[yaz-moved-to-github.git] / doc / book.xml

<?xml version="1.0" standalone="no"?>
<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.4//EN"
    "http://www.oasis-open.org/docbook/xml/4.4/docbookx.dtd"
[
     <!ENTITY % local SYSTEM "local.ent">
     %local;
     <!ENTITY % entities SYSTEM "entities.ent">
     %entities;
     <!ENTITY % idcommon SYSTEM "common/common.ent">
     %idcommon;
]>
<book>
 <bookinfo>
  <title>YAZ User&apos;s Guide and Reference</title>
  <authorgroup>
   <author><firstname>Sebastian</firstname><surname>Hammer</surname></author>
   <author><firstname>Adam</firstname><surname>Dickmeiss</surname></author>
   <author><firstname>Mike</firstname><surname>Taylor</surname></author>
   <author><firstname>Heikki</firstname><surname>Levanto</surname></author>
   <author><firstname>Dennis</firstname><surname>Schafroth</surname></author>
  </authorgroup>
  <releaseinfo>&version;</releaseinfo>
  <copyright>
   <year>&copyright-year;</year>
   <holder>Index Data</holder>
  </copyright>
  <abstract>
   <simpara>
    This document is the programmer's guide and reference to the &yaz;
    package version &version;. &yaz; is a compact toolkit that provides
    access to the Z39.50 and SRU/Solr protocols, as well as a set of
    higher-level tools for implementing the server and client
    roles, respectively.
    The documentation can be used on its own, or as a reference when
    looking at the example applications provided with the package.
   </simpara>
   <simpara>
    <inlinemediaobject>
     <imageobject>
      <imagedata fileref="common/id.png" format="PNG"/>
     </imageobject>
     <imageobject>
      <imagedata fileref="common/id.eps" format="EPS"/>
     </imageobject>
    </inlinemediaobject>
   </simpara></abstract>
 </bookinfo>
 <chapter id="introduction">
  <title>Introduction</title>
  <para>
   &yaz; is a C/C++ library for information retrieval applications
   using the Z39.50/SRU/Solr protocols for information retrieval.
  </para>
  <para>
   Properties of &yaz;:
   <itemizedlist>
    <listitem>
     <para>
      Complete
      <ulink url="&url.z39.50;">Z39.50</ulink> version 3 support.
      Amendments and Z39.50-2002 revision is supported.
    </para>
    </listitem>
    <listitem>
     <para>
      Supports
      <ulink url="&url.sru;">SRU GET/POST/SOAP</ulink>
      version 1.1, 1.2 and 2.0 (over HTTP and HTTPS).
    </para>
    </listitem>
    <listitem>
     <para>
      Includes BER encoders/decoders for the
      <ulink url="&url.ill;">ISO ILL</ulink>
      protocol.
    </para>
    </listitem>
    <listitem>
     <para>
      Supports
      <ulink url="&url.solr;">Solr</ulink> Web Service version 1.4.x
      (client side only)
     </para>
    </listitem>
    <listitem>
     <para>
      Supports the following transports: BER over TCP/IP
      (<ulink url="&url.ber.over.tcpip;">RFC1729</ulink>),
      BER over unix local socket, and
      <ulink url="&url.http.1.1;">HTTP 1.1</ulink>.
    </para>
    </listitem>
    <listitem>
     <para>
      Secure Socket Layer support using
      <ulink url="&url.gnutls;">GnuTLS</ulink>.
      If enabled, &yaz; uses HTTPS transport (for SOAP) or
      "Secure BER" (for Z39.50).
     </para>
    </listitem>
    <listitem>
     <para>
      Offers
      <ulink url="&url.zoom;">ZOOM</ulink> C API implementing
      Z39.50, SRU and Solr Web Service.
    </para>
    </listitem>
    <listitem>
     <para>
      The &yaz; library offers a set of useful utilities
      related to the protocols, such as MARC (ISO2709) parser,
      CCL (ISO8777) parser,
      <ulink url="&url.cql;">CQL</ulink>
      parser, memory management routines, character set conversion.
     </para>
    </listitem>
    <listitem>
     <para>
      Portable code. &yaz; compiles out-of-the box on most Unixes and
      on Windows using Microsoft Visual C++.
     </para>
    </listitem>
    <listitem>
     <para>
      Fast operation. The C based BER encoders/decoders as well
      as the server component of &yaz; is very fast.
    </para>
    </listitem>
    <listitem>
     <para>
      Liberal license that allows for commercial use of &yaz;.
     </para>
    </listitem>
   </itemizedlist>
  </para>

