path: root/lib/erl_interface/doc/src/ei.xml

<?xml version="1.0" encoding="utf-8" ?>
<!DOCTYPE cref SYSTEM "cref.dtd">

<cref>
  <header>
    <copyright>
      <year>2001</year><year>2017</year>
      <holder>Ericsson AB. All Rights Reserved.</holder>
    </copyright>
    <legalnotice>
      Licensed under the Apache License, Version 2.0 (the "License");
      you may not use this file except in compliance with the License.
      You may obtain a copy of the License at

          http://www.apache.org/licenses/LICENSE-2.0

      Unless required by applicable law or agreed to in writing, software
      distributed under the License is distributed on an "AS IS" BASIS,
      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
      See the License for the specific language governing permissions and
      limitations under the License.

    </legalnotice>

    <title>ei</title>
    <prepared>Jakob Cederlund</prepared>
    <responsible>Kent Boortz</responsible>
    <docno>1</docno>
    <approved>Kenneth Lundin</approved>
    <checked></checked>
    <date>2000-11-27</date>
    <rev>PA1</rev>
    <file>ei.xml</file>
  </header>
  <lib>ei</lib>
  <libsummary>Routines for handling the Erlang binary term format.</libsummary>
  <description>
    <p>The library <c>ei</c> contains macros and functions to encode
      and decode the Erlang binary term format.</p>

    <p><c>ei</c> allows you to convert atoms, lists, numbers, and
      binaries to and from the binary format. This is useful when
      writing port programs and drivers. <c>ei</c> uses a given
      buffer, no dynamic memory (except
      <c>ei_decode_fun()</c>) and is often quite fast.</p>

    <p><c>ei</c> also handles C-nodes, C-programs that talks Erlang
      distribution with Erlang nodes (or other C-nodes) using the
      Erlang distribution format. The difference between <c>ei</c>
      and <c>erl_interface</c> is that <c>ei</c> uses
      the binary format directly when sending and receiving terms. It is also
      thread safe, and using threads, one process can handle multiple
      C-nodes. The <c>erl_interface</c> library is built on top of
      <c>ei</c>, but of legacy reasons, it does not allow for
      multiple C-nodes. In general, <c>ei</c> is the preferred way
      of doing C-nodes.</p>

    <p>The decode and encode functions use a buffer and an index into the
      buffer, which points at the point where to encode and
      decode. The index is updated to point right after the term
      encoded/decoded. No checking is done whether the term fits in
      the buffer or not. If encoding goes outside the buffer, the
      program can crash.</p>

    <p>All functions take two parameters:</p>

    <list type="bulleted">
      <item><p><c>buf</c> is a pointer to
        the buffer where the binary data is or will be.</p>
      </item>
      <item><p><c>index</c> is a pointer to an index into the
        buffer. This parameter is incremented with the size of the term
        decoded/encoded.</p>
      </item>
    </list>

    <p>The data is thus at <c>buf[*index]</c> when an
      <c>ei</c> function is called.</p>

    <p>All encode functions assume that the <c>buf</c> and
      <c>index</c> parameters point to a buffer large enough for
      the data. To get the size of an encoded term, without encoding it,
      pass <c>NULL</c> instead of a buffer pointer. Parameter
      <c>index</c> is incremented, but nothing will be encoded. This
      is the way in <c>ei</c> to "preflight" term encoding.</p>

    <p>There are also encode functions that use a dynamic buffer. It
      is often more convenient to use these to encode data. All encode
      functions comes in two versions; those starting with
      <c>ei_x</c> use a dynamic buffer.</p>

    <p>All functions return <c>0</c> if successful, otherwise
      <c>-1</c> (for example, if a term is not of the expected
      type, or the data to decode is an invalid Erlang term).</p>

    <p>Some of the decode functions need a pre-allocated buffer. This
      buffer must be allocated large enough, and for non-compound types
      the <c>ei_get_type()</c>
      function returns the size required (notice that for strings an
      extra byte is needed for the <c>NULL</c>-terminator).</p>
  </description>

