<?xml version="1.0" encoding="latin1" ?>
<!DOCTYPE chapter SYSTEM "chapter.dtd">

<chapter>
  <header>
    <copyright>
      <year>1996</year><year>2009</year>
      <holder>Ericsson AB. All Rights Reserved.</holder>
    </copyright>
    <legalnotice>
      The contents of this file are subject to the Erlang Public License,
      Version 1.1, (the "License"); you may not use this file except in
      compliance with the License. You should have received a copy of the
      Erlang Public License along with this software. If not, it can be
      retrieved online at http://www.erlang.org/.
    
      Software distributed under the License is distributed on an "AS IS"
      basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
      the License for the specific language governing rights and limitations
      under the License.
    
    </legalnotice>

    <title>The Erl_Interface Library</title>
    <prepared>Torbj&ouml;rn T&ouml;rnkvist</prepared>
    <responsible>Torbj&ouml;rn T&ouml;rnkvist</responsible>
    <docno></docno>
    <approved>Bjarne D&auml;cker</approved>
    <checked>K.Lundin</checked>
    <date>990113</date>
    <rev>A</rev>
    <file>erl_interface.sgml</file>
  </header>
  <p>The Erl_Interface library contains functions. which help you
    integrate programs written in C and Erlang. The functions in
    Erl_Interface support the following:</p>
  <list type="bulleted">
    <item>manipulation of data represented as Erlang data types</item>
    <item>conversion of data between C and Erlang formats</item>
    <item>encoding and decoding of Erlang data types for transmission or storage</item>
    <item>communication between C nodes and Erlang processes</item>
    <item>backup and restore of C node state to and from Mnesia</item>
  </list>
  <p>In the following sections, these topics are described:</p>
  <list type="bulleted">
    <item>compiling your code for use with Erl_Interface</item>
    <item>initializing Erl_Interface</item>
    <item>encoding, decoding, and sending Erlang terms</item>
    <item>building terms and patterns</item>
    <item>pattern matching</item>
    <item>connecting to a distributed Erlang node</item>
    <item>using EPMD</item>
    <item>sending and receiving Erlang messages</item>
    <item>remote procedure calls</item>
    <item>global names</item>
    <item>the registry</item>
  </list>

  <section>
    <title>Compiling and Linking Your Code</title>
    <p>In order to use any of the Erl_Interface functions, include the
      following lines in your code:</p>
    <code type="none"><![CDATA[
#include "erl_interface.h"
#include "ei.h"    ]]></code>
    <p>Determine where the top directory of your OTP installation is. You
      can find this out by starting Erlang and entering the following
      command at the Eshell prompt:</p>
    <code type="none"><![CDATA[
Eshell V4.7.4  (abort with ^G)
1> code:root_dir().
/usr/local/otp    ]]></code>
    <p>To compile your code, make sure that your C compiler knows where
      to find <c><![CDATA[erl_interface.h]]></c> by specifying an appropriate <c><![CDATA[-I]]></c>
      argument on the command line, or by adding it to the <c><![CDATA[CFLAGS]]></c>
      definition in your <c><![CDATA[Makefile]]></c>. The correct value for this path is
      <c><![CDATA[$OTPROOT/lib/erl_interface]]></c><em>Vsn</em><c><![CDATA[/include]]></c>, where <c><![CDATA[$OTPROOT]]></c> is the path
      reported by <c><![CDATA[code:root_dir/0]]></c> in the above example, and <em>Vsn</em> is
      the version of the Erl_interface application, for example 
      <c><![CDATA[erl_interface-3.2.3]]></c></p>
    <code type="none"><![CDATA[
$ cc -c -I/usr/local/otp/lib/erl_interface-3.2.3/include myprog.c    ]]></code>
    <p>When linking, you will need to specify the path to
      <c><![CDATA[liberl_interface.a]]></c> and <c><![CDATA[libei.a]]></c> with
      <c><![CDATA[-L$OTPROOT/lib/erl_interface-3.2.3/lib]]></c>, and you will need to specify the
      name of the libraries with <c><![CDATA[-lerl_interface -lei]]></c>. You can do
      this on the command line or by adding the flags to the <c><![CDATA[LDFLAGS]]></c>
      definition in your <c><![CDATA[Makefile]]></c>.</p>
    <code type="none"><![CDATA[
$ ld -L/usr/local/otp/lib/erl_interface-3.2.3/
                            lib myprog.o -lerl_interface -lei -o myprog    ]]></code>
    <p>Also, on some systems it may be necessary to link with some
      additional libraries (e.g. <c><![CDATA[libnsl.a]]></c> and <c><![CDATA[libsocket.a]]></c> on
      Solaris, or <c><![CDATA[wsock32.lib]]></c> on Windows) in order to use the
      communication facilities of Erl_Interface.</p>
    <p>If you are using Erl_Interface functions in a threaded
      application based on POSIX threads or Solaris threads, then
      Erl_Interface needs access to some of the synchronization
      facilities in your threads package, and you will need to specify
      additional compiler flags in order to indicate which of the packages
      you are using. Define <c><![CDATA[_REENTRANT]]></c> and either <c><![CDATA[STHREADS]]></c> or
      <c><![CDATA[PTHREADS]]></c>. The default is to use POSIX threads if
      <c><![CDATA[_REENTRANT]]></c> is specified.</p>
    <p>Note that both single threaded and default versions of the Erl_interface 
      and Ei libraries are provided. (The single threaded versions are named 
      <c><![CDATA[liberl_interface_st]]></c> and <c><![CDATA[libei_st]]></c>). Whether the default
      versions of the libraries have support for threads or not is determined by if 
      the platform in question has support for POSIX or Solaris threads. To check this, 
      have a look in the <c><![CDATA[eidefs.mk]]></c> file in the erl_interface src directory.</p>
  </section>