  <sect1 id="introduction.reading">
   <title>Reading this Manual</title>
   <para>
    Most implementors only need to read a fraction of the
    material in thie manual, so a quick walkthrough of the chapters
    is in order.
   </para>
   <itemizedlist>
    <listitem>
     <para>
      <xref linkend="installation"/> contains installation
      instructions for &yaz;. You don't need reading this
      if you expect to download &yaz; binaries.
      However, the chapter contains information about how
      to make <emphasis>your</emphasis> application link
      with &yaz;.
     </para>
    </listitem>
    <listitem>
     <para>
      <xref linkend="zoom"/> describes the ZOOM API of &yaz;.
      This is definitely worth a read if you wish to develop a Z39.50/SRU
      client.
     </para>
    </listitem>
    <listitem>
     <para>
      <xref linkend="server"/> describes the generic frontend server
      and explains how to develop server Z39.50/SRU applications for &yaz;.
      Obviously worth reading if you're to develop a server.
    </para>
    </listitem>
    <listitem>
     <para>
      <xref linkend="yaz-client"/> describes how to use the &yaz; Z39.50
      client. If you're developer and wish to test your server
      or a server from another party, you might find this chapter
      useful.
    </para>
    </listitem>
    <listitem>
     <para>
      <xref linkend="asn"/> documents the most commonly used Z39.50
      C data structures offered by the &yaz; API. Client
      developers using ZOOM and non-Z39.50 implementors may skip this.
     </para>
    </listitem>
    <listitem>
     <para>
      <xref linkend="soap"/> describes how SRU and SOAP is used
      in &yaz;. Only if you're developing SRU applications
      this section is a must.
     </para>
    </listitem>
    <listitem>
     <para>
      <xref linkend="tools"/> contains sections for the various
      tools offered by &yaz;. Scan through the material quickly
      and see what's relevant to you! SRU implementors
      might find the <link linkend="cql">CQL</link> section
      particularly useful.
     </para>
    </listitem>
    <listitem>
     <para>
      <xref linkend="odr"/> goes through the details of the
      ODR module which is the work horse that encodes and decodes
      BER packages. Implementors using ZOOM only, do <emphasis>not</emphasis>
      need reading this.
      Most other Z39.50 implementors only need to read the first two
      sections (<xref linkend="odr.introduction"/> and
      <xref linkend="odr.use"/>).
     </para>
    </listitem>
    <listitem>
     <para>
      <xref linkend="comstack"/> describes the network layer module
      COMSTACK. Implementors using ZOOM or the generic frontend server
      may skip this. Others, presumably, handling client/server
     communication on their own should read this.
     </para>
    </listitem>
   </itemizedlist>
  </sect1>
  <sect1 id="introduction.api">
   <title>The API</title>
   <para>
    The <ulink url="&url.yaz;">&yaz;</ulink>
    toolkit offers several different levels of access to the
    <ulink url="&url.z39.50;">ISO23950/Z39.50</ulink>,
    <ulink url="&url.ill;">ILL</ulink> and
    <ulink url="&url.sru;">SRU</ulink>
    protocols.
    The level that you need to use depends on your requirements, and
    the role (server or client) that you want to implement.
    If you're developing a client application you should consider the
    <link linkend="zoom">ZOOM</link> API.
    It is, by far, the easiest way to develop clients in C.
    Server implementers should consider the
    <link linkend="server">generic frontend server</link>.
    None of those high-level APIs support the whole protocol, but
    they do include most facilities used in existing Z39.50 applications.
   </para>
   <para>
    If you're using 'exotic' functionality (meaning anything not included in
    the high-level APIs), developing non-standard extensions to Z39.50 or
    you're going to develop an ILL application you'll have to learn the lower
    level APIs of &yaz;.
   </para>
   <para>
    The YAZ toolkit modules is shown in figure <xref linkend="yaz.layer"/>.
   </para>
   <figure id="yaz.layer">
    <title>YAZ layers</title>
    <mediaobject>
     <imageobject>
      <imagedata fileref="apilayer.png" format="PNG"/>
     </imageobject>
     <imageobject>
      <imagedata fileref="apilayer.eps" format="EPS"/>
     </imageobject>
    </mediaobject>
   </figure>
   <para>
    There are four layers.
    <itemizedlist>
     <listitem>
      <para>A client or server application (or both).
       This layer includes ZOOM and the generic frontend server.
      </para>
     </listitem>
     <listitem>
      <para>
       The second layer provides a C represenation of the
       protocol units (packages) for Z39.50 ASN.1, ILL ASN.1,
       SRU.
      </para>
     </listitem>
     <listitem>
      <para>
       The third layer encodes and decodes protocol data units to
       simple packages (buffer with certain length). The &odr; module
       encodes and decodes BER whereas the HTTP modules encodes and
       decodes HTTP ruquests/responses.
      </para>
     </listitem>
     <listitem>
      <para>
       The lowest layer is &comstack; which exchanges the encoded packages
       with a peer process over a network.
      </para>
     </listitem>
    </itemizedlist>
   </para>
   <para>
    The &asn; module represents the ASN.1 definition of
    the Z39.50 protocol. It establishes a set of type and
    structure definitions, with one structure for each of the top-level
    PDUs, and one structure or type for each of the contained ASN.1 types.
    For primitive types, or other types that are defined by the ASN.1
    standard itself (such as the EXTERNAL type), the C representation is
    provided by the &odr; (Open Data Representation) subsystem.
  </para>
   <para>
     &odr; is a basic mechanism for representing an
    ASN.1 type in the C programming language, and for implementing BER
    encoders and decoders for values of that type. The types defined in
    the &asn; module generally have the prefix <literal>Z_</literal>, and
    a suffix corresponding to the name of the type in the ASN.1
    specification of the protocol (generally Z39.50-1995). In the case of
    base types (those originating in the ASN.1 standard itself), the prefix
    <literal>Odr_</literal> is sometimes seen. Either way, look for
    the actual definition in either <filename>z-core.h</filename> (for the types
    from the protocol), <filename>odr.h</filename> (for the primitive ASN.1
    types).
    The &asn; library also provides functions (which are, in turn,
    defined using &odr; primitives) for encoding and decoding data values.
    Their general form is
    <funcsynopsis>
     <funcprototype><funcdef>int <function>z_<replaceable>xxx</replaceable></function></funcdef>
      <paramdef>ODR <parameter>o</parameter></paramdef>
      <paramdef>Z_<replaceable>xxx</replaceable> **<parameter>p</parameter></paramdef>
      <paramdef>int <parameter>optional</parameter></paramdef>
      <paramdef>const char *<parameter>name</parameter></paramdef>
     </funcprototype>
    </funcsynopsis>
    (note the lower-case &quot;z&quot; in the function name)
   </para>
   <note>
    <para>
     If you are using the premade definitions of the &asn; module, and you
     are not adding new protocol of your own, the only parts of &odr; that you
     need to worry about are documented in
     <xref linkend="odr.use"/>.
    </para>
   </note>
   <para>
    When you have created a BER-encoded buffer, you can use the &comstack;
    subsystem to transmit (or receive) data over the network. The &comstack;
    module provides simple functions for establishing a connection
    (passively or actively, depending on the role of your application),
    and for exchanging BER-encoded PDUs over that connection. When you
    create a connection endpoint, you need to specify what transport to
    use (TCP/IP, SSL or UNIX sockets).
    For the remainder of the connection's lifetime, you don't have
    to worry about the underlying transport protocol at all - the &comstack;
    will ensure that the correct mechanism is used.
   </para>
   <para>
    We call the combined interfaces to &odr;, &asn;, and &comstack; the service
    level API. It's the API that most closely models the Z39.50
    service/protocol definition, and it provides unlimited access to all
    fields and facilities of the protocol definitions.
   </para>
   <para>
    The reason that the &yaz; service-level API is a conglomerate of the
    APIs from three different submodules is twofold. First, we wanted to allow
    the user a choice of different options for each major task. For instance,
    if you don't like the protocol API provided by &odr;/&asn;, you
    can use SNACC or BERUtils instead, and still have the benefits of the
    transparent transport approach of the &comstack; module. Secondly,
    we realize that you may have to fit the toolkit into an existing
    event-processing structure, in a way that is incompatible with
    the &comstack; interface or some other part of &yaz;.
   </para>
  </sect1>
 </chapter>
 <chapter id="installation">
  <title>Compilation and Installation</title>
  <sect1 id="installation-introduction">
   <title>Introduction</title>
   <para>
    The latest version of the software will generally be found at:
   </para>
   <para>
    <ulink url="&url.yaz.download;"/>
   </para>
   <para>
    We have tried our best to keep the software portable, and on many
    platforms, you should be able to compile everything with little or
    no changes.
   </para>
   <para>
    The software is regularly tested on
    <ulink url="&url.debian;">Debian GNU/Linux</ulink>,
    <ulink url="&url.centos;">CentOS</ulink>,
    <ulink url="&url.ubuntu;">Ubuntu Linux</ulink>,
    <ulink url="&url.freebsd;">FreeBSD (i386)</ulink>,
    <ulink url="&url.macosx;">MAC OSX</ulink>,
    <ulink url="&url.solaris;">Solaris</ulink>,
    Windows 7, Windows XP.
   </para>
   <para>
    Some versions have be known to work on HP/UX,
    DEC Unix, <ulink url="&url.netbsd;">NetBSD</ulink>,
    <ulink url="&url.openbsd;">OpenBSD</ulink>,
    IBM AIX,
    Data General DG/UX (with some CFLAGS tinkering),
    SGI/IRIX, DDE Supermax, Apple Macintosh (using the Codewarrior programming
    environment and the GUSI socket libraries),
    IBM AS/400 .
   </para>
   <para>
    If you move the software to other platforms, we'd be grateful if you'd
    let us know about it. If you run into difficulties, we will try to help
    if we can, and if you solve the problems, we would be happy to include
    your fixes in the next release. So far, we have mostly avoided
    <literal>#ifdefs</literal> for individual platforms, and we'd
    like to keep it that way as far as it makes sense.
   </para>
   <para>
    We maintain a mailing-list for the purpose of announcing new releases and
    bug-fixes, as well as general discussion. Subscribe by
    filling-in the form
    <ulink url="&url.yaz.mailinglist;">here</ulink>.
    General questions and problems can be directed at
    <ulink url="&url.yaz.mail;"/>, or the address given at the top of
     this document.
   </para>
  </sect1>
  <sect1 id="installation.unix"><title>UNIX</title>
   <para>
    We provide
    <ulink url="&url.debian;">Debian GNU/Linux</ulink> (i386 and amd64),
    <ulink url="&url.ubuntu;">Ubuntu</ulink> (i386 and amd64)
    and
    <ulink url="&url.centos;">CentOS</ulink> (amd64 only) packages for &yaz;.
    You should be able to create packages for other CPUs by building
    them from the source package.
   </para>
   <para>
    YAZ is also part of several packages repositories. Some of them are
   </para>
   <itemizedlist>
    <listitem>
     <para>
      Solaris CSW: <ulink url="http://www.opencsw.org/packages/yaz/"/>
     </para>
    </listitem>
    <listitem>
     <para>
      Solaris: <ulink url="http://unixpackages.com"/>
     </para>
    </listitem>
    <listitem>
     <para>
      FreeBSD: <ulink url="http://www.freshports.org/net/yaz"/>
     </para>
    </listitem>
    <listitem>
     <para>
      Debian: <ulink url="http://packages.debian.org/search?keywords=yaz"/>
     </para>
    </listitem>
    <listitem>
     <para>
      Ubuntu: <ulink url="https://launchpad.net/ubuntu/+source/yaz"/>
     </para>
    </listitem>
    <listitem>
     <para>
      NetBSD:
      <ulink url="http://ftp.netbsd.org/pub/pkgsrc/current/pkgsrc/net/yaz/README.html"/>
     </para>
    </listitem>
   </itemizedlist>
   <sect2 id="installation.source.unix">
    <title>Compiling from source on Unix</title>
    <para>
     Note that if your system doesn't have a native ANSI C compiler, you may
     have to acquire one separately. We recommend
     <ulink url="&url.gcc;">GCC</ulink>.
    </para>
    <para>
     If you wish to use character set conversion facilities in &yaz; or if you
     are compiling &yaz; for use with Zebra it is a good idea to ensure that
     the iconv library is installed. Some Unixes today already have it
     - if not, we suggest
     <ulink url="&url.libiconv;">GNU libiconv</ulink>.
    </para>
    <para>
     YAZ 3.0.16 and later includes a wrapper for the
     <ulink url="&url.icu;">ICU</ulink>
     (International Components for Unicode).
     In order to use this, the developer version of the ICU library
     must be available. ICU support is recommended for applications
     such as Pazpar2 and Zebra.
    </para>
    <para>
     The <ulink url="&url.libxslt;">libxslt</ulink>,
     <ulink url="&url.libxml2;">libxml2</ulink> librararies are required
     if &yaz; is to support SRU/Solr.
     These libraries are very portable and should compile out-of-the
     box on virtually all Unix platforms. It is available in binary
     forms for Linux and others.
    </para>
    <para>
     The GNU tools
     <ulink url="&url.autoconf;">Autoconf</ulink>,
     <ulink url="&url.automake;">Automake</ulink> and
     <ulink url="&url.libtool;">Libtool</ulink>
     are used to generate Makefiles and configure &yaz; for the system.
     You do <emphasis>not</emphasis> these tools unless you're using the
     Git version of &yaz;.
    </para>
    <para>
     The CQL parser for &yaz; is built using
     GNU <ulink url="&url.bison;">Bison</ulink>.
     This tool is only needed if you're using the Git version of &yaz;.
    </para>
    <para>
     &yaz; includes a tiny ASN.1 compiler. This compiler is
     written in <ulink url="&url.tcl;">Tcl</ulink>.
     But as for Bison you do not need it unless you're using Git
     version of &yaz; or you're using the compiler to built own codecs
     for private ASN.1.
    </para>
    <para>
     Generally it should be sufficient to run configure without options,
     like this:
    </para>
    <screen>
     ./configure
    </screen>
    <para>
     The configure script attempts to use use the C compiler specified by
     the <literal>CC</literal> environment variable. If not set, GNU C will be
     used if it is available. The <literal>CFLAGS</literal> environment
     variable holds options to be passed to the C compiler. If you're using
     Bourne-compatible shell you may pass something like this to use a
     particular C compiler with optimization enabled:
    </para>
    <screen>
     CC=/opt/ccs/bin/cc CFLAGS=-O ./configure
    </screen>
    <para>
     To customize &yaz;, the configure script also accepts a set of options.
     The most important are:
     <variablelist>
      <varlistentry>
       <term>
        <literal>--prefix</literal>=<replaceable>prefix</replaceable>
       </term>
       <listitem>
        <para>Specifies installation prefix for &yaz;. This is
        only needed if you run <literal>make install</literal> later to
        perform a "system" installation. The prefix is
        <literal>/usr/local</literal> if not specified.
        </para>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term>
        <literal>--enable-tcpd</literal>
       </term>
       <listitem>
        <para>The front end server will be built using Wietse's
        <ulink url="&url.tcpwrapper;">TCP wrapper library</ulink>.
        It allows you to allow/deny clients depending on IP number.
        The TCP wrapper library is often used in GNU/Linux and
        BSD distributions.
        See
        <citerefentry>
         <refentrytitle>hosts_access</refentrytitle>
         <manvolnum>5</manvolnum>
        </citerefentry>
        and
        <citerefentry>
         <refentrytitle>tcpd</refentrytitle>
         <manvolnum>8</manvolnum>
         </citerefentry>.
        </para>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term>
        <literal>--enable-threads</literal>
       </term>
       <listitem>
        <para>&yaz; will be built using POSIX threads.
        Specifically, <constant>_REENTRANT</constant> will be defined during
        compilation.
        </para>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term>
        <literal>--disable-shared</literal>
       </term>
       <listitem>
        <para>The make process will not create shared
        libraries (also known as shared objects <filename>.so</filename>).
        By default, shared libraries are created -
        equivalent to <literal>--enable-shared</literal>.
       </para>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term>
        <literal>--disable-shared</literal>
       </term>
       <listitem>
        <para>The make process will not create
        static libraries (<filename>.a</filename>).
        By default, static libraries are created -
        equivalent to <literal>--enable-static</literal>.
        </para>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term>
        <literal>--with-iconv</literal>[=<replaceable>prefix</replaceable>]
       </term>
       <listitem>
        <para>Compile &yaz; with iconv library in directory
        <replaceable>prefix</replaceable>. By default configure will
        search for iconv on the system. Use this option if it
        doesn't find iconv. Alternatively,
        <literal>--without-iconv</literal>, can be uset to force &yaz;
        not to use iconv.
        </para>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term>
        <literal>--with-xslt</literal>[=<replaceable>prefix</replaceable>]
       </term>
       <listitem>
        <para>Compile &yaz; with
        <ulink url="&url.libxslt;">libxslt</ulink> in directory
        <replaceable>prefix</replaceable>.
        Use this option if you want XSLT and XML support.
        By default, configure will
        search for libxslt on the system. Use this option if it
        libxslt is not found automatically. Alternatively,
        <literal>--without-xslt</literal>, can be used to force &yaz;
        not to use libxslt.
        </para>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term>
        <literal>--with-xml2</literal>[=<replaceable>prefix</replaceable>]
       </term>
       <listitem>
        <para>Compile &yaz; with
        <ulink url="&url.libxml2;">libxml2</ulink> in directory
        <replaceable>prefix</replaceable>.
        Use this option if you want &yaz; to use XML and support SRU/Solr.
        By default, configure will
        search for libxml2 on the system. Use this option if it
        libxml2 is not found automatically. Alternatively,
        <literal>--without-xml2</literal>, can be used to force &yaz;
        not to use libxml2.
        </para>
        <para>
         Note that option <literal>--with-xslt</literal>
         also enables libxml2.
        </para>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term>
        <literal>--with-gnutls</literal>[=<replaceable>prefix</replaceable>]
       </term>
       <listitem>
        <para>&yaz; will be linked with the GNU TLS libraries and
        an SSL COMSTACK will be provided. By default configure enables
        SSL support for YAZ if the GNU TLS development libraries are found
        on the system.
        </para>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term>
        <literal>--with-icu</literal>[=<replaceable>prefix</replaceable>]
       </term>
       <listitem>
        <para>&yaz; will be linked the
        <ulink url="&url.icu;">ICU</ulink> library in the prefix if given.
        If prefix is not given, the libraries exposed by the script
        <application>icu-config</application> will be used if found.
        </para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>
        <literal>--with-libgcrypt</literal>[=<replaceable>prefix</replaceable>]
       </term>
       <listitem>
        <para>&yaz; will be linked with
        <ulink url="&url.libgcrypt;">Libgcrypt</ulink> in the prefix if given.
        If prefix is not given, the libraries exposed by the script
        <application>libgcrypt-config</application> will be used if found.
        </para>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term>
        <literal>--with-memcached</literal>
       </term>
       <listitem>
        <para>&yaz; will be linked with
        <ulink url="&url.libmemcached;">libMemcached</ulink> to allow
        for result-set caching for ZOOM.
        The prefix can not be given. Note that YAZ will only search
        for libMemcached if Libgcrypt is also enabled.
        Note that 0.40 of libmemcached is required.
       </para>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term>
        <literal>--with-redis</literal>
       </term>
       <listitem>
        <para>&yaz; will be linked with the hiredis C library
        to allow for result-set caching for ZOOM on a
        <ulink url="&url.redis;">redis</ulink> server.
        The prefix can not be given. Note that YAZ will only search
        for hiredis if Libgcrypt is also enabled.
       </para>
       </listitem>
      </varlistentry>