  <section>
    <title>Data Types</title>
    <taglist>
      <tag><marker id="erlang_char_encoding"/>erlang_char_encoding</tag>
       <item>
	 <code type="none">
typedef enum {
    ERLANG_ASCII = 1,
    ERLANG_LATIN1 = 2,
    ERLANG_UTF8 = 4
} erlang_char_encoding;</code>
         <p>The character encodings used for atoms. <c>ERLANG_ASCII</c>
           represents 7-bit ASCII. Latin-1 and UTF-8 are different extensions
           of 7-bit ASCII. All 7-bit ASCII characters are valid Latin-1 and
           UTF-8 characters. ASCII and Latin-1 both represent each character
           by one byte. An UTF-8 character can consist of 1-4 bytes.
           Notice that these constants are bit-flags and can be combined with
           bitwise OR.</p>
      </item>
    </taglist>
  </section>

  <funcs>
      <func>
          <name><ret>int</ret><nametext>ei_decode_atom(const char *buf, int *index, char *p)</nametext></name>
          <fsummary>Decode an atom.</fsummary>
          <desc>
              <p>Decodes an atom from the binary format. The <c>NULL</c>-terminated
                  name of the atom is placed at <c>p</c>. At most
                  <c>MAXATOMLEN</c> bytes can be placed in the buffer.</p>
          </desc>
      </func>

      <func>
          <name><ret>int</ret><nametext>ei_decode_atom_as(const char *buf, int *index, char *p, int plen, erlang_char_encoding want, erlang_char_encoding* was, erlang_char_encoding* result)</nametext></name>
          <fsummary>Decode an atom.</fsummary>
          <desc>
              <p>Decodes an atom from the binary format. The <c>NULL</c>-terminated
                  name of the atom is placed in buffer at <c>p</c> of length <c>plen</c>
                  bytes.</p>
              <p>The wanted string encoding is specified by
                  <seealso marker="#erlang_char_encoding"><c>want</c></seealso>.
                  The original encoding used in the binary format (Latin-1 or UTF-8) can
                  be obtained from <c>*was</c>. The encoding of the resulting string
                  (7-bit ASCII, Latin-1, or UTF-8) can be obtained from <c>*result</c>.
                  Both <c>was</c> and <c>result</c> can be <c>NULL</c>. <c>*result</c>
                  can differ from <c>want</c> if <c>want</c> is a bitwise OR'd
                  combination like <c>ERLANG_LATIN1|ERLANG_UTF8</c> or if
                  <c>*result</c> turns out to be pure 7-bit ASCII
                  (compatible with both Latin-1 and UTF-8).</p>
              <p>This function fails if the atom is too long for the buffer
                  or if it cannot be represented with encoding <c>want</c>.</p>
              <p>This function was introduced in Erlang/OTP R16 as part of a first
                  step to support UTF-8 atoms.</p>
          </desc>
      </func>

      <func>
          <name><ret>int</ret><nametext>ei_decode_bignum(const char *buf, int *index, mpz_t obj)</nametext></name>
          <fsummary>Decode a GMP arbitrary precision integer.</fsummary>
          <desc>
              <p>Decodes an integer in the binary format to a GMP
                  <c>mpz_t</c> integer. To use this function, the <c>ei</c>
                  library must be configured and compiled to use the GMP library.</p>
          </desc>
      </func>

      <func>
          <name><ret>int</ret><nametext>ei_decode_binary(const char *buf, int *index, void *p, long *len)</nametext></name>
          <fsummary>Decode a binary.</fsummary>
          <desc>
              <p>Decodes a binary from the binary format. Parameter
                  <c>len</c> is set to the actual size of the
                  binary. Notice that <c>ei_decode_binary()</c> assumes that
                  there is enough room for the binary. The size required can be
                  fetched by <c>ei_get_type()</c>.</p>
          </desc>
      </func>

      <func>
          <name><ret>int</ret><nametext>ei_decode_boolean(const char *buf, int *index, int *p)</nametext></name>
          <fsummary>Decode a boolean.</fsummary>
          <desc>
              <p>Decodes a boolean value from the binary format.
                  A boolean is actually an atom, <c>true</c> decodes 1
                  and <c>false</c> decodes 0.</p>
          </desc>
      </func>

      <func>
          <name><ret>int</ret><nametext>ei_decode_char(const char *buf, int *index, char *p)</nametext></name>
          <fsummary>Decode an 8-bit integer between 0-255.</fsummary>
          <desc>
              <p>Decodes a char (8-bit) integer between 0-255 from the binary format.
                  For historical reasons the returned integer is of
                  type <c>char</c>. Your C code is to  consider the
                  returned value to be of type <c>unsigned char</c> even if
                  the C compilers and system can define <c>char</c> to be
                  signed.</p>
          </desc>
      </func>