  <section>
    <title>Initializing the erl_interface Library</title>
    <p>Before calling any of the other Erl_Interface functions, you
      must call <c><![CDATA[erl_init()]]></c> exactly once to initialize the library.
      <c><![CDATA[erl_init()]]></c> takes two arguments, however the arguments are no
      longer used by Erl_Interface, and should therefore be specified
      as <c><![CDATA[erl_init(NULL,0)]]></c>.</p>
  </section>

  <section>
    <title>Encoding, Decoding and Sending Erlang Terms</title>
    <p>Data sent between distributed Erlang nodes is encoded in the
      Erlang external format. Consequently, you have to encode and decode
      Erlang terms into byte streams if you want to use the distribution
      protocol to communicate between a C program and Erlang. </p>
    <p>The Erl_Interface library supports this activity. It has a
      number of C functions which create and manipulate Erlang data
      structures. The library also contains an encode and a decode function.
      The example below shows how to create and encode an Erlang tuple
      <c><![CDATA[{tobbe,3928}]]></c>:</p>
    <code type="none"><![CDATA[

ETERM *arr[2], *tuple;
char buf[BUFSIZ];
int i;
  
arr[0] = erl_mk_atom("tobbe");
arr[1] = erl_mk_integer(3928);
tuple  = erl_mk_tuple(arr, 2);
i = erl_encode(tuple, buf);    ]]></code>
    <p>Alternatively, you can use <c><![CDATA[erl_send()]]></c> and
      <c><![CDATA[erl_receive_msg]]></c>, which handle the encoding and decoding of
      messages transparently.</p>
    <p>Refer to the Reference Manual for a complete description of the
      following modules:</p>
    <list type="bulleted">
      <item>the <c><![CDATA[erl_eterm]]></c> module for creating Erlang terms</item>
      <item>the <c><![CDATA[erl_marshal]]></c> module for encoding and decoding routines.</item>
    </list>
  </section>

  <section>
    <title>Building Terms and Patterns</title>
    <p>The previous example can be simplified by using 
      <c><![CDATA[erl_format()]]></c> to create an Erlang term.</p>
    <code type="none"><![CDATA[

ETERM *ep;
ep = erl_format("{~a,~i}", "tobbe", 3928);    ]]></code>
    <p>Refer to the Reference Manual, the <c><![CDATA[erl_format]]></c> module, for a
      full description of the different format directives. The following
      example is more complex:</p>
    <code type="none"><![CDATA[