     </variablelist>
    </para>
    <para>
     When configured, build the software by typing:
     <screen>
      make
     </screen>
    </para>
    <para>
     The following files are generated by the make process:
     <variablelist>
      <varlistentry>
       <term><filename>src/libyaz.la</filename></term>
       <listitem><para>
        Main &yaz; library. This is no ordinary library. It's
        a Libtool archive.
        By default, &yaz; creates a static library in
        <filename>lib/.libs/libyaz.a</filename>.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>src/libyaz_server.la</filename></term>
       <listitem><para>
         Generic Frontend server. This is an add-on for libyaz.la.
         Code in this library uses POSIX threads functions - if POSIX
         threads are available on the platform.
        </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>src/libyaz_icu.la</filename></term>
       <listitem><para>
        Functions that wrap the ICU library.
        </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>ztest/yaz-ztest</filename></term>
       <listitem><para>Test Z39.50 server.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>client/yaz-client</filename></term>
       <listitem><para>Z39.50 client for testing the protocol.
       See chapter <link linkend="yaz-client">
       YAZ client</link> for more information.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>util/yaz-config</filename></term>
       <listitem><para>A Bourne-shell script, generated by configure, that
       specifies how external applications should compile - and link with
       &yaz;.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>util/yaz-asncomp</filename></term>
       <listitem><para>The ASN.1 compiler for &yaz;. Requires the
       Tcl Shell, <application>tclsh</application>, in
       <literal>PATH</literal> to operate.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>util/yaz-iconv</filename></term>
       <listitem><para>This program converts data in one character set to
       another. This command exercises the YAZ character set
       conversion API.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>util/yaz-marcdump</filename></term>
       <listitem><para>This program parses ISO2709 encoded MARC records
       and prints them in line-format or XML.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>util/yaz-icu</filename></term>
       <listitem><para>This program exposes the ICU wrapper library if that
       is enabled for YAZ. Only if ICU is available this program is
       useful.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>util/yaz-url</filename></term>
       <listitem><para>This program is a simple HTTP page fetcher ala
       wget or curl.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>zoom/zoomsh</filename></term>
       <listitem><para>
        A simple shell implemented on top of the
        <link linkend="zoom">ZOOM</link> functions.
        The shell is a command line application that allows you to enter
        simple commands to perform ZOOM operations.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>zoom/zoomtst1</filename>,
       <filename>zoom/zoomtst2</filename>, ..</term>
       <listitem><para>
        Several small applications that demonstrates the ZOOM API.
       </para></listitem>
      </varlistentry>
     </variablelist>
    </para>
    <para>
     If you wish to install &yaz; in system directories
     <filename>/usr/local/bin</filename>,
     <filename>/usr/local/lib</filename> .. etc, you can type:
    </para>
    <screen>
     make install
    </screen>
    <para>
     You probably need to have root access in order to perform this.
     You must specify the <literal>--prefix</literal> option for configure if
     you wish to install &yaz; in other directories than the default
     <filename>/usr/local/</filename>.
    </para>
    <para>
     If you wish to perform an un-installation of &yaz;, use:
    </para>
    <screen>
     make uninstall
    </screen>
    <para>
     This will only work if you haven't reconfigured &yaz; (and therefore
     changed installation prefix). Note that uninstall will not
     remove directories created by make install, e.g.
     <filename>/usr/local/include/yaz</filename>.
    </para>
   </sect2>
   <sect2 id="installation-linking-yaz-unix">
    <title>How to make apps using YAZ on UNIX</title>
    <para>
     This section describes how to compile - and link your own
     applications using the &yaz; toolkit.
     If you're used to Makefiles this shouldn't be hard. As for
     other libraries you have used before, you have to set a proper include
     path for your C/C++ compiler and specify the location of
     &yaz; libraries. You can do it by hand, but generally we suggest
     you use the <filename>yaz-config</filename> that is generated
     by <filename>configure</filename>. This is especially
     important if you're using the threaded version of &yaz; which
     require you to pass more options to your linker/compiler.
    </para>
    <para>
     The <filename>yaz-config</filename> script accepts command line
     options that makes the <filename>yaz-config</filename> script print
     options that you should use in your make process.
     The most important ones are:
     <literal>--cflags</literal>, <literal>--libs</literal>
     which prints C compiler flags, and linker flags respectively.
    </para>
    <para>
     A small and complete <literal>Makefile</literal> for a C
     application consisting of one source file,
     <filename>myprog.c</filename>, may look like this:
     <screen>
      YAZCONFIG=/usr/local/bin/yaz-config
      CFLAGS=`$(YAZCONFIG) --cflags`
      LIBS=`$(YAZCONFIG) --libs`
      myprog: myprog.o
         $(CC) $(CFLAGS) -o myprog myprog.o $(LIBS)
      </screen>
    </para>
    <para>
     The CFLAGS variable consists of a C compiler directive that will set
     the include path to the <emphasis>parent</emphasis> directory
     of <filename>yaz</filename>. That is, if &yaz; header files were
     installed in <filename>/usr/local/include/yaz</filename>,
     then include path is set to <filename>/usr/local/include</filename>.
     Therefore, in your applications you should use
     <screen>
      #include &lt;yaz/proto.h>
     </screen>
     and <emphasis>not</emphasis>
     <screen>
      #include &lt;proto.h>
     </screen>
    </para>
    <para>
     For Libtool users, the <filename>yaz-config</filename> script provides
     a different variant of option <literal>--libs</literal>, called
     <literal>--lalibs</literal> that returns the name of the
     Libtool archive(s) for &yaz; rather than the ordinary ones.
    </para>
    <para>
     For applications using the threaded version of &yaz;,
     specify <literal>threads</literal> after the
     other options. When <literal>threads</literal> is given,
     more flags and linker flags will be printed by
     <filename>yaz-config</filename>. If our previous example was
      using threads, you'd have to modify the lines that set
     <literal>CFLAGS</literal> and <literal>LIBS</literal> as
     follows:
     <screen>
      CFLAGS=`$(YAZCONFIG) --cflags threads`
      LIBS=`$(YAZCONFIG) --libs threads`
     </screen>
     There is no need specify POSIX thread libraries in your Makefile.
     The <literal>LIBS</literal> variable includes that as well.
    </para>
   </sect2>
  </sect1>
  <sect1 id="installation.win32">
   <title>Windows</title>
   <para>The easiest way to install YAZ on Windows is by downloading
   an installer from
   <ulink url="&url.yaz.download.win32;">here</ulink>.
   The installer comes with source too - in case you wish to
   compile YAZ with different compiler options, etc.
   </para>