      <func>
          <name><ret>int</ret><nametext>ei_decode_double(const char *buf, int *index, double *p)</nametext></name>
          <fsummary>Decode a double.</fsummary>
          <desc>
              <p>Decodes a double-precision (64-bit) floating
                  point number from the binary format.</p>
          </desc>
      </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_ei_term(const char* buf, int* index, ei_term* term)</nametext></name>
      <fsummary>Decode a term, without previous knowledge of type.</fsummary>
      <desc>
        <p>Decodes any term, or at least tries to. If the term
          pointed at by <c>*index</c> in <c>buf</c> fits
          in the <c>term</c> union, it is decoded, and the
          appropriate field in <c>term->value</c> is set, and
          <c>*index</c> is incremented by the term size.</p>
        <p>The function returns <c>1</c> on successful decoding, <c>-1</c> on
          error, and <c>0</c> if the term seems alright, but does not fit in the
          <c>term</c> structure. If <c>1</c> is returned, the
          <c>index</c> is incremented, and <c>term</c>
          contains the decoded term.</p>
        <p>The <c>term</c> structure contains the arity for a tuple
          or list, size for a binary, string, or atom. It contains
          a term if it is any of the following: integer, float, atom,
          pid, port, or ref.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_fun(const char *buf, int *index, erlang_fun *p)</nametext></name>
      <name><ret>void</ret><nametext>free_fun(erlang_fun* f)</nametext></name>
      <fsummary>Decode a fun.</fsummary>
      <desc>
        <p>Decodes a fun from the binary format. Parameter
          <c>p</c> is to be <c>NULL</c> or point to an
          <c>erlang_fun</c> structure. This is the only decode
          function that allocates memory. When the <c>erlang_fun</c>
          is no longer needed, it is to be freed with
          <c>free_fun</c>. (This has to do with the arbitrary size
          of the environment for a fun.)</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_list_header(const char *buf, int *index, int *arity)</nametext></name>
      <fsummary>Decode a list.</fsummary>
      <desc>
        <p>Decodes a list header from the binary
          format. The number of elements is returned in
          <c>arity</c>. The <c>arity+1</c> elements
          follow (the last one is the tail of the list, normally an empty list).
          If <c>arity</c> is <c>0</c>, it is an empty
          list.</p>
        <p>Notice that lists are encoded as strings if they consist
          entirely of integers in the range 0..255. This function do
          not decode such strings, use <c>ei_decode_string()</c>
          instead.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_long(const char *buf, int *index, long *p)</nametext></name>
      <fsummary>Decode integer.</fsummary>
      <desc>
        <p>Decodes a long integer from the binary format.
          If the code is 64 bits, the function <c>ei_decode_long()</c> is
          the same as <c>ei_decode_longlong()</c>.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_longlong(const char *buf, int *index, long long *p)</nametext></name>
      <fsummary>Decode integer.</fsummary>
      <desc>
        <p>Decodes a GCC <c>long long</c> or Visual C++
          <c>__int64</c>
          (64-bit) integer from the binary format. This
          function is missing in the VxWorks port.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_map_header(const char *buf, int *index, int *arity)</nametext></name>
      <fsummary>Decode a map.</fsummary>
      <desc>
        <p>Decodes a map header from the binary
          format. The number of key-value pairs is returned in
          <c>*arity</c>. Keys and values follow in this order:
	  <c>K1, V1, K2, V2, ..., Kn, Vn</c>. This makes a total of
	  <c>arity*2</c> terms. If <c>arity</c> is zero, it is an empty map.
	  A correctly encoded map does not have duplicate keys.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_pid(const char *buf, int *index, erlang_pid *p)</nametext></name>
      <fsummary>Decode a <c>pid</c>.</fsummary>
      <desc>
        <p>Decodes a process identifier (pid) from the binary format.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_port(const char *buf, int *index, erlang_port *p)</nametext></name>
      <fsummary>Decode a port.</fsummary>
      <desc>
        <p>Decodes a port identifier from the binary format.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_ref(const char *buf, int *index, erlang_ref *p)</nametext></name>
      <fsummary>Decode a reference.</fsummary>
      <desc>
        <p>Decodes a reference from the binary format.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_string(const char *buf, int *index, char *p)</nametext></name>