ETERM *ep;
ep = erl_format("[{name,~a},{age,~i},{data,~w}]",
                 "madonna", 
                 21, 
                 erl_format("[{adr,~s,~i}]", "E-street", 42));
erl_free_compound(ep);      ]]></code>
    <p>As in previous examples, it is your responsibility to free the
      memory allocated for Erlang terms. In this example, 
      <c><![CDATA[erl_free_compound()]]></c> ensures that the complete term pointed to
      by <c><![CDATA[ep]]></c> is released. This is necessary, because the pointer from
      the second call to <c><![CDATA[erl_format()]]></c> is lost. </p>
    <p>The following
      example shows a slightly different solution:</p>
    <code type="none"><![CDATA[

ETERM *ep,*ep2;
ep2 = erl_format("[{adr,~s,~i}]","E-street",42);
ep  = erl_format("[{name,~a},{age,~i},{data,~w}]",
                 "madonna", 21, ep2);
erl_free_term(ep);  
erl_free_term(ep2);      ]]></code>
    <p>In this case, you free the two terms independently. The order in
      which you free the terms <c><![CDATA[ep]]></c> and <c><![CDATA[ep2]]></c> is not important,
      because the Erl_Interface library uses reference counting to
      determine when it is safe to actually remove objects. </p>
    <p>If you are not sure whether you have freed the terms properly, you
      can use the following function to see the status of the fixed term
      allocator:</p>
    <code type="none"><![CDATA[
long allocated, freed;

erl_eterm_statistics(&allocated,&freed);
printf("currently allocated blocks: %ld\
",allocated);
printf("length of freelist: %ld\
",freed);

/* really free the freelist */
erl_eterm_release();
    ]]></code>
    <p>Refer to the Reference Manual, the <c><![CDATA[erl_malloc]]></c> module for more
      information.</p>
  </section>

  <section>
    <title>Pattern Matching</title>
    <p>An Erlang pattern is a term that may contain unbound variables or
      <c><![CDATA["do not care"]]></c> symbols. Such a pattern can be matched against a
      term and, if the match is successful, any unbound variables in the
      pattern will be bound as a side effect. The content of a bound
      variable can then be retrieved.</p>
    <code type="none"><![CDATA[

ETERM *pattern;
pattern = erl_format("{madonna,Age,_}");    ]]></code>
    <p><c><![CDATA[erl_match()]]></c> is used to perform pattern matching. It takes a
      pattern and a term and tries to match them. As a side effect any unbound
      variables in the pattern will be bound. In the following example, we
      create a pattern with a variable <em>Age</em> which appears at two
      positions in the tuple. The pattern match is performed as follows:</p>
    <list type="ordered">
      <item><c><![CDATA[erl_match()]]></c> will bind the contents of
      <em>Age</em> to <em>21</em> the first time it reaches the variable</item>
      <item>the second occurrence of <em>Age</em> will cause a test for
       equality between the terms since <em>Age</em> is already bound to
      <em>21</em>. Since <em>Age</em> is bound to 21, the equality test will
       succeed and the match continues until the end of the pattern.</item>
      <item>if the end of the pattern is reached, the match succeeds and you
       can retrieve the contents of the variable</item>
    </list>
    <code type="none"><![CDATA[
ETERM *pattern,*term;
pattern = erl_format("{madonna,Age,Age}");
term    = erl_format("{madonna,21,21}");
if (erl_match(pattern, term)) {
  fprintf(stderr, "Yes, they matched: Age = ");
  ep = erl_var_content(pattern, "Age"); 
  erl_print_term(stderr, ep);
  fprintf(stderr,"\
");
  erl_free_term(ep);
}
erl_free_term(pattern);
erl_free_term(term);    ]]></code>
    <p>Refer to the Reference Manual, the <c><![CDATA[erl_match()]]></c> function for
      more information.</p>
  </section>

  <section>
    <title>Connecting to a Distributed Erlang Node</title>
    <p>In order to connect to a distributed Erlang node you need to first
      initialize the connection routine with <c><![CDATA[erl_connect_init()]]></c>,
      which stores information such as the host name, node name, and IP
      address for later use:</p>
    <code type="none"><![CDATA[
int identification_number = 99;
int creation=1;
char *cookie="a secret cookie string"; /* An example */
erl_connect_init(identification_number, cookie, creation);    ]]></code>
    <p>Refer to the Reference Manual, the <c><![CDATA[erl_connect]]></c> module for more information.</p>
    <p>After initialization, you set up the connection to the Erlang node.
      Use <c><![CDATA[erl_connect()]]></c> to specify the Erlang node you want to
      connect to. The following example sets up the connection and should
      result in a valid socket file descriptor:</p>
    <code type="none"><![CDATA[
int sockfd;
char *nodename="xyz@chivas.du.etx.ericsson.se"; /* An example */
if ((sockfd = erl_connect(nodename)) < 0)
  erl_err_quit("ERROR: erl_connect failed");    ]]></code>
    <p><c><![CDATA[erl_err_quit()]]></c> prints the specified string and terminates
      the program. Refer to the Reference Manual, the <c><![CDATA[erl_error()]]></c>
      function for more information.</p>
  </section>

  <section>
    <title>Using EPMD</title>
    <p><c><![CDATA[Epmd]]></c> is the Erlang Port Mapper Daemon. Distributed Erlang nodes
      register with <c><![CDATA[epmd]]></c> on the localhost to indicate to other nodes that
      they exist and can accept connections. <c><![CDATA[Epmd]]></c> maintains a register of
      node and port number information, and when a node wishes to connect to
      another node, it first contacts <c><![CDATA[epmd]]></c> in order to find out the correct
      port number to connect to.</p>
    <p>When you use <c><![CDATA[erl_connect()]]></c> to connect to an Erlang node, a
      connection is first made to <c><![CDATA[epmd]]></c> and, if the node is known, a
      connection is then made to the Erlang node.</p>
    <p>C nodes can also register themselves with <c><![CDATA[epmd]]></c> if they want other
      nodes in the system to be able to find and connect to them.</p>
    <p>Before registering with <c><![CDATA[epmd]]></c>, you need to first create a listen socket
      and bind it to a port. Then:</p>
    <code type="none"><![CDATA[
int pub;

pub = erl_publish(port);    ]]></code>
    <p><c><![CDATA[pub]]></c> is a file descriptor now connected to <c><![CDATA[epmd]]></c>. <c><![CDATA[Epmd]]></c>
      monitors the other end of the connection, and if it detects that the
      connection has been closed, the node will be unregistered. So, if you
      explicitly close the descriptor or if your node fails, it will be
      unregistered from <c><![CDATA[epmd]]></c>.</p>
    <p>Be aware that on some systems (such as VxWorks), a failed node will
      not be detected by this mechanism since the operating system does not
      automatically close descriptors that were left open when the node
      failed. If a node has failed in this way, <c><![CDATA[epmd]]></c> will prevent you from
      registering a new node with the old name, since it thinks that the old
      name is still in use. In this case, you must unregister the name
      explicitly:</p>
    <code type="none"><![CDATA[
erl_unpublish(node);    ]]></code>
    <p>This will cause <c><![CDATA[epmd]]></c> to close the connection from the far end. Note
      that if the name was in fact still in use by a node, the results of
      this operation are unpredictable. Also, doing this does not cause the
      local end of the connection to close, so resources may be consumed.</p>
  </section>

  <section>
    <title>Sending and Receiving Erlang Messages</title>
    <p>Use one of the following two functions to send messages:</p>
    <list type="bulleted">
      <item><c><![CDATA[erl_send()]]></c></item>
      <item><c><![CDATA[erl_reg_send()]]></c></item>
    </list>
    <p>As in Erlang, it is possible to send messages to a 
      <em>Pid</em> or to a registered name. It is easier to send a
      message to a registered name because it avoids the problem of finding
      a suitable <em>Pid</em>.</p>
    <p>Use one of the following two functions to receive messages:</p>
    <list type="bulleted">
      <item><c><![CDATA[erl_receive()]]></c></item>
      <item><c><![CDATA[erl_receive_msg()]]></c></item>
    </list>
    <p><c><![CDATA[erl_receive()]]></c> receives the message into a buffer, while
      <c><![CDATA[erl_receive_msg()]]></c> decodes the message into an Erlang term. </p>

    <section>
      <title>Example of Sending Messages</title>
      <p>In the following example, <c><![CDATA[{Pid, hello_world}]]></c> is
        sent to a registered process <c><![CDATA[my_server]]></c>. The message is encoded
        by <c><![CDATA[erl_send()]]></c>:</p>
      <code type="none"><![CDATA[
extern const char *erl_thisnodename(void);
extern short erl_thiscreation(void);
#define SELF(fd) erl_mk_pid(erl_thisnodename(),fd,0,erl_thiscreation())
ETERM *arr[2], *emsg;
int sockfd, creation=1;
  
arr[0] = SELF(sockfd);
arr[1] = erl_mk_atom("Hello world");
emsg   = erl_mk_tuple(arr, 2);
  
erl_reg_send(sockfd, "my_server", emsg);
erl_free_term(emsg);      ]]></code>
      <p>The first element of the tuple that is sent is your own
        <em>Pid</em>. This enables <c><![CDATA[my_server]]></c> to reply. Refer to the
        Reference Manual, the <c><![CDATA[erl_connect]]></c> module for more information
        about send primitives.</p>
    </section>

    <section>
      <title>Example of Receiving Messages</title>
      <p>In this example <c><![CDATA[{Pid, Something}]]></c> is received. The
        received Pid is then used to return <c><![CDATA[{goodbye,Pid}]]></c></p>
      <code type="none"><![CDATA[
ETERM *arr[2], *answer;
int sockfd,rc;
char buf[BUFSIZE];
ErlMessage emsg;
  
if ((rc = erl_receive_msg(sockfd , buf, BUFSIZE, &emsg)) == ERL_MSG) {
   arr[0] = erl_mk_atom("goodbye");
   arr[1] = erl_element(1, emsg.msg); 
   answer = erl_mk_tuple(arr, 2);
   erl_send(sockfd, arr[1], answer);
   erl_free_term(answer);
   erl_free_term(emsg.msg);
   erl_free_term(emsg.to);
}
}      ]]></code>
      <p>In order to provide robustness, a distributed Erlang node
        occasionally polls all its connected neighbours in an attempt to
        detect failed nodes or communication links. A node which receives such
        a message is expected to respond immediately with an <c><![CDATA[ERL_TICK]]></c> message.
        This is done automatically by <c><![CDATA[erl_receive()]]></c>, however when this
        has occurred <c><![CDATA[erl_receive]]></c> returns <c><![CDATA[ERL_TICK]]></c> to the caller
        without storing a message into the <c><![CDATA[ErlMessage]]></c> structure.</p>
      <p>When a message has been received, it is the caller's responsibility
        to free the received message <c><![CDATA[emsg.msg]]></c> as well as <c><![CDATA[emsg.to]]></c>
        or <c><![CDATA[emsg.from]]></c>, depending on the type of message received.</p>
      <p>Refer to the Reference Manual for additional information about the
        following modules:</p>
      <list type="bulleted">
        <item><c><![CDATA[erl_connect]]></c></item>
        <item><c><![CDATA[erl_eterm]]></c>.</item>
      </list>
    </section>
  </section>

  <section>
    <title>Remote Procedure Calls</title>
    <p>An Erlang node acting as a client to another Erlang node
      typically sends a request and waits for a reply. Such a request is
      included in a function call at a remote node and is called a remote
      procedure call. The following example shows how the
      Erl_Interface library supports remote procedure calls:</p>
    <code type="none"><![CDATA[

char modname[]=THE_MODNAME;
ETERM *reply,*ep;
ep = erl_format("[~a,[]]", modname);
if (!(reply = erl_rpc(fd, "c", "c", ep)))
  erl_err_msg("<ERROR> when compiling file: %s.erl !\
", modname);
erl_free_term(ep);
ep = erl_format("{ok,_}");
if (!erl_match(ep, reply))
  erl_err_msg("<ERROR> compiler errors !\
");
erl_free_term(ep);
erl_free_term(reply);    ]]></code>
    <p><c><![CDATA[c:c/1]]></c> is called to compile the specified module on the
      remote node. <c><![CDATA[erl_match()]]></c> checks that the compilation was
      successful by testing for the expected <c><![CDATA[ok]]></c>.</p>
    <p>Refer to the Reference Manual, the <c><![CDATA[erl_connect]]></c> module for
      more information about <c><![CDATA[erl_rpc()]]></c>, and its companions
      <c><![CDATA[erl_rpc_to()]]></c> and <c><![CDATA[erl_rpc_from()]]></c>.</p>
  </section>

  <section>
    <title>Using Global Names</title>
    <p>A C node has access to names registered through the Erlang Global
      module. Names can be looked up, allowing the C node to send messages
      to named Erlang services. C nodes can also register global names,
      allowing them to provide named services to Erlang processes or other C
      nodes. </p>
    <p>Erl_Interface does not provide a native implementation of the global
      service. Instead it uses the global services provided by a "nearby"
      Erlang node. In order to use the services described in this section,
      it is necessary to first open a connection to an Erlang node.</p>
    <p>To see what names there are:</p>
    <code type="none"><![CDATA[
char **names;
int count;
int i;

names = erl_global_names(fd,&count);

if (names) 
  for (i=0; i<count; i++) 
    printf("%s\
",names[i]);

free(names);    ]]></code>
    <p><c><![CDATA[erl_global_names()]]></c> allocates and returns a buffer containing
      all the names known to global. <c><![CDATA[count]]></c> will be initialized to
      indicate how many names are in the array. The array of strings in
      names is terminated by a NULL pointer, so it is not necessary to use
      <c><![CDATA[count]]></c> to determine when the last name is reached.</p>
    <p>It is the caller's responsibility to free the array.
      <c><![CDATA[erl_global_names()]]></c> allocates the array and all of the strings
      using a single call to <c><![CDATA[malloc()]]></c>, so <c><![CDATA[free(names)]]></c> is all
      that is necessary.</p>
    <p>To look up one of the names:</p>
    <code type="none"><![CDATA[
ETERM *pid;
char node[256];

pid = erl_global_whereis(fd,"schedule",node);    ]]></code>
    <p>If <c><![CDATA["schedule"]]></c> is known to global, an Erlang pid is returned
      that can be used to send messages to the schedule service.
      Additionally, <c><![CDATA[node]]></c> will be initialized to contain the name of
      the node where the service is registered, so that you can make a
      connection to it by simply passing the variable to <c><![CDATA[erl_connect()]]></c>.</p>
    <p>Before registering a name, you should already have registered your
      port number with <c><![CDATA[epmd]]></c>. This is not strictly necessary, but if you
      neglect to do so, then other nodes wishing to communicate with your
      service will be unable to find or connect to your process.</p>
    <p>Create a pid that Erlang processes can use to communicate with your
      service:</p>
    <code type="none"><![CDATA[
ETERM *pid;

pid = erl_mk_pid(thisnode,14,0,0);
erl_global_register(fd,servicename,pid);    ]]></code>
    <p>After registering the name, you should use <c><![CDATA[erl_accept()]]></c> to wait for
      incoming connections.</p>
    <p>Do not forget to free <c><![CDATA[pid]]></c> later with <c><![CDATA[erl_free_term()]]></c>!</p>
    <p>To unregister a name:</p>
    <code type="none"><![CDATA[
erl_global_unregister(fd,servicename);    ]]></code>
  </section>

  <section>
    <title>The Registry</title>
    <p>This section describes the use of the registry, a simple mechanism
      for storing key-value pairs in a C-node, as well as backing them up or
      restoring them from a Mnesia table on an Erlang node. More detailed
      information about the individual API functions can be found in the
      Reference Manual.</p>
    <p>Keys are strings, i.e. 0-terminated arrays of characters, and values
      are arbitrary objects. Although integers and floating point numbers
      are treated specially by the registry, you can store strings or binary
      objects of any type as pointers.</p>
    <p>To start, you need to open a registry:</p>
    <code type="none"><![CDATA[
ei_reg *reg;

reg = ei_reg_open(45);    ]]></code>
    <p>The number 45 in the example indicates the approximate number of
      objects that you expect to store in the registry. Internally the
      registry uses hash tables with collision chaining, so there is no
      absolute upper limit on the number of objects that the registry can
      contain, but if performance or memory usage are important, then you
      should choose a number accordingly. The registry can be resized later.</p>
    <p>You can open as many registries as you like (if memory permits).</p>
    <p>Objects are stored and retrieved through set and get functions. In
      the following examples you see how to store integers, floats, strings
      and arbitrary binary objects:</p>
    <code type="none"><![CDATA[
struct bonk *b = malloc(sizeof(*b));
char *name = malloc(7);

ei_reg_setival(reg,"age",29); 
ei_reg_setfval(reg,"height",1.85);

strcpy(name,"Martin");
ei_reg_setsval(reg,"name",name); 

b->l = 42;
b->m = 12;
ei_reg_setpval(reg,"jox",b,sizeof(*b));    ]]></code>
    <p>If you attempt to store an object in the registry and there is an
      existing object with the same key, the new value will replace the old
      one. This is done regardless of whether the new object and the old one
      have the same type, so you can, for example, replace a string with an
      integer. If the existing value is a string or binary, it will be freed
      before the new value is assigned.</p>
    <p>Stored values are retrieved from the registry as follows:</p>
    <code type="none"><![CDATA[
long i;
double f;
char *s;
struct bonk *b;
int size;

i = ei_reg_getival(reg,"age");
f = ei_reg_getfval(reg,"height");
s = ei_reg_getsval(reg,"name");
b = ei_reg_getpval(reg,"jox",&size);    ]]></code>
    <p>In all of the above examples, the object must exist and it must be of
      the right type for the specified operation. If you do not know the
      type of a given object, you can ask:</p>
    <code type="none"><![CDATA[
struct ei_reg_stat buf;

ei_reg_stat(reg,"name",&buf);    ]]></code>
    <p>Buf will be initialized to contain object attributes.</p>
    <p>Objects can be removed from the registry:</p>
    <code type="none"><![CDATA[
ei_reg_delete(reg,"name");    ]]></code>
    <p>When you are finished with a registry, close it to remove all the
      objects and free the memory back to the system:</p>
    <code type="none"><![CDATA[
ei_reg_close(reg);    ]]></code>

    <section>
      <title>Backing Up the Registry to Mnesia</title>
      <p>The contents of a registry can be backed up to Mnesia on a "nearby"
        Erlang node. You need to provide an open connection to the Erlang node
        (see <c><![CDATA[erl_connect()]]></c>). Also, Mnesia 3.0 or later must be running
        on the Erlang node before the backup is initiated:</p>
      <code type="none"><![CDATA[
ei_reg_dump(fd, reg, "mtab", dumpflags);      ]]></code>
      <p>The example above will backup the contents of the registry to the
        specified Mnesia table <c><![CDATA["mtab"]]></c>. Once a registry has been backed
        up to Mnesia in this manner, additional backups will only affect
        objects that have been modified since the most recent backup, i.e.
        objects that have been created, changed or deleted. The backup
        operation is done as a single atomic transaction, so that the entire
        backup will be performed or none of it will.</p>
      <p>In the same manner, a registry can be restored from a Mnesia table:</p>
      <code type="none"><![CDATA[
ei_reg_restore(fd, reg, "mtab");      ]]></code>
      <p>This will read the entire contents of <c><![CDATA["mtab"]]></c> into the specified
        registry. After the restore, all of the objects in the registry will
        be marked as unmodified, so a subsequent backup will only affect
        objects that you have modified since the restore.</p>
      <p>Note that if you restore to a non-empty registry, objects in the
        table will overwrite objects in the registry with the same keys. Also,
        the <em>entire</em> contents of the registry is marked as unmodified
        after the restore, including any modified objects that were not
        overwritten by the restore operation. This may not be your intention.</p>
    </section>

    <section>
      <title>Storing Strings and Binaries</title>
      <p>When string or binary objects are stored in the registry it is
        important that a number of simple guidelines are followed. </p>
      <p>Most importantly, the object must have been created with a single call
        to <c><![CDATA[malloc()]]></c> (or similar), so that it can later be removed by a
        single call to <c><![CDATA[free()]]></c>. Objects will be freed by the registry
        when it is closed, or when you assign a new value to an object that
        previously contained a string or binary.</p>
      <p>You should also be aware that if you store binary objects that are
        context-dependent (e.g. containing pointers or open file descriptors),
        they will lose their meaning if they are backed up to a Mnesia table
        and subsequently restored in a different context.</p>
      <p>When you retrieve a stored string or binary value from the registry,
        the registry maintains a pointer to the object and you are passed a
        copy of that pointer. You should never free an object retrieved in
        this manner because when the registry later attempts to free it, a
        runtime error will occur that will likely cause the C-node to crash.</p>
      <p>You are free to modify the contents of an object retrieved this way.
        However when you do so, the registry will not be aware of the changes
        you make, possibly causing it to be missed the next time you make a
        Mnesia backup of the registry contents. This can be avoided if you
        mark the object as dirty after any such changes with
        <c><![CDATA[ei_reg_markdirty()]]></c>, or pass appropriate flags to
        <c><![CDATA[ei_reg_dump()]]></c>.</p>
    </section>
  </section>
</chapter>