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>
    <note><p>The support for VxWorks is deprecated as of OTP 22, and
    will be removed in OTP 23.</p></note>
    
    <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 since=""><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 since="OTP R16B"><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 since=""><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 since=""><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 since="OTP 22.0"><ret>int</ret><nametext>ei_decode_bitstring(const char *buf, int *index, const char **pp, unsigned int *bitoffsp, size_t *nbitsp)</nametext></name>
          <fsummary>Decode a bitstring.</fsummary>
          <desc>
            <p>Decodes a bit string from the binary format.</p>
	    <taglist>
	      <tag><c>pp</c></tag>
	      <item><p>Either <c>NULL</c> or <c>*pp</c> returns a pointer to
	       the first byte of the bit string. The returned bit string is
	       readable as long as the buffer pointed to by <c>buf</c> is
	       readable and not written to.</p>
	      </item>
	      <tag><c>bitoffsp</c></tag>
	      <item><p>Either <c>NULL</c> or <c>*bitoffsp</c> returns the
	       number of unused bits in the first byte pointed to by
	       <c>*pp</c>. The value of <c>*bitoffsp</c> is between 0 and 7.
	       Unused bits in the first byte are the most significant bits.</p>
	      </item>
	      <tag><c>nbitsp</c></tag>
	      <item><p>Either <c>NULL</c> or <c>*nbitsp</c> returns the length
	       of the bit string in <em>bits</em>.</p>
	      </item>
	    </taglist>
	    <p>Returns <c>0</c> if it was a bit string term.</p>
	    <p>The number of <em>bytes</em> pointed to by <c>*pp</c>, which are
	      part of the bit string, is <c>(*bitoffsp + *nbitsp + 7)/8</c>. If
	      <c>(*bitoffsp + *bitsp)%8 > 0</c> then only <c>(*bitoffsp +
	      *bitsp)%8</c> bits of the last byte are used. Unused bits in
	      the last byte are the least significant bits.</p>
	    <p>The values of unused bits in the first and last byte are undefined
	      and cannot be relied on.</p>
	    <p>Number of bits may be divisible by 8, which means a binary
	      decodable by <c>ei_decode_binary</c> is also decodable by
	      <c>ei_decode_bitstring</c>.</p>
          </desc>
      </func>

      <func>
          <name since=""><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 since=""><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 since=""><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 since=""><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 since=""><ret>int</ret><nametext>ei_decode_fun(const char *buf, int *index, erlang_fun *p)</nametext></name>
      <name since=""><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 since=""><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 since=""><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 since=""><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 since="OTP 17.0"><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 since=""><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 since=""><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 since=""><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 since=""><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 since=""><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>
	<note><p>This function is deprecated as of OTP 22 and will be removed in
	  OTP 23 together with the old legacy <c>erl_interface</c> library (functions
	  with prefix <c>erl_</c>).</p>
	</note>
      </desc>
    </func>

    <func>
      <name since=""><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 since=""><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 since=""><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 since=""><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 since=""><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 since=""><ret>int</ret><nametext>ei_encode_atom(char *buf, int *index, const char *p)</nametext></name>
      <name since=""><ret>int</ret><nametext>ei_encode_atom_len(char *buf, int *index, const char *p, int len)</nametext></name>
      <name since=""><ret>int</ret><nametext>ei_x_encode_atom(ei_x_buff* x, const char *p)</nametext></name>
      <name since=""><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 since="OTP R16B"><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 since="OTP R16B"><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 since="OTP R16B"><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 since="OTP R16B"><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 since=""><ret>int</ret><nametext>ei_encode_bignum(char *buf, int *index, mpz_t obj)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_binary(char *buf, int *index, const void *p, long len)</nametext></name>
      <name since=""><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 since="OTP 22.0"><ret>int</ret>
      <nametext>ei_encode_bitstring(char *buf, int *index, const char *p, size_t bitoffs, size_t nbits)</nametext></name>
      <name since="OTP 22.0"><ret>int</ret>
      <nametext>ei_x_encode_bitstring(ei_x_buff* x, const char *p, size_t bitoffs, size_t nbits)</nametext></name>
      <fsummary>Encode a bitstring.</fsummary>
      <desc>
        <p>Encodes a bit string in the binary format.</p>
	<p>The data is at <c>p</c>. The length of the bit string is <c>nbits</c>
	  bits. The first <c>bitoffs</c> bits of the data at <c>p</c> are unused.
	  The first byte which is part of the bit string is
	  <c>p[bitoffs/8]</c>. The <c>bitoffs%8</c> most significant bits of
	  the first byte <c>p[bitoffs/8]</c> are unused.</p>
	<p>The number of bytes which is part of the bit string is <c>(bitoffs +
	  nbits + 7)/8</c>. If <c>(bitoffs + nbits)%8 > 0</c> then only <c>(bitoffs +
	  nbits)%8</c> bits of the last byte are used. Unused bits in
	  the last byte are the least significant bits.</p>
	<p>The values of unused bits are disregarded and does not need to be
	  cleared.</p>
      </desc>
    </func>