      <fsummary>Decode a string.</fsummary>
      <desc>
        <p>Decodes a string from the binary format. A
          string in Erlang is a list of integers between 0 and
          255. Notice that as the string is just a list, sometimes
          lists are encoded as strings by <c>term_to_binary/1</c>,
          even if it was not intended.</p>
        <p>The string is copied to <c>p</c>, and enough space must
          be allocated. The returned string is <c>NULL</c>-terminated, so you
          must add an extra byte to the memory requirement.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_term(const char *buf, int *index, void *t)</nametext></name>
      <fsummary>Decode a <c>ETERM</c>.</fsummary>
      <desc>
        <p>Decodes a term from the binary format. The term
          is return in <c>t</c> as a <c>ETERM*</c>, so
          <c>t</c> is actually an <c>ETERM**</c> (see
          <seealso marker="erl_eterm"><c>erl_eterm</c></seealso>).
          The term is later to be deallocated.</p>
        <p>Notice that this function is located in the <c>Erl_Interface</c>
          library.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_trace(const char *buf, int *index, erlang_trace *p)</nametext></name>
      <fsummary>Decode a trace token.</fsummary>
      <desc>
        <p>Decodes an Erlang trace token from the binary format.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_tuple_header(const char *buf, int *index, int *arity)</nametext></name>
      <fsummary>Decode a tuple.</fsummary>
      <desc>
        <p>Decodes a tuple header, the number of elements
          is returned in <c>arity</c>. The tuple elements follow
          in order in the buffer.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_ulong(const char *buf, int *index, unsigned long *p)</nametext></name>
      <fsummary>Decode unsigned integer.</fsummary>
      <desc>
        <p>Decodes an unsigned long integer from the binary format.
          If the code is 64 bits, the function <c>ei_decode_ulong()</c> is
          the same as <c>ei_decode_ulonglong()</c>.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_ulonglong(const char *buf, int *index, unsigned long long *p)</nametext></name>
      <fsummary>Decode unsigned integer.</fsummary>
      <desc>
        <p>Decodes a GCC <c>unsigned long long</c> or Visual C++
          <c>unsigned __int64</c> (64-bit) integer from the binary
          format. This function is missing in the VxWorks port.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_decode_version(const char *buf, int *index, int *version)</nametext></name>
      <fsummary>Decode an empty list (<c>nil</c>).</fsummary>
      <desc>
        <p>Decodes the version magic number for the
          Erlang binary term format. It must be the first token in a
          binary term.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_atom(char *buf, int *index, const char *p)</nametext></name>
      <name><ret>int</ret><nametext>ei_encode_atom_len(char *buf, int *index, const char *p, int len)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_atom(ei_x_buff* x, const char *p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_atom_len(ei_x_buff* x, const char *p, int len)</nametext></name>
      <fsummary>Encode an atom.</fsummary>
      <desc>
        <p>Encodes an atom in the binary format. Parameter <c>p</c>
          is the name of the atom in Latin-1 encoding. Only up to
          <c>MAXATOMLEN-1</c> bytes
          are encoded. The name is to be <c>NULL</c>-terminated, except for
          the <c>ei_x_encode_atom_len()</c> function.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_atom_as(char *buf, int *index, const char *p, erlang_char_encoding from_enc, erlang_char_encoding to_enc)</nametext></name>
      <name><ret>int</ret><nametext>ei_encode_atom_len_as(char *buf, int *index, const char *p, int len, erlang_char_encoding from_enc, erlang_char_encoding to_enc)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_atom_as(ei_x_buff* x, const char *p, erlang_char_encoding from_enc, erlang_char_encoding to_enc)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_atom_len_as(ei_x_buff* x, const char *p, int len, erlang_char_encoding from_enc, erlang_char_encoding to_enc)</nametext></name>
      <fsummary>Encode an atom.</fsummary>
      <desc>
        <p>Encodes an atom in the binary format. Parameter <c>p</c> is the name of the atom with
          character encoding
          <seealso marker="#erlang_char_encoding"><c>from_enc</c></seealso>
          (ASCII, Latin-1, or UTF-8). The name must either be <c>NULL</c>-terminated or
          a function variant with a <c>len</c> parameter must be used.</p>
	<p>The encoding fails if <c>p</c> is not a valid string in encoding
          <c>from_enc</c>.</p>

