Update libxml to 2.9.13

ref:f572491d236694e847142c36f0f5546c649e05d7

1 files changed, 95 insertions, 37 deletions

diff --git a/contrib/libs/libxml/encoding.c b/contrib/libs/libxml/encoding.c
index 89ac8dd2c9..945d2c4470 100644
--- a/contrib/libs/libxml/encoding.c
+++ b/contrib/libs/libxml/encoding.c
@@ -80,7 +80,7 @@ static int xmlLittleEndian = 1;
 
 /**
  * xmlEncodingErrMemory:
- * @extra:  extra informations
+ * @extra:  extra information
  *
  * Handle an out of memory condition
  */
@@ -174,7 +174,7 @@ closeIcuConverter(uconv_t *conv)
  * Returns 0 if success, or -1 otherwise
  * The value of @inlen after return is the number of octets consumed
  *     if the return value is positive, else unpredictable.
- * The value of @outlen after return is the number of octets consumed.
+ * The value of @outlen after return is the number of octets produced.
  */
 static int
 asciiToUTF8(unsigned char* out, int *outlen,
@@ -221,7 +221,7 @@ asciiToUTF8(unsigned char* out, int *outlen,
  * Returns 0 if success, -2 if the transcoding fails, or -1 otherwise
  * The value of @inlen after return is the number of octets consumed
  *     if the return value is positive, else unpredictable.
- * The value of @outlen after return is the number of octets consumed.
+ * The value of @outlen after return is the number of octets produced.
  */
 static int
 UTF8Toascii(unsigned char* out, int *outlen,
@@ -305,7 +305,7 @@ UTF8Toascii(unsigned char* out, int *outlen,
  * Returns the number of bytes written if success, or -1 otherwise
  * The value of @inlen after return is the number of octets consumed
  *     if the return value is positive, else unpredictable.
- * The value of @outlen after return is the number of octets consumed.
+ * The value of @outlen after return is the number of octets produced.
  */
 int
 isolat1ToUTF8(unsigned char* out, int *outlen,
@@ -377,6 +377,11 @@ UTF8ToUTF8(unsigned char* out, int *outlen,
     if (len < 0)
 	return(-1);
 
+    /*
+     * FIXME: Conversion functions must assure valid UTF-8, so we have
+     * to check for UTF-8 validity. Preferably, this converter shouldn't
+     * be used at all.
+     */
     memcpy(out, inb, len);
 
     *outlen = len;
@@ -400,7 +405,7 @@ UTF8ToUTF8(unsigned char* out, int *outlen,
            or -1 otherwise
  * The value of @inlen after return is the number of octets consumed
  *     if the return value is positive, else unpredictable.
- * The value of @outlen after return is the number of octets consumed.
+ * The value of @outlen after return is the number of octets produced.
  */
 int
 UTF8Toisolat1(unsigned char* out, int *outlen,
@@ -500,13 +505,18 @@ UTF16LEToUTF8(unsigned char* out, int *outlen,
 {
     unsigned char* outstart = out;
     const unsigned char* processed = inb;
-    unsigned char* outend = out + *outlen;
+    unsigned char* outend;
     unsigned short* in = (unsigned short*) inb;
     unsigned short* inend;
     unsigned int c, d, inlen;
     unsigned char *tmp;
     int bits;
 
+    if (*outlen == 0) {
+        *inlenb = 0;
+        return(0);
+    }
+    outend = out + *outlen;
     if ((*inlenb % 2) == 1)
         (*inlenb)--;
     inlen = *inlenb / 2;
@@ -521,7 +531,7 @@ UTF16LEToUTF8(unsigned char* out, int *outlen,
 	    in++;
 	}
         if ((c & 0xFC00) == 0xD800) {    /* surrogates */
-	    if (in >= inend) {           /* (in > inend) shouldn't happens */
+	    if (in >= inend) {           /* handle split mutli-byte characters */
 		break;
 	    }
 	    if (xmlLittleEndian) {
@@ -738,38 +748,39 @@ UTF16BEToUTF8(unsigned char* out, int *outlen,
 {
     unsigned char* outstart = out;
     const unsigned char* processed = inb;
-    unsigned char* outend = out + *outlen;
+    unsigned char* outend;
     unsigned short* in = (unsigned short*) inb;
     unsigned short* inend;
     unsigned int c, d, inlen;
     unsigned char *tmp;
     int bits;
 
+    if (*outlen == 0) {
+        *inlenb = 0;
+        return(0);
+    }
+    outend = out + *outlen;
     if ((*inlenb % 2) == 1)
         (*inlenb)--;
     inlen = *inlenb / 2;
     inend= in + inlen;
-    while (in < inend) {
+    while ((in < inend) && (out - outstart + 5 < *outlen)) {
 	if (xmlLittleEndian) {
 	    tmp = (unsigned char *) in;
 	    c = *tmp++;
-	    c = c << 8;
-	    c = c | (unsigned int) *tmp;
+	    c = (c << 8) | (unsigned int) *tmp;
 	    in++;
 	} else {
 	    c= *in++;
 	}
         if ((c & 0xFC00) == 0xD800) {    /* surrogates */
-	    if (in >= inend) {           /* (in > inend) shouldn't happens */
-		*outlen = out - outstart;
-		*inlenb = processed - inb;
-	        return(-2);
+	    if (in >= inend) {           /* handle split mutli-byte characters */
+                break;
 	    }
 	    if (xmlLittleEndian) {
 		tmp = (unsigned char *) in;
 		d = *tmp++;
-		d = d << 8;
-		d = d | (unsigned int) *tmp;
+		d = (d << 8) | (unsigned int) *tmp;
 		in++;
 	    } else {
 		d= *in++;
@@ -1493,16 +1504,25 @@ xmlRegisterCharEncodingHandler(xmlCharEncodingHandlerPtr handler) {
     if ((handler == NULL) || (handlers == NULL)) {
         xmlEncodingErr(XML_I18N_NO_HANDLER,
 		"xmlRegisterCharEncodingHandler: NULL handler !\n", NULL);
-	return;
+        goto free_handler;
     }
 
     if (nbCharEncodingHandler >= MAX_ENCODING_HANDLERS) {
         xmlEncodingErr(XML_I18N_EXCESS_HANDLER,
 	"xmlRegisterCharEncodingHandler: Too many handler registered, see %s\n",
 	               "MAX_ENCODING_HANDLERS");
-	return;
+        goto free_handler;
     }
     handlers[nbCharEncodingHandler++] = handler;
+    return;
+
+free_handler:
+    if (handler != NULL) {
+        if (handler->name != NULL) {
+            xmlFree(handler->name);
+        }
+        xmlFree(handler);
+    }
 }
 
 /**
@@ -1794,7 +1814,7 @@ xmlFindCharEncodingHandler(const char *name) {
  * @cd:		iconv converter data structure
  * @out:  a pointer to an array of bytes to store the result
  * @outlen:  the length of @out
- * @in:  a pointer to an array of ISO Latin 1 chars
+ * @in:  a pointer to an array of input bytes
  * @inlen:  the length of @in
  *
  * Returns 0 if success, or
@@ -1805,7 +1825,7 @@ xmlFindCharEncodingHandler(const char *name) {
  *
  * The value of @inlen after return is the number of octets consumed
  *     as the return value is positive, else unpredictable.
- * The value of @outlen after return is the number of octets consumed.
+ * The value of @outlen after return is the number of octets produced.
  */
 static int
 xmlIconvWrapper(iconv_t cd, unsigned char *out, int *outlen,
@@ -1813,7 +1833,7 @@ xmlIconvWrapper(iconv_t cd, unsigned char *out, int *outlen,
     size_t icv_inlen, icv_outlen;
     const char *icv_in = (const char *) in;
     char *icv_out = (char *) out;
-    int ret;
+    size_t ret;
 
     if ((out == NULL) || (outlen == NULL) || (inlen == NULL) || (in == NULL)) {
         if (outlen != NULL) *outlen = 0;
@@ -1824,7 +1844,7 @@ xmlIconvWrapper(iconv_t cd, unsigned char *out, int *outlen,
     ret = iconv(cd, (ICONV_CONST char **) &icv_in, &icv_inlen, &icv_out, &icv_outlen);
     *inlen -= icv_inlen;
     *outlen -= icv_outlen;
-    if ((icv_inlen != 0) || (ret == -1)) {
+    if ((icv_inlen != 0) || (ret == (size_t) -1)) {
 #ifdef EILSEQ
         if (errno == EILSEQ) {
             return -2;
@@ -1861,7 +1881,7 @@ xmlIconvWrapper(iconv_t cd, unsigned char *out, int *outlen,
  * @toUnicode : non-zero if toUnicode. 0 otherwise.
  * @out:  a pointer to an array of bytes to store the result
  * @outlen:  the length of @out
- * @in:  a pointer to an array of ISO Latin 1 chars
+ * @in:  a pointer to an array of input bytes
  * @inlen:  the length of @in
  * @flush: if true, indicates end of input
  *
@@ -1873,7 +1893,7 @@ xmlIconvWrapper(iconv_t cd, unsigned char *out, int *outlen,
  *
  * The value of @inlen after return is the number of octets consumed
  *     as the return value is positive, else unpredictable.
- * The value of @outlen after return is the number of octets consumed.
+ * The value of @outlen after return is the number of octets produced.
  */
 static int
 xmlUconvWrapper(uconv_t *cd, int toUnicode, unsigned char *out, int *outlen,
@@ -1922,6 +1942,25 @@ xmlUconvWrapper(uconv_t *cd, int toUnicode, unsigned char *out, int *outlen,
  *									*
  ************************************************************************/
 
+/**
+ * xmlEncInputChunk:
+ * @handler:  encoding handler
+ * @out:  a pointer to an array of bytes to store the result
+ * @outlen:  the length of @out
+ * @in:  a pointer to an array of input bytes
+ * @inlen:  the length of @in
+ * @flush:  flush (ICU-related)
+ *
+ * Returns 0 if success, or
+ *     -1 by lack of space, or
+ *     -2 if the transcoding fails (for *in is not valid utf8 string or
+ *        the result of transformation can't fit into the encoding we want), or
+ *     -3 if there the last byte can't form a single output char.
+ *
+ * The value of @inlen after return is the number of octets consumed
+ *     as the return value is 0, else unpredictable.
+ * The value of @outlen after return is the number of octets produced.
+ */
 static int
 xmlEncInputChunk(xmlCharEncodingHandler *handler, unsigned char *out,
                  int *outlen, const unsigned char *in, int *inlen, int flush) {
@@ -1930,6 +1969,8 @@ xmlEncInputChunk(xmlCharEncodingHandler *handler, unsigned char *out,
 
     if (handler->input != NULL) {
         ret = handler->input(out, outlen, in, inlen);
+        if (ret > 0)
+           ret = 0;
     }
 #ifdef LIBXML_ICONV_ENABLED
     else if (handler->iconv_in != NULL) {
@@ -1951,7 +1992,25 @@ xmlEncInputChunk(xmlCharEncodingHandler *handler, unsigned char *out,
     return(ret);
 }
 
-/* Returns -4 if no output function was found. */
+/**
+ * xmlEncOutputChunk:
+ * @handler:  encoding handler
+ * @out:  a pointer to an array of bytes to store the result
+ * @outlen:  the length of @out
+ * @in:  a pointer to an array of input bytes
+ * @inlen:  the length of @in
+ *
+ * Returns 0 if success, or
+ *     -1 by lack of space, or
+ *     -2 if the transcoding fails (for *in is not valid utf8 string or
+ *        the result of transformation can't fit into the encoding we want), or
+ *     -3 if there the last byte can't form a single output char.
+ *     -4 if no output function was found.
+ *
+ * The value of @inlen after return is the number of octets consumed
+ *     as the return value is 0, else unpredictable.
+ * The value of @outlen after return is the number of octets produced.
+ */
 static int
 xmlEncOutputChunk(xmlCharEncodingHandler *handler, unsigned char *out,
                   int *outlen, const unsigned char *in, int *inlen) {
@@ -1959,6 +2018,8 @@ xmlEncOutputChunk(xmlCharEncodingHandler *handler, unsigned char *out,
 
     if (handler->output != NULL) {
         ret = handler->output(out, outlen, in, inlen);
+        if (ret > 0)
+           ret = 0;
     }
 #ifdef LIBXML_ICONV_ENABLED
     else if (handler->iconv_out != NULL) {
@@ -1968,7 +2029,7 @@ xmlEncOutputChunk(xmlCharEncodingHandler *handler, unsigned char *out,
 #ifdef LIBXML_ICU_ENABLED
     else if (handler->uconv_out != NULL) {
         ret = xmlUconvWrapper(handler->uconv_out, 0, out, outlen, in, inlen,
-                              TRUE);
+                              1);
     }
 #endif /* LIBXML_ICU_ENABLED */
     else {
@@ -2064,7 +2125,7 @@ xmlCharEncFirstLineInt(xmlCharEncodingHandler *handler, xmlBufferPtr out,
      */
     if (ret == -3) ret = 0;
     if (ret == -1) ret = 0;
-    return(ret);
+    return(written ? written : ret);
 }
 
 /**
@@ -2194,7 +2255,7 @@ xmlCharEncFirstLineInput(xmlParserInputBufferPtr input, int len)
      */
     if (ret == -3) ret = 0;
     if (ret == -1) ret = 0;
-    return(ret);
+    return(c_out ? c_out : ret);
 }
 
 /**
@@ -2404,7 +2465,7 @@ xmlCharEncOutput(xmlOutputBufferPtr output, int init)
 {
     int ret;
     size_t written;
-    size_t writtentot = 0;
+    int writtentot = 0;
     size_t toconv;
     int c_in;
     int c_out;
@@ -2437,7 +2498,7 @@ retry:
 	xmlGenericError(xmlGenericErrorContext,
 		"initialized encoder\n");
 #endif
-        return(0);
+        return(c_out);
     }
 
     /*
@@ -2445,7 +2506,7 @@ retry:
      */
     toconv = xmlBufUse(in);
     if (toconv == 0)
-        return (0);
+        return (writtentot);
     if (toconv > 64 * 1024)
         toconv = 64 * 1024;
     if (toconv * 4 >= written) {
@@ -2550,7 +2611,7 @@ retry:
             goto retry;
 	}
     }
-    return(ret);
+    return(writtentot ? writtentot : ret);
 }
 #endif
 
@@ -2579,7 +2640,6 @@ xmlCharEncOutFunc(xmlCharEncodingHandler *handler, xmlBufferPtr out,
     int written;
     int writtentot = 0;
     int toconv;
-    int output = 0;
 
     if (handler == NULL) return(-1);
     if (out == NULL) return(-1);
@@ -2632,8 +2692,6 @@ retry:
         ret = -3;
     }
 
-    if (ret >= 0) output += ret;
-
     /*
      * Attempt to handle error cases
      */
@@ -2715,7 +2773,7 @@ retry:
             goto retry;
 	}
     }
-    return(ret);
+    return(writtentot ? writtentot : ret);
 }
 
 /**