTER) (void *userData, const char *p);
-
-ENCODING *
-XmlInitUnknownEncoding(void *mem,
-                       int *table,
-                       CONVERTER convert,
-                       void *userData);
+ENCODING *XmlInitUnknownEncoding(void *mem, int *table, CONVERTER convert,
+                                 void *userData);
 
-int XmlParseXmlDeclNS(int isGeneralTextEntity,
-                      const ENCODING *enc,
-                      const char *ptr,
-                      const char *end,
-                      const char **badPtr,
-                      const char **versionPtr,
-                      const char **versionEndPtr,
+int XmlParseXmlDeclNS(int isGeneralTextEntity, const ENCODING *enc,
+                      const char *ptr, const char *end, const char **badPtr,
+                      const char **versionPtr, const char **versionEndPtr,
                       const char **encodingNamePtr,
-                      const ENCODING **namedEncodingPtr,
-                      int *standalonePtr);
+                      const ENCODING **namedEncodingPtr, int *standalonePtr);
 
 int XmlInitEncodingNS(INIT_ENCODING *, const ENCODING **, const char *name);
 const ENCODING *XmlGetUtf8InternalEncodingNS(void);
 const ENCODING *XmlGetUtf16InternalEncodingNS(void);
-ENCODING *
-XmlInitUnknownEncodingNS(void *mem,
-                         int *table,
-                         CONVERTER convert,
-                         void *userData);
+ENCODING *XmlInitUnknownEncodingNS(void *mem, int *table, CONVERTER convert,
+                                   void *userData);
 #ifdef __cplusplus
 }
 #endif