	<p>Argument <c>to_enc</c> is ignored. As from Erlang/OTP 20 the encoding is always
	  done in UTF-8 which is readable by nodes as old as Erlang/OTP R16.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_bignum(char *buf, int *index, mpz_t obj)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_bignum(ei_x_buff *x, mpz_t obj)</nametext></name>
      <fsummary>Encode an arbitrary precision integer.</fsummary>
      <desc>
        <p>Encodes a GMP <c>mpz_t</c> integer to binary format.
          To use this function, the <c>ei</c> library must be configured and
          compiled to use the GMP library.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_binary(char *buf, int *index, const void *p, long len)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_binary(ei_x_buff* x, const void *p, long len)</nametext></name>
      <fsummary>Encode a binary.</fsummary>
      <desc>
        <p>Encodes a binary in the binary format. The data is at
          <c>p</c>, of <c>len</c> bytes length.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_boolean(char *buf, int *index, int p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_boolean(ei_x_buff* x, int p)</nametext></name>
      <fsummary>Encode a boolean.</fsummary>
      <desc>
        <p>Encodes a boolean value as the atom <c>true</c> if
          <c>p</c> is not zero, or <c>false</c> if <c>p</c> is
          zero.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_char(char *buf, int *index, char p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_char(ei_x_buff* x, char p)</nametext></name>
      <fsummary>Encode an 8-bit integer between 0-255.</fsummary>
      <desc>
        <p>Encodes a char (8-bit) as an integer between 0-255 in the binary
          format. For historical reasons the integer argument is of
          type <c>char</c>. Your C code is to consider the specified
          argument to be of type <c>unsigned char</c> even if
          the C compilers and system may define <c>char</c> to be
          signed.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_double(char *buf, int *index, double p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_double(ei_x_buff* x, double p)</nametext></name>
      <fsummary>Encode a double float.</fsummary>
      <desc>
        <p>Encodes a double-precision (64-bit) floating point number in
          the binary format.</p>
        <p>Returns <c>-1</c> if the floating point
          number is not finite.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_empty_list(char* buf, int* index)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_empty_list(ei_x_buff* x)</nametext></name>
      <fsummary>Encode an empty list (<c>nil</c>).</fsummary>
      <desc>
        <p>Encodes an empty list. It is often used at the tail of a list.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_fun(char *buf, int *index, const erlang_fun *p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_fun(ei_x_buff* x, const erlang_fun* fun)</nametext></name>
      <fsummary>Encode a fun.</fsummary>
      <desc>
        <p>Encodes a fun in the binary format. Parameter <c>p</c>
          points to an <c>erlang_fun</c> structure. The
          <c>erlang_fun</c> is not freed automatically, the
          <c>free_fun</c> is to be called if the fun is not needed
          after encoding.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_list_header(char *buf, int *index, int arity)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_list_header(ei_x_buff* x, int arity)</nametext></name>
      <fsummary>Encode a list.</fsummary>
      <desc>
        <p>Encodes a list header, with a specified
          arity. The next <c>arity+1</c> terms are the elements
          (actually its <c>arity</c> cons cells) and the tail of the
          list. Lists and tuples are encoded recursively, so that a
          list can contain another list or tuple.</p>
        <p>For example, to encode the list
          <c>[c, d, [e | f]]</c>:</p>
        <pre>
ei_encode_list_header(buf, &amp;i, 3);
ei_encode_atom(buf, &amp;i, "c");
ei_encode_atom(buf, &amp;i, "d");
ei_encode_list_header(buf, &amp;i, 1);
ei_encode_atom(buf, &amp;i, "e");
ei_encode_atom(buf, &amp;i, "f");
ei_encode_empty_list(buf, &amp;i);</pre>
        <note>
          <p>It may seem that there is no way to create a list without
            knowing the number of elements in advance. But indeed
            there is a way. Notice that the list <c>[a, b, c]</c>
            can be written as <c>[a | [b | [c]]]</c>.
            Using this, a list can be written as conses.</p>
        </note>
        <p>To encode a list, without knowing the arity in advance:</p>
        <pre>
while (something()) {
    ei_x_encode_list_header(&amp;x, 1);
    ei_x_encode_ulong(&amp;x, i); /* just an example */
}
ei_x_encode_empty_list(&amp;x);</pre>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_long(char *buf, int *index, long p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_long(ei_x_buff* x, long p)</nametext></name>