    <func>
      <name since=""><ret>int</ret><nametext>ei_encode_boolean(char *buf, int *index, int p)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_char(char *buf, int *index, char p)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_double(char *buf, int *index, double p)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_empty_list(char* buf, int* index)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_fun(char *buf, int *index, const erlang_fun *p)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_list_header(char *buf, int *index, int arity)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_long(char *buf, int *index, long p)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_longlong(char *buf, int *index, long long p)</nametext></name>
      <name since=""><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 since="OTP 17.0"><ret>int</ret><nametext>ei_encode_map_header(char *buf, int *index, int arity)</nametext></name>
      <name since="OTP 17.0"><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 since=""><ret>int</ret><nametext>ei_encode_pid(char *buf, int *index, const erlang_pid *p)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_port(char *buf, int *index, const erlang_port *p)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_ref(char *buf, int *index, const erlang_ref *p)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_string(char *buf, int *index, const char *p)</nametext></name>
      <name since=""><ret>int</ret><nametext>ei_encode_string_len(char *buf, int *index, const char *p, int len)</nametext></name>
      <name since=""><ret>int</ret><nametext>ei_x_encode_string(ei_x_buff* x, const char *p)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_term(char *buf, int *index, void *t)</nametext></name>
      <name since=""><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>
	<note><p>These functions are deprecated as of OTP 22 and will be removed in
	  OTP 23 together with the old legacy <c>erl_interface</c> library
	  (functions with prefix <c>erl_</c>).</p>
	</note>
      </desc>
    </func>
    <func>
      <name since=""><ret>int</ret><nametext>ei_encode_trace(char *buf, int *index, const erlang_trace *p)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_tuple_header(char *buf, int *index, int arity)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_ulong(char *buf, int *index, unsigned long p)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_ulonglong(char *buf, int *index, unsigned long long p)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_encode_version(char *buf, int *index)</nametext></name>
      <name since=""><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 since=""><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 and bitstrings, <c>*size</c> is
	  the number of bytes. For lists, tuples and maps, <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 since="OTP 21.3"><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 since=""><ret>int</ret><nametext>ei_print_term(FILE* fp, const char* buf, int* index)</nametext></name>
      <name since=""><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 since=""><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>In general, the <c>ei</c> library is guaranteed
          to be compatible with other Erlang/OTP components that are 2 major
	  releases older or newer than the <c>ei</c> library itself.</p>
        <p>Sometimes an exception to the above rule has to be made to make new
	  features (or even bug fixes) possible. A call to
	  <c>ei_set_compat_rel(release_number)</c> sets
          the <c>ei</c> library in compatibility mode of OTP release
          <c>release_number</c>.</p>
	<p>The only useful value for <c>release_number</c> is currently
	  <c>21</c>. This will only be useful and have an effect if <em>bit
	  strings</em> or <em>export funs</em> are received from a connected
	  node. Before OTP 22, bit strings and export funs were not supported by
	  <c>ei</c>. They were instead encoded using an undocumented fallback
	  tuple format when sent from the emulator to <c>ei</c>:</p>
	<taglist>
	  <tag><c>Bit string</c></tag>
	  <item><p>The term <c>&lt;&lt;42, 1:1>></c> was encoded as
	    <c>{&lt;&lt;42, 128>>, 1}</c>. The first element of the tuple is a
	    binary and the second element denotes how many bits of the last bytes
	    are part of the bit string. In this example only the most significant
	    bit of the last byte (128) is part of the bit string.</p>
	  </item>
	  <tag><c>Export fun</c></tag>
	  <item><p>The term <c>fun lists:map/2</c> was encoded as
	    <c>{lists,map}</c>. A tuple with the module, function and a missing
	    arity.</p>
	  </item>
	</taglist>
	<p>If <c>ei_set_compat_rel(21)</c> is <em>not</em> called then a connected
	  emulator will send bit strings and export funs correctly encoded. The
	  functions <seealso marker="#ei_decode_bitstring"><c>ei_decode_bitstring</c></seealso>
	  and <seealso marker="#ei_decode_fun"><c>ei_decode_fun</c></seealso>
	  has to be used to decode such terms. Calling
	  <c>ei_set_compat_rel(21)</c> should only be done as a workaround to
	  keep an old implementation alive, which expects to receive the
	  undocumented tuple formats for bit strings and/or export funs.
	</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>
      </desc>
    </func>

    <func>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_x_append(ei_x_buff* x, const ei_x_buff* x2)</nametext></name>
      <name since=""><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 since=""><ret>int</ret><nametext>ei_x_format(ei_x_buff* x, const char* fmt, ...)</nametext></name>
      <name since=""><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 since=""><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 since=""><ret>int</ret><nametext>ei_x_new(ei_x_buff* x)</nametext></name>
      <name since=""><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>