view release on metacpan or  search on metacpan

src/subversion/subversion/libsvn_subr/xml.c  view on Meta::CPAN

{
  /** the expat parser */
  XML_Parser parser;

  /** the SVN callbacks to call from the Expat callbacks */
  svn_xml_start_elem start_handler;
  svn_xml_end_elem end_handler;
  svn_xml_char_data data_handler;

  /** the user's baton for private data */
  void *baton;

  /** if non-@c NULL, an error happened while parsing */
  svn_error_t *error;

  /** where this object is allocated, so we can free it easily */
  apr_pool_t *pool;

};


/*** XML character validation ***/

svn_boolean_t
svn_xml_is_xml_safe(const char *data, apr_size_t len)
{
  const char *end = data + len;
  const char *p;

  if (! svn_utf__is_valid(data, len))
    return FALSE;

  for (p = data; p < end; p++)
    {
      unsigned char c = *p;

      if (svn_ctype_iscntrl(c))
        {
          if ((c != SVN_CTYPE_ASCII_TAB)
              && (c != SVN_CTYPE_ASCII_LINEFEED)
              && (c != SVN_CTYPE_ASCII_CARRIAGERETURN)
              && (c != SVN_CTYPE_ASCII_DELETE))
            return FALSE;
        }
    }
  return TRUE;
}





/*** XML escaping. ***/

/* ### ...?
 *
 * If *OUTSTR is @c NULL, set *OUTSTR to a new stringbuf allocated
 * in POOL, else append to the existing stringbuf there.
 */
static void
xml_escape_cdata(svn_stringbuf_t **outstr,
                 const char *data,
                 apr_size_t len,
                 apr_pool_t *pool)
{
  const char *end = data + len;
  const char *p = data, *q;

  if (*outstr == NULL)
    *outstr = svn_stringbuf_create_empty(pool);

  while (1)
    {
      /* Find a character which needs to be quoted and append bytes up
         to that point.  Strictly speaking, '>' only needs to be
         quoted if it follows "]]", but it's easier to quote it all
         the time.

         So, why are we escaping '\r' here?  Well, according to the
         XML spec, '\r\n' gets converted to '\n' during XML parsing.
         Also, any '\r' not followed by '\n' is converted to '\n'.  By
         golly, if we say we want to escape a '\r', we want to make
         sure it remains a '\r'!  */
      q = p;
      while (q < end && *q != '&' && *q != '<' && *q != '>' && *q != '\r')
        q++;
      svn_stringbuf_appendbytes(*outstr, p, q - p);

      /* We may already be a winner.  */
      if (q == end)
        break;

      /* Append the entity reference for the character.  */
      if (*q == '&')
        svn_stringbuf_appendcstr(*outstr, "&amp;");
      else if (*q == '<')
        svn_stringbuf_appendcstr(*outstr, "&lt;");
      else if (*q == '>')
        svn_stringbuf_appendcstr(*outstr, "&gt;");
      else if (*q == '\r')
        svn_stringbuf_appendcstr(*outstr, "&#13;");

      p = q + 1;
    }
}

/* Essentially the same as xml_escape_cdata, with the addition of
   whitespace and quote characters. */
static void
xml_escape_attr(svn_stringbuf_t **outstr,
                const char *data,
                apr_size_t len,
                apr_pool_t *pool)
{
  const char *end = data + len;
  const char *p = data, *q;

  if (*outstr == NULL)
    *outstr = svn_stringbuf_create_ensure(len, pool);

  while (1)
    {
      /* Find a character which needs to be quoted and append bytes up
         to that point. */
      q = p;
      while (q < end && *q != '&' && *q != '<' && *q != '>'
             && *q != '"' && *q != '\'' && *q != '\r'
             && *q != '\n' && *q != '\t')
        q++;
      svn_stringbuf_appendbytes(*outstr, p, q - p);

      /* We may already be a winner.  */
      if (q == end)
        break;

      /* Append the entity reference for the character.  */
      if (*q == '&')
        svn_stringbuf_appendcstr(*outstr, "&amp;");
      else if (*q == '<')
        svn_stringbuf_appendcstr(*outstr, "&lt;");
      else if (*q == '>')
        svn_stringbuf_appendcstr(*outstr, "&gt;");
      else if (*q == '"')
        svn_stringbuf_appendcstr(*outstr, "&quot;");
      else if (*q == '\'')
        svn_stringbuf_appendcstr(*outstr, "&apos;");
      else if (*q == '\r')
        svn_stringbuf_appendcstr(*outstr, "&#13;");
      else if (*q == '\n')
        svn_stringbuf_appendcstr(*outstr, "&#10;");
      else if (*q == '\t')
        svn_stringbuf_appendcstr(*outstr, "&#9;");

      p = q + 1;
    }
}


void
svn_xml_escape_cdata_stringbuf(svn_stringbuf_t **outstr,
                               const svn_stringbuf_t *string,
                               apr_pool_t *pool)
{
  xml_escape_cdata(outstr, string->data, string->len, pool);
}