      <fsummary>Encode integer.</fsummary>
      <desc>
        <p>Encodes a long integer in the binary format.
          If the code is 64 bits, the function <c>ei_encode_long()</c> is
          the same as <c>ei_encode_longlong()</c>.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_longlong(char *buf, int *index, long long p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_longlong(ei_x_buff* x, long long p)</nametext></name>
      <fsummary>Encode integer.</fsummary>
      <desc>
        <p>Encodes a GCC <c>long long</c> or Visual C++
          <c>__int64</c> (64-bit) integer in the binary format.
          This function is missing in the VxWorks port.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_map_header(char *buf, int *index, int arity)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_map_header(ei_x_buff* x, int arity)</nametext></name>
      <fsummary>Encode a map.</fsummary>
      <desc>
        <p>Encodes a map header, with a specified arity. The next
	   <c>arity*2</c> terms encoded will be the keys and values of the map
	   encoded in the following order: <c>K1, V1, K2, V2, ..., Kn, Vn</c>.
	</p>
        <p>For example, to encode the map <c>#{a => "Apple", b =>
          "Banana"}</c>:</p>
        <pre>
ei_x_encode_map_header(&amp;x, 2);
ei_x_encode_atom(&amp;x, "a");
ei_x_encode_string(&amp;x, "Apple");
ei_x_encode_atom(&amp;x, "b");
ei_x_encode_string(&amp;x, "Banana");</pre>
	<p>A correctly encoded map cannot have duplicate keys.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_pid(char *buf, int *index, const erlang_pid *p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_pid(ei_x_buff* x, const erlang_pid *p)</nametext></name>
      <fsummary>Encode a pid.</fsummary>
      <desc>
        <p>Encodes an Erlang process identifier (pid) in the binary
          format. Parameter <c>p</c> points to an
          <c>erlang_pid</c> structure (which should have been
          obtained earlier with <c>ei_decode_pid()</c>).</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_port(char *buf, int *index, const erlang_port *p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_port(ei_x_buff* x, const erlang_port *p)</nametext></name>
      <fsummary>Encode a port.</fsummary>
      <desc>
        <p>Encodes an Erlang port in the binary format. Parameter
          <c>p</c> points to a <c>erlang_port</c>
          structure (which should have been obtained earlier with
          <c>ei_decode_port()</c>).</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_ref(char *buf, int *index, const erlang_ref *p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_ref(ei_x_buff* x, const erlang_ref *p)</nametext></name>
      <fsummary>Encode a ref.</fsummary>
      <desc>
        <p>Encodes an Erlang reference in the binary format. Parameter
          <c>p</c> points to a <c>erlang_ref</c>
          structure (which should have been obtained earlier with
          <c>ei_decode_ref()</c>).</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_string(char *buf, int *index, const char *p)</nametext></name>
      <name><ret>int</ret><nametext>ei_encode_string_len(char *buf, int *index, const char *p, int len)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_string(ei_x_buff* x, const char *p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_string_len(ei_x_buff* x, const char* s, int len)</nametext></name>
      <fsummary>Encode a string.</fsummary>
      <desc>
        <p>Encodes a string in the binary format. (A string in Erlang
          is a list, but is encoded as a character array in the binary
          format.) The string is to be <c>NULL</c>-terminated, except for
          the <c>ei_x_encode_string_len()</c> function.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_term(char *buf, int *index, void *t)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_term(ei_x_buff* x, void *t)</nametext></name>
      <fsummary>Encode an <c>erl_interface</c> term.</fsummary>
      <desc>
        <p>Encodes an <c>ETERM</c>, as obtained from
          <c>erl_interface</c>. Parameter <c>t</c> is
          actually an <c>ETERM</c> pointer. This function
          does not free the <c>ETERM</c>.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>ei_encode_trace(char *buf, int *index, const erlang_trace *p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_trace(ei_x_buff* x, const erlang_trace *p)</nametext></name>
      <fsummary>Encode a trace token.</fsummary>
      <desc>
        <p>Encodes an Erlang trace token in the binary format.
          Parameter <c>p</c> points to a
          <c>erlang_trace</c> structure (which should have been
          obtained earlier with <c>ei_decode_trace()</c>).</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_tuple_header(char *buf, int *index, int arity)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_tuple_header(ei_x_buff* x, int arity)</nametext></name>