   <sect2 id="installation.win32.source">
    <title>Compiling from Source on Windows</title>
    <para>
     &yaz; is shipped with "makefiles" for the NMAKE tool that comes
     with <ulink url="&url.vstudio;">
     Microsoft Visual Studio</ulink>. It has been tested with
     Microsoft Visual Studio 2013.
    </para>
    <para>
     Start a command prompt and switch the sub directory
     <filename>WIN</filename> where the file <filename>makefile</filename>
     is located. Customize the installation by editing the
     <filename>makefile</filename> file (for example by using notepad).
     The following summarizes the most important settings in that file:
     <variablelist>
      <varlistentry>
       <term><literal>DEBUG</literal></term>
       <listitem><para>
        If set to 1, the software is
        compiled with debugging libraries (code generation is
        multi-threaded debug DLL).
        If set to 0, the software is compiled with release libraries
        (code generation is multi-threaded DLL).
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><literal>HAVE_TCL</literal>, <literal>TCL</literal></term>
       <listitem><para>
        If <literal>HAVE_TCL</literal> is set to 1, nmake will
        use the ASN.1 compiler (<ulink url="&url.tcl;">Tcl</ulink> based).
        You must set <literal>TCL</literal> to the full path of the Tcl
        interpreter. A Windows version of Tcl is part of
        <ulink url="&url.gitwindows;">Git for Windows</ulink>.
       </para>
       <para>
        If you do not have Tcl installed, set
        <literal>HAVE_TCL</literal> to 0.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><literal>HAVE_BISON</literal>,
       <literal>BISON</literal></term>
       <listitem><para>
        If GNU Bison is present, you might set <literal>HAVE_BISON</literal>
        to 1 and specify the Bison executable in <literal>BISON</literal>.
        Bison is only required if you use the Git version of
        YAZ or if you modify the grammar for CQL
        (<filename>cql.y</filename>).
       </para>
       <para>
        A Windows version of GNU Bison is part of
        <ulink url="&url.gitwindows;">Git for Windows</ulink>.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><literal>HAVE_ICONV</literal>,
       <literal>ICONV_DIR</literal></term>
       <listitem><para>
        If <literal>HAVE_ICONV</literal> is set to 1, YAZ is compiled
        with iconv support. In this configuration, set
        <literal>ICONV_DIR</literal> to the iconv source directory.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><literal>HAVE_LIBXML2</literal>,
       <literal>LIBXML2_DIR</literal></term>
       <listitem>
        <para>
         If <literal>HAVE_LIBXML2</literal> is set to 1, YAZ is compiled
         with SRU support. In this configuration, set
         <literal>LIBXML2_DIR</literal> to the
         <ulink url="&url.libxml2;">libxml2</ulink> source directory.
        </para>
        <para>
         You can get pre-compiled Libxml2+Libxslt DLLs and headers from
         <ulink url="&url.libxml2.download.windows;">here</ulink>.
         Should you with to compile those libraries yourself, refer to
         to <xref linkend="installation.windows.libxml2"/>
        </para>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term><literal>HAVE_LIBXSLT</literal>,
       <literal>LIBXSLT_DIR</literal></term>
       <listitem>
        <para>
         If <literal>HAVE_LIBXSLT</literal> is set to 1, YAZ is compiled
         with XSLT support. In this configuration, set
         <literal>LIBXSLT_DIR</literal> to the
         <ulink url="&url.libxslt;">libxslt</ulink> source directory.
        </para>
        <note>
         <para>
          libxslt depends libxml2.
         </para>
        </note>
       </listitem>
      </varlistentry>
      <varlistentry>
       <term><literal>HAVE_ICU</literal>,
       <literal>ICU_DIR</literal></term>
       <listitem>
        <para>
         If <literal>HAVE_ICU</literal> is set to 1, YAZ is compiled
         with <ulink url="&url.icu;">ICU</ulink> support.
         In this configuration, set
         <literal>ICU_DIR</literal> to the
         <ulink url="&url.icu;">ICU</ulink> source directory.
        </para>
       </listitem>
      </varlistentry>
     </variablelist>
    </para>
    <para>
     When satisfied with the settings in the makefile, type
     <screen>
      nmake
     </screen>
    </para>
    <note>
     <para>
      If the <filename>nmake</filename> command is not found on your system
      you probably haven't defined the environment variables required to
      use that tool. To fix that, find and run the batch file
      <filename>vcvars32.bat</filename>. You need to run it from within
      the command prompt or set the environment variables "globally";
      otherwise it doesn't work.
     </para>
    </note>
    <para>
     If you wish to recompile &yaz; - for example if you modify
     settings in the <filename>makefile</filename> you can delete
     object files, etc by running.
     <screen>
      nmake clean
     </screen>
    </para>
    <para>
     The following files are generated upon successful compilation:
     <variablelist>
      <varlistentry>
       <term><filename>bin/yaz&soversion;.dll</filename> /
       <filename>bin/yaz&soversion;d.dll</filename></term>
       <listitem><para>
        &yaz; Release/Debug DLL.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>lib/yaz&soversion;.lib</filename> /
       <filename>lib/yaz&soversion;d.lib</filename></term>
       <listitem><para>
        Import library for <filename>yaz&soversion;.dll</filename> /
        <filename>yaz&soversion;d.dll</filename>.
      </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>bin/yaz_cond&soversion;.dll</filename> /
       <filename>bin/yaz_cond&soversion;d.dll</filename></term>
       <listitem><para>
        Release/Debug DLL for condition variable utilities (condvar.c).
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>lib/yaz_cond&soversion;.lib</filename> /
       <filename>lib/yaz_cond&soversion;d.lib</filename></term>
       <listitem><para>
        Import library for <filename>yaz_cond&soversion;.dll</filename> /
        <filename>yaz_cond&soversion;d.dll</filename>.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>bin/yaz_icu&soversion;.dll</filename> /
       <filename>bin/yaz_icu&soversion;d.dll</filename></term>
       <listitem><para>
        Release/Debug DLL for the ICU wrapper utility.
        Only build if HAVE_ICU is 1.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>lib/yaz_icu&soversion;.lib</filename> /
       <filename>lib/yaz_icu&soversion;d.lib</filename></term>
       <listitem><para>
        Import library for <filename>yaz_icu&soversion;.dll</filename> /
        <filename>yaz_icu&soversion;d.dll</filename>.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>bin/yaz-ztest.exe</filename></term>
       <listitem><para>
        Z39.50 multi-threaded test/example server. It's a WIN32
        console application.
      </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>bin/yaz-client.exe</filename></term>
       <listitem><para>
        &yaz; Z39.50 client application. It's a WIN32 console application.
        See chapter <link linkend="yaz-client">YAZ client</link> for more
        information.
      </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>bin/yaz-icu.exe</filename></term>
       <listitem><para>This program exposes the ICU wrapper library if that
       is enabled for YAZ. Only if ICU is available this program is
       build.
      </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>bin/zoomsh.exe</filename></term>
       <listitem><para>
        Simple console application implemented on top of the
        <link linkend="zoom">ZOOM</link> functions.
        The application is a command line shell that allows you to enter
        simple commands to perform ZOOM operations.
      </para></listitem>
      </varlistentry>
      <varlistentry>
       <term><filename>bin/zoomtst1.exe</filename>,
       <filename>bin/zoomtst2.exe</filename>, ..</term>
       <listitem><para>
        Several small applications that demonstrates the ZOOM API.
      </para></listitem>
      </varlistentry>
     </variablelist>
    </para>
   </sect2>