void
svn_xml_escape_cdata_string(svn_stringbuf_t **outstr,
                            const svn_string_t *string,
                            apr_pool_t *pool)
{
  xml_escape_cdata(outstr, string->data, string->len, pool);
}


void
svn_xml_escape_cdata_cstring(svn_stringbuf_t **outstr,
                             const char *string,
                             apr_pool_t *pool)
{
  xml_escape_cdata(outstr, string, (apr_size_t) strlen(string), pool);
}


void
svn_xml_escape_attr_stringbuf(svn_stringbuf_t **outstr,
                              const svn_stringbuf_t *string,
                              apr_pool_t *pool)
{
  xml_escape_attr(outstr, string->data, string->len, pool);
}


void
svn_xml_escape_attr_string(svn_stringbuf_t **outstr,
                           const svn_string_t *string,
                           apr_pool_t *pool)
{
  xml_escape_attr(outstr, string->data, string->len, pool);
}


void
svn_xml_escape_attr_cstring(svn_stringbuf_t **outstr,
                            const char *string,
                            apr_pool_t *pool)
{
  xml_escape_attr(outstr, string, (apr_size_t) strlen(string), pool);
}


const char *
svn_xml_fuzzy_escape(const char *string, apr_pool_t *pool)
{
  const char *end = string + strlen(string);
  const char *p = string, *q;
  svn_stringbuf_t *outstr;
  char escaped_char[6];   /* ? \ u u u \0 */

  for (q = p; q < end; q++)
    {
      if (svn_ctype_iscntrl(*q)
          && ! ((*q == '\n') || (*q == '\r') || (*q == '\t')))
        break;
    }

  /* Return original string if no unsafe characters found. */
  if (q == end)
    return string;

  outstr = svn_stringbuf_create_empty(pool);
  while (1)
    {
      q = p;

      /* Traverse till either unsafe character or eos. */
      while ((q < end)
             && ((! svn_ctype_iscntrl(*q))
                 || (*q == '\n') || (*q == '\r') || (*q == '\t')))
        q++;

src/subversion/subversion/libsvn_subr/xml.c  view on Meta::CPAN

  return ht;
}


void
svn_xml_hash_atts_overlaying(const char **atts,
                             apr_hash_t *ht,
                             apr_pool_t *pool)
{
  amalgamate(atts, ht, 0, pool);
}


void
svn_xml_hash_atts_preserving(const char **atts,
                             apr_hash_t *ht,
                             apr_pool_t *pool)
{
  amalgamate(atts, ht, 1, pool);
}



/*** Making XML tags. ***/


void
svn_xml_make_open_tag_hash(svn_stringbuf_t **str,
                           apr_pool_t *pool,
                           enum svn_xml_open_tag_style style,
                           const char *tagname,
                           apr_hash_t *attributes)
{
  apr_hash_index_t *hi;
  apr_size_t est_size = strlen(tagname) + 4 + apr_hash_count(attributes) * 30;

  if (*str == NULL)
    *str = svn_stringbuf_create_ensure(est_size, pool);

  svn_stringbuf_appendcstr(*str, "<");
  svn_stringbuf_appendcstr(*str, tagname);

  for (hi = apr_hash_first(pool, attributes); hi; hi = apr_hash_next(hi))
    {
      const void *key;
      void *val;

      apr_hash_this(hi, &key, NULL, &val);
      assert(val != NULL);

      svn_stringbuf_appendcstr(*str, "\n   ");
      svn_stringbuf_appendcstr(*str, key);
      svn_stringbuf_appendcstr(*str, "=\"");
      svn_xml_escape_attr_cstring(str, val, pool);
      svn_stringbuf_appendcstr(*str, "\"");
    }

  if (style == svn_xml_self_closing)
    svn_stringbuf_appendcstr(*str, "/");
  svn_stringbuf_appendcstr(*str, ">");
  if (style != svn_xml_protect_pcdata)
    svn_stringbuf_appendcstr(*str, "\n");
}


void
svn_xml_make_open_tag_v(svn_stringbuf_t **str,
                        apr_pool_t *pool,
                        enum svn_xml_open_tag_style style,
                        const char *tagname,
                        va_list ap)
{
  apr_pool_t *subpool = svn_pool_create(pool);
  apr_hash_t *ht = svn_xml_ap_to_hash(ap, subpool);

  svn_xml_make_open_tag_hash(str, pool, style, tagname, ht);
  svn_pool_destroy(subpool);
}



void
svn_xml_make_open_tag(svn_stringbuf_t **str,
                      apr_pool_t *pool,
                      enum svn_xml_open_tag_style style,
                      const char *tagname,
                      ...)
{
  va_list ap;

  va_start(ap, tagname);
  svn_xml_make_open_tag_v(str, pool, style, tagname, ap);
  va_end(ap);
}


void svn_xml_make_close_tag(svn_stringbuf_t **str,
                            apr_pool_t *pool,
                            const char *tagname)
{
  if (*str == NULL)
    *str = svn_stringbuf_create_empty(pool);

  svn_stringbuf_appendcstr(*str, "</");
  svn_stringbuf_appendcstr(*str, tagname);
  svn_stringbuf_appendcstr(*str, ">\n");
}