      <fsummary>Encode a tuple.</fsummary>
      <desc>
        <p>Encodes a tuple header, with a specified
          arity. The next <c>arity</c> terms encoded will be the
          elements of the tuple. Tuples and lists are encoded
          recursively, so that a tuple can contain another tuple or list.</p>
        <p>For example, to encode the tuple <c>{a, {b, {}}}</c>:</p>
        <pre>
ei_encode_tuple_header(buf, &amp;i, 2);
ei_encode_atom(buf, &amp;i, "a");
ei_encode_tuple_header(buf, &amp;i, 2);
ei_encode_atom(buf, &amp;i, "b");
ei_encode_tuple_header(buf, &amp;i, 0);</pre>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_ulong(char *buf, int *index, unsigned long p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_ulong(ei_x_buff* x, unsigned long p)</nametext></name>
      <fsummary>Encode unsigned integer.</fsummary>
      <desc>
        <p>Encodes an unsigned long integer in the binary format.
          If the code is 64 bits, the function <c>ei_encode_ulong()</c> is
          the same as <c>ei_encode_ulonglong()</c>.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_ulonglong(char *buf, int *index, unsigned long long p)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_ulonglong(ei_x_buff* x, unsigned long long p)</nametext></name>
      <fsummary>Encode unsigned integer.</fsummary>
      <desc>
        <p>Encodes a GCC <c>unsigned long long</c> or Visual C++
          <c>unsigned __int64</c> (64-bit) integer in the binary
          format. This function is missing in the VxWorks port.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_encode_version(char *buf, int *index)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_encode_version(ei_x_buff* x)</nametext></name>
      <fsummary>Encode version.</fsummary>
      <desc>
        <p>Encodes a version magic number for the binary format. Must
          be the first token in a binary term.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_get_type(const char *buf, const int *index, int *type, int *size)</nametext></name>
      <fsummary>Fetch the type and size of an encoded term.</fsummary>
      <desc>
        <p>Returns the type in <c>type</c> and size in
          <c>size</c> of the encoded term. For strings and atoms,
          size is the number of characters <em>not</em> including the
          terminating <c>NULL</c>. For binaries, <c>size</c> is the number of
          bytes. For lists and tuples, <c>size</c> is the arity of
          the object. For other types, <c>size</c> is 0. In all
          cases, <c>index</c> is left unchanged.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_init(void)</nametext></name>
      <fsummary>Initialize the ei library.</fsummary>
      <desc>
        <p>Initialize the <c>ei</c> library. This function should be called once
	(and only once) before calling any other functionality in the <c>ei</c>
	library. However, note the exception below.</p>
	<p>If the <c>ei</c> library is used together with the <c>erl_interface</c>
	library, this function should <em>not</em> be called directly. It will be
	called by the <c>erl_init()</c> function which should be used to initialize
	the combination of the two libraries instead.</p>
	<p>On success zero is returned. On failure a posix error code is returned.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_print_term(FILE* fp, const char* buf, int* index)</nametext></name>
      <name><ret>int</ret><nametext>ei_s_print_term(char** s, const char* buf, int* index)</nametext></name>
      <fsummary>Print a term in clear text.</fsummary>
      <desc>
        <p>Prints a term, in clear text, to the file
          specified by <c>fp</c>, or the buffer pointed to by
          <c>s</c>. It
          tries to resemble the term printing in the Erlang shell.</p>
        <p>In <c>ei_s_print_term()</c>, parameter
          <c>s</c> is to
          point to a dynamically (malloc) allocated string of
          <c>BUFSIZ</c> bytes or a <c>NULL</c> pointer. The string
          can be reallocated (and <c>*s</c> can be updated) by this
          function if the result is more than <c>BUFSIZ</c>
          characters. The string returned is <c>NULL</c>-terminated.</p>
        <p>The return value is the number of characters written to the file
          or string, or <c>-1</c> if <c>buf[index]</c> does not
          contain a valid term.
          Unfortunately, I/O errors on <c>fp</c> is not checked.</p>
        <p>Argument <c>index</c> is updated, that is, this function
          can be viewed as a decode function that decodes a term into a
          human-readable format.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>ei_set_compat_rel(release_number)</nametext></name>
      <fsummary>Set the ei library in compatibility mode.</fsummary>
      <type>
        <v>unsigned release_number;</v>
      </type>
      <desc>
        <marker id="ei_set_compat_rel"></marker>
        <p>By default, the <c>ei</c> library is only guaranteed
          to be compatible with other Erlang/OTP components from the same
          release as the <c>ei</c> library itself. For example,
          <c>ei</c> from
          Erlang/OTP R10 is not compatible with an Erlang emulator
          from Erlang/OTP R9 by default.</p>
        <p>A call to <c>ei_set_compat_rel(release_number)</c> sets
          the <c>ei</c> library in compatibility mode of release
          <c>release_number</c>. Valid range of
          <c>release_number</c>
          is <c>[7, current release]</c>. This makes it possible to
          communicate with Erlang/OTP components from earlier releases.</p>
        <note>
          <p>If this function is called, it can only be called once
            and must be called before any other functions in the
            <c>ei</c> library are called.</p>
        </note>
        <warning>
          <p>You can run into trouble if this feature is used
            carelessly. Always ensure that all communicating
            components are either from the same Erlang/OTP release, or
            from release X and release Y where all components
            from release Y are in compatibility mode of release X.</p>
        </warning>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_skip_term(const char* buf, int* index)</nametext></name>
      <fsummary>Skip a term.</fsummary>
      <desc>
        <p>Skips a term in the specified buffer;
          recursively skips elements of lists and tuples, so that a
          full term is skipped. This is a way to get the size of an
          Erlang term.</p>
        <p><c>buf</c> is the buffer.</p>
        <p><c>index</c> is updated to point right after the term
          in the buffer.</p>
        <note>
          <p>This can be useful when you want to hold arbitrary
            terms: skip them and copy the binary term data to some
            buffer.</p>
        </note>
        <p>Returns <c>0</c> on success, otherwise
          <c>-1</c>.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_x_append(ei_x_buff* x, const ei_x_buff* x2)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_append_buf(ei_x_buff* x, const char* buf, int len)</nametext></name>
      <fsummary>Append a buffer at the end.</fsummary>
      <desc>
        <p>Appends data at the end of buffer <c>x</c>.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_x_format(ei_x_buff* x, const char* fmt, ...)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_format_wo_ver(ei_x_buff* x, const char *fmt, ... )</nametext></name>
      <fsummary>Format a term from a format string and parameters.</fsummary>
      <desc>
        <p>Formats a term, given as a string, to a buffer.
          Works like a sprintf for Erlang terms.
          <c>fmt</c> contains a format string, with arguments like
          <c>~d</c>, to insert terms from variables. The following
          formats are supported (with the C types given):</p>
        <pre>
~a  An atom, char*
~c  A character, char
~s  A string, char*
~i  An integer, int
~l  A long integer, long int
~u  A unsigned long integer, unsigned long int
~f  A float, float
~d  A double float, double float
~p  An Erlang pid, erlang_pid*</pre>
        <p>For example, to encode a tuple with some stuff:</p>
        <pre>
ei_x_format("{~a,~i,~d}", "numbers", 12, 3.14159)
encodes the tuple {numbers,12,3.14159}</pre>
        <p><c>ei_x_format_wo_ver()</c> formats into a buffer,
          without the initial version byte.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_x_free(ei_x_buff* x)</nametext></name>
      <fsummary>Free a buffer.</fsummary>
      <desc>
        <p>Frees an <c>ei_x_buff</c> buffer.
          The memory used by the buffer is returned to the OS.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>ei_x_new(ei_x_buff* x)</nametext></name>
      <name><ret>int</ret><nametext>ei_x_new_with_version(ei_x_buff* x)</nametext></name>
      <fsummary>Allocate a new buffer.</fsummary>
      <desc>
        <p>Allocates a new <c>ei_x_buff</c> buffer. The
          fields of the structure pointed to by parameter <c>x</c>
          is filled in, and a default buffer is allocated.
          <c>ei_x_new_with_version()</c> also puts an initial
          version byte, which is used in the binary format (so that
          <c>ei_x_encode_version()</c> will not be needed.)</p>
      </desc>
    </func>
  </funcs>

  <section>
    <title>Debug Information</title>
    <p>Some tips on what to check when the emulator does not seem to
      receive the terms that you send:</p>

    <list type="bulleted">
      <item>Be careful with the version header, use
        <c>ei_x_new_with_version()</c> when appropriate.</item>
      <item>Turn on distribution tracing on the Erlang node.</item>
      <item>Check the result codes from <c>ei_decode_-calls</c>.</item>
    </list>
  </section>

  <section>
    <title>See Also</title>
    <p><seealso marker="erl_eterm"><c>erl_eterm</c></seealso></p>
  </section>
</cref>