   <sect2 id="installation-linking-yaz-win32">
    <title>How to make apps using YAZ on Windows</title>
    <para>
     This section will go though the process of linking your Windows
     applications with &yaz;.
    </para>
    <para>
     Some people are confused by the fact that we use the nmake
     tool to build &yaz;. They think they have to do that too - in order
     to make their Windows applications work with &yaz;. The good news is that
     you don't have to. You can use the integrated environment of
     Visual Studio if desired for your own application.
    </para>
    <para>
     When setting up a project or Makefile you have to set the following:
     <variablelist>
      <varlistentry>
       <term>include path</term>
       <listitem><para>
        Set it to the <filename>include</filename> directory of &yaz;.
        </para></listitem>
      </varlistentry>
      <varlistentry>
       <term>import library <filename>yaz&soversion;.lib</filename></term>
       <listitem><para>
        You must link with this library. It's located in the
        sub directory <filename>lib</filename> of &yaz;.
        If you want to link with the debug version of &yaz;, you must
        link against <filename>yaz&soversion;d.lib</filename> instead.
       </para></listitem>
      </varlistentry>
      <varlistentry>
       <term>dynamic link library
       <filename>yaz&soversion;.dll</filename>
       </term>
       <listitem><para>
        This DLL must be in your execution path when you invoke
        your application. Specifically, you should distribute this
        DLL with your application.
       </para></listitem>
      </varlistentry>
     </variablelist>
    </para>
   </sect2>

   <sect2 id="installation.windows.libxml2">
    <title>Compiling Libxml2 and Libxslt on windows</title>
    <para>
     Download libxml2 and Libxslt source and unpack it.
     In the example below we install  Libxml2 2.9.2 and Libxslt 1.1.28
     for 32-bit, so we  use the destination directories
     libxml2.2.9.2.win32 and libxslt-1.1.28.win32 to reflect both
     version and architecture.
     <screen>
      cd win32
      cscript configure.js prefix=c:\libxml2-2.9.2.win32 iconv=no
      nmake
      nmake install
     </screen>
    </para>
    <para>
     For Libxslt it is similar. We must ensure that compilation of
     Libxslt links against the already installed libxml2.
     <screen>
      cd win32
      cscript configure.js prefix=c:\libxslt-1.1.28.win32 iconv=no \
          lib=c:\libxmlt-2.9.2.win32\lib \
          include=c:\libxmlt-2.9.2.win32\include\libxml2
      nmake
      nmake install
     </screen>
    </para>
   </sect2>

  </sect1>
 </chapter>
 <!--
     ### Still to document:
     ZOOM_connection_errcode(c)
     ZOOM_connection_errmsg(c)
     ZOOM_connection_addinfo(c)
     ZOOM_connection_addinfo(c)
     ZOOM_connection_diagset(c);
     ZOOM_connection_save_apdu_wrbuf
     ZOOM_diag_str(error)
     ZOOM_resultset_record_immediate(s, pos)
     ZOOM_resultset_cache_reset(r)
     ZOOM_options_set_callback(opt, function, handle)
     ZOOM_options_create_with_parent2(parent1, parent2)
     ZOOM_options_getl(opt, name, len)
     ZOOM_options_setl(opt, name, value, len)
     ZOOM_options_get_bool(opt, name, defa)
     ZOOM_options_get_int(opt, name, defa)
     ZOOM_options_set_int(opt, name, value)
 -->
 <chapter id="zoom">
  <title>ZOOM</title>
  <para>
   &zoom; is an acronym for 'Z39.50 Object-Orientation Model' and is
   an initiative started by Mike Taylor (Mike is from the UK, which
   explains the peculiar name of the model). The goal of &zoom; is to
   provide a common Z39.50 client API not bound to a particular
   programming language or toolkit.
  </para>
  <para>
   From YAZ version 2.1.12, <ulink url="&url.sru;">SRU</ulink> is supported.
   You can make SRU ZOOM connections by specifying scheme
   <literal>http://</literal> for the hostname for a connection.
   The dialect of SRU used is specified by the value of the
   connection's <literal>sru</literal> option, which may be SRU over
   HTTP GET (<literal>get</literal>),
   SRU over HTTP POST (<literal>post</literal>), (SRU over
   SOAP) (<literal>soap</literal>) or <literal>solr</literal>
   (<ulink url="&url.solr;">Solr</ulink> Web Service).
   Using the facility for embedding options in target strings, a
   connection can be forced to use SRU rather the SRW (the default) by
   prefixing the target string with <literal>sru=get,</literal>, like this:
   <literal>sru=get,http://sru.miketaylor.org.uk:80/sru.pl</literal>
  </para>
  <para>
   <ulink url="&url.solr;">Solr</ulink>  protocol support was added to
   YAZ in version 4.1.0, as a dialect of a SRU protocol, since both are
   HTTP based protocols.
  </para>
  <para>
   The lack of a simple Z39.50 client API for &yaz; has become more
   and more apparent over time. So when the first &zoom; specification
   became available,
   an implementation for &yaz; was quickly developed. For the first time, it is
   now as easy (or easier!) to develop clients than servers with &yaz;. This
   chapter describes the &zoom; C binding. Before going further, please
   reconsider whether C is the right programming language for the job.
   There are other language bindings available for &yaz;, and still
   more
   are in active development. See the
   <ulink url="&url.zoom;">ZOOM web-site</ulink> for
   more information.
  </para>
  <para>
   In order to fully understand this chapter you should read and
   try the example programs <literal>zoomtst1.c</literal>,
   <literal>zoomtst2.c</literal>, .. in the <literal>zoom</literal>
   directory.
  </para>
  <para>
   The C language misses features found in object oriented languages
   such as C++, Java, etc. For example, you'll have to manually,
   destroy all objects you create, even though you may think of them as
   temporary. Most objects has a <literal>_create</literal> - and a
   <literal>_destroy</literal> variant.
   All objects are in fact pointers to internal stuff, but you don't see
   that because of typedefs. All destroy methods should gracefully ignore a
   <literal>NULL</literal> pointer.
  </para>
  <para>
   In each of the sections below you'll find a sub section called
   protocol behavior, that describes how the API maps to the Z39.50
   protocol.
  </para>
  <sect1 id="zoom-connections">
   <title>Connections</title>
   <para>The Connection object is a session with a target.
   </para>
   <synopsis>
    #include &lt;yaz/zoom.h>

    ZOOM_connection ZOOM_connection_new(const char *host, int portnum);

    ZOOM_connection ZOOM_connection_create(ZOOM_options options);

    void ZOOM_connection_connect(ZOOM_connection c, const char *host,
                                 int portnum);
    void ZOOM_connection_destroy(ZOOM_connection c);
   </synopsis>
   <para>
    Connection objects are created with either function
    <function>ZOOM_connection_new</function> or
    <function>ZOOM_connection_create</function>.
    The former creates and automatically attempts to establish a network
    connection with the target. The latter doesn't establish
    a connection immediately, thus allowing you to specify options
    before establishing network connection using the function
    <function>ZOOM_connection_connect</function>.
    If the port number, <literal>portnum</literal>, is zero, the
    <literal>host</literal> is consulted for a port specification.
    If no port is given, 210 is used. A colon denotes the beginning of
    a port number in the host string. If the host string includes a
    slash, the following part specifies a database for the connection.
   </para>
   <para>
    You can prefix the host with a scheme followed by colon. The
    default scheme is <literal>tcp</literal> (Z39.50 protocol).
    The scheme <literal>http</literal> selects SRU/get over HTTP by default,
    but can overridded to use SRU/post, SRW and the Solr protocol.
   </para>
   <para>
    You can prefix the scheme-qualified host-string with one or more
    comma-separated
    <literal><parameter>key</parameter>=<parameter>value</parameter></literal>
    sequences, each of which represents an option to be set into the
    connection structure <emphasis>before</emphasis> the
    protocol-level connection is forged and the initialization
    handshake takes place.  This facility can be used to provide
    authentication credentials, as in host-strings such as:
    <literal>user=admin,password=halfAm4n,tcp:localhost:8017/db</literal>
   </para>
   <para>
    Connection objects should be destroyed using the function
    <function>ZOOM_connection_destroy</function>.
   </para>
   <synopsis>
    void ZOOM_connection_option_set(ZOOM_connection c,
                                    const char *key, const char *val);

    void ZOOM_connection_option_setl(ZOOM_connection c,
                                     const char *key,
                                     const char *val, int len);

    const char *ZOOM_connection_option_get(ZOOM_connection c,
                                           const char *key);
    const char *ZOOM_connection_option_getl(ZOOM_connection c,
                                            const char *key,
                                            int *lenp);
   </synopsis>
   <para>
    The functions <function>ZOOM_connection_option_set</function> and
    <function>ZOOM_connection_option_setl</function> allows you to
    set an option given by <parameter>key</parameter> to the value
    <parameter>value</parameter> for the connection.
    For <function>ZOOM_connection_option_set</function>, the
    value is assumed to be a 0-terminated string. Function
    <function>ZOOM_connection_option_setl</function> specifies a
    value of a certain size (len).
   </para>
   <para>
    Functions <function>ZOOM_connection_option_get</function> and
    <function>ZOOM_connection_option_getl</function> returns
    the value for an option given by <parameter>key</parameter>.
   </para>
   <table id="zoom-connection-options" frame="top">
    <title>ZOOM Connection Options</title>
    <tgroup cols="3">
     <colspec colwidth="4*" colname="name"></colspec>
     <colspec colwidth="7*" colname="description"></colspec>
     <colspec colwidth="3*" colname="default"></colspec>
     <thead>
      <row>
       <entry>Option</entry>
       <entry>Description</entry>
       <entry>Default</entry>
      </row>
     </thead>
     <tbody>
      <row><entry>
       implementationName</entry><entry>Name of Your client
      </entry><entry>none</entry></row>
      <row><entry>
       user</entry><entry>Authentication user name
      </entry><entry>none</entry></row>
      <row><entry>
       group</entry><entry>Authentication group name
      </entry><entry>none</entry></row>
      <row><entry>
       password</entry><entry>Authentication password.
      </entry><entry>none</entry></row>
      <row><entry>
       authenticationMode</entry><entry>How authentication is encoded.
      </entry><entry>basic</entry></row>
      <row><entry>
       host</entry><entry>Target host. This setting is "read-only".
       It's automatically set internally when connecting to a target.
      </entry><entry>none</entry></row>
      <row><entry>
       proxy</entry><entry>Proxy host. If set, the logical host
       is encoded in the otherInfo area of the Z39.50 Init PDU
       with OID 1.2.840.10003.10.1000.81.1.
      </entry><entry>none</entry></row>
      <row><entry>
       clientIP</entry><entry>Client IP. If set, is
       encoded in the otherInfo area of a Z39.50 PDU with OID
       1.2.840.10003.10.1000.81.3. Holds the original IP addreses
       of a client. Is used of ZOOM is used in a gateway of some sort.
      </entry><entry>none</entry></row>
      <row><entry>
       async</entry><entry>If true (1) the connection operates in
       asynchronous operation which means that all calls are non-blocking
       except
       <link linkend="zoom.events"><function>ZOOM_event</function></link>.
      </entry><entry>0</entry></row>
      <row><entry>
       maximumRecordSize</entry><entry> Maximum size of single record.
      </entry><entry>1 MB</entry></row>
      <row><entry>
       preferredMessageSize</entry><entry> Maximum size of multiple records.
      </entry><entry>1 MB</entry></row>
      <row><entry>
       lang</entry><entry> Language for negotiation.
      </entry><entry>none</entry></row>
      <row><entry>
       charset</entry><entry> Character set for negotiation.
      </entry><entry>none</entry></row>
      <row><entry>
       serverImplementationId</entry><entry>
       Implementation ID of server.  (The old targetImplementationId
       option is also supported for the benefit of old applications.)
      </entry><entry>none</entry></row>
      <row><entry>
       targetImplementationName</entry><entry>
       Implementation Name of server.  (The old
       targetImplementationName option is also supported for the
       benefit of old applications.)
      </entry><entry>none</entry></row>
      <row><entry>
       serverImplementationVersion</entry><entry>
       Implementation Version of server.  (the old
       targetImplementationVersion option is also supported for the
       benefit of old applications.)
      </entry><entry>none</entry></row>
      <row><entry>
       databaseName</entry><entry>One or more database names
       separated by character plus (<literal>+</literal>), which to
       be used by subsequent search requests on this Connection.
      </entry><entry>Default</entry></row>
      <row><entry>
       piggyback</entry><entry>True (1) if piggyback should be
       used in searches; false (0) if not.
      </entry><entry>1</entry></row>
      <row><entry>
       smallSetUpperBound</entry><entry>If hits is less than or equal to this
       value, then target will return all records using small element set name
      </entry><entry>0</entry></row>
      <row><entry>
       largeSetLowerBound</entry><entry>If hits is greater than this
       value, the target will return no records.
      </entry><entry>1</entry></row>
      <row><entry>
       mediumSetPresentNumber</entry><entry>This value represents
       the number of records to be returned as part of a search when when
       hits is less than or equal to large set lower bound and if hits
       is greater than small set upper bound.
      </entry><entry>0</entry></row>
      <row><entry>
       smallSetElementSetName</entry><entry>
       The element set name to be used for small result sets.
      </entry><entry>none</entry></row>
      <row><entry>
       mediumSetElementSetName</entry><entry>
       The element set name to be for medium-sized result sets.
      </entry><entry>none</entry></row>
      <row><entry>
       init_opt_search, init_opt_present, init_opt_delSet, etc.</entry><entry>
       After a successful Init, these options may be interrogated to
       discover whether the server claims to support the specified
       operations.
      </entry><entry>none</entry></row>
      <row>
       <entry>sru</entry><entry>
       SRU/Solr transport type. Must be either <literal>soap</literal>,
       <literal>get</literal>, <literal>post</literal>, or
       <literal>solr</literal>.
      </entry><entry>soap</entry></row>
      <row><entry>
       sru_version</entry><entry>
       SRU/SRW version. Should be <literal>1.1</literal>, or
       <literal>1.2</literal>. This is , prior to connect, the version
       to offer (highest version). And following connect (in fact
       first operation), holds the negotiated version with the server
       (same or lower version).
      </entry><entry>1.2</entry></row>
      <row id="zoom.facets.option"><entry>
       facets</entry><entry>
       Requested or recommend facets may be given before a search is sent.
       The value of this setting is described in <xref linkend="facets"/>
       For inspection of the facets returned, refer to the functions
       described in <xref linkend="zoom.facets"/>.
      </entry><entry>none</entry></row>
      <row><entry>
       apdulog</entry><entry>
       If set to a true value such as "1", a log of low-level
       protocol packets is emitted on standard error stream.  This
       can be very useful for debugging.
      </entry><entry>0</entry></row>
      <row><entry>
       saveAPDU</entry><entry>
       If set to a true value such as "1", a log of low-level
       protocol packets is saved. The log can be retrieved by reading
       option APDU. Setting saveAPDU always has the side effect of
       resetting the currently saved log. This setting is
       <emphasis>write-only</emphasis>. If read, NULL will be returned.
       It is only recognized in
       <function>ZOOM_connection_option_set</function>.
      </entry><entry>0</entry></row>
      <row><entry>
       APDU</entry><entry>
       Returns the log of protocol packets. Will be empty if logging
       is not enabled (see saveAPDU above). This setting is
       <emphasis>read-only</emphasis>. It is only recognized if used
       in call to <function>ZOOM_connection_option_get</function> or
       <function>ZOOM_connection_option_getl</function>.
      </entry><entry></entry></row>
      <row><entry>
       memcached</entry><entry>
       If given and non-empty,
       <ulink url="&url.libmemcached;">libMemcached</ulink>
       will be configured for the connection.
       This option is inspected by ZOOM when a connection is established.
       If the <literal>memcached</literal> option is given
       and YAZ is compiled without libMemcached support, an internal
       diagnostic (10018) will be thrown.
       libMemcached support is available for YAZ 5.0.13 or later. If this
       option is supplied for an earlier version of YAZ, it is
       <emphasis>ignored</emphasis>.
       The value of this option is a list options - each is of the
       form <literal>--name=value</literal>.
       Option <literal>--server=</literal>host[:port] specifies a memcached
       server. It may be repeated for multiple memcached servers.
       Option <literal>--expire=</literal>seconds sets expiry time in seconds
       for how long result sets are to be cached.
      </entry><entry>none</entry></row>
      <row><entry>
       redis</entry><entry>
       If given and non-empty,
       a <ulink url="&url.redis;">redis</ulink> context will be created
       for the connection.
       This option is inspected by ZOOM when a connection is established.
       If the <literal>redis</literal> option is given
       and YAZ is compiled without redis support, an internal
       diagnostic (10018) will be thrown.
       redis support is available for YAZ 5.2.0 or later. If this
       option is supplied for an earlier version of YAZ, it is
       <emphasis>ignored</emphasis>.
       The value of this option is a set options, similar to that
       of the memcached setting. At this stage only --server=host[:port]
       and --expire=seconds is supported.
      </entry><entry>none</entry></row>
     </tbody>
    </tgroup>
   </table>
   <para>
    If either option <literal>lang</literal> or <literal>charset</literal>
    is set, then
    <ulink url="&url.z39.50.charneg;">
     Character Set and Language Negotiation</ulink> is in effect.
   </para>
   <synopsis>
     int ZOOM_connection_error(ZOOM_connection c, const char **cp,
                               const char **addinfo);
     int ZOOM_connection_error_x(ZOOM_connection c, const char **cp,
                                 const char **addinfo, const char **dset);
   </synopsis>
   <para>
    Function <function>ZOOM_connection_error</function> checks for
    errors for the last operation(s) performed. The function returns
    zero if no errors occurred; non-zero otherwise indicating the error.
    Pointers <parameter>cp</parameter> and <parameter>addinfo</parameter>
    holds messages for the error and additional-info if passed as
    non-<literal>NULL</literal>. Function
    <function>ZOOM_connection_error_x</function> is an extended version
    of <function>ZOOM_connection_error</function> that is capable of
    returning name of diagnostic set in <parameter>dset</parameter>.
   </para>
   <sect2 id="zoom-connection-z39.50">
    <title>Z39.50 Protocol behavior</title>
    <para>
     The calls <function>ZOOM_connection_new</function> and
     <function>ZOOM_connection_connect</function> establishes a TCP/IP
     connection and sends an Initialize Request to the target if
     possible. In addition, the calls waits for an Initialize Response
     from the target and the result is inspected (OK or rejected).
    </para>
    <para>
     If <literal>proxy</literal> is set then the client will establish
     a TCP/IP connection with the peer as specified by the
     <literal>proxy</literal> host and the hostname as part of the
     connect calls will be set as part of the Initialize Request.
     The proxy server will then "forward" the PDU's transparently
     to the target behind the proxy.
    </para>
    <para>
     For the authentication parameters, if option <literal>user</literal>
     is set and both options <literal>group</literal> and
     <literal>pass</literal> are unset, then Open style
     authentication is used (Version 2/3) in which case the username
     is usually followed by a slash, then by a password.
     If either <literal>group</literal>
     or <literal>pass</literal> is set then idPass authentication
     (Version 3 only) is used. If none of the options are set, no
     authentication parameters are set as part of the Initialize Request
     (obviously).
    </para>
    <para>
     When option <literal>async</literal> is 1, it really means that
     all network operations are postponed (and queued) until the
     function <literal>ZOOM_event</literal> is invoked. When doing so
     it doesn't make sense to check for errors after
     <literal>ZOOM_connection_new</literal> is called since that
     operation "connecting - and init" is still incomplete and the
     API cannot tell the outcome (yet).
    </para>
    </sect2>
   <sect2 id="zoom.sru.init.behavior">
    <title>SRU/Solr Protocol behavior</title>
    <para>
     The HTTP based protocols (SRU, SRW, Solr) doesn't feature an
     Inititialize Request, so  the connection phase merely establishes a
     TCP/IP connection with the HTTP server.
    </para>
    <para>Most of the ZOOM connection options do not
     affect SRU/Solr and they are ignored. However, future versions
     of &yaz; might honor <literal>implementationName</literal> and
     put that as part of User-Agent header for HTTP requests.
     </para>
    <para>
     The <literal>charset</literal> is used in the Content-Type header
     of HTTP requests.
    </para>
    <para>
     Setting <literal>authentcationMode</literal> specifies how
     authentication parameters are encoded for HTTP. The default is
     "<literal>basic</literal>" where <literal>user</literal> and
     <literal>password</literal> are encoded by using HTTP basic
     authentication.
     </para>
    <para>
     If <literal>authentcationMode</literal> is "<literal>url</literal>", then
     user and password are encoded in the URL by parameters
     <literal>x-username</literal> and <literal>x-password</literal> as
     given by the SRU standard.
    </para>
   </sect2>
  </sect1>
  <sect1 id="zoom.query">
   <title>Queries</title>
   <para>
    Query objects represents queries.
   </para>
   <synopsis>
     ZOOM_query ZOOM_query_create(void);

     void ZOOM_query_destroy(ZOOM_query q);

     int ZOOM_query_prefix(ZOOM_query q, const char *str);

     int ZOOM_query_cql(ZOOM_query s, const char *str);

     int ZOOM_query_sortby(ZOOM_query q, const char *criteria);

     int ZOOM_query_sortby2(ZOOM_query q, const char *strategy,
                            const char *criteria);
   </synopsis>
   <para>
    Create query objects using <function>ZOOM_query_create</function>
    and destroy them by calling <function>ZOOM_query_destroy</function>.
    RPN-queries can be specified in <link linkend="PQF">PQF</link>
    notation by using the
    function <function>ZOOM_query_prefix</function>.
    The <function>ZOOM_query_cql</function> specifies a CQL
    query to be sent to the server/target.
    More query types will be added in future versions of &yaz;, such as
    <link linkend="CCL">CCL</link> to RPN-mapping, native CCL query,
    etc. In addition to a search, a sort criteria may be set. Function
    <function>ZOOM_query_sortby</function> enables Z39.50 sorting and
    it takes sort criteria using the same string notation as
    yaz-client's <link linkend="sortspec">sort command</link>.
   </para>
   <para id="zoom.query.sortby2">
    <function>ZOOM_query_sortby2</function> is similar to
    <function>ZOOM_query_sortby</function> but allows a strategy for
    sorting. The reason for the strategy parameter is that some
    protocols offers multiple ways of performing sorting.
    For example, Z39.50 has the standard sort, which is performed after
    search on an existing result set.
    It's also possible to use CQL in Z39.50 as the query type and use
    CQL's SORTBY keyword. Finally, Index Data's
    Zebra server also allows sorting to be specified as part of RPN (Type 7).
   </para>
   <table id="zoom-sort-strategy" frame="top">
    <title>ZOOM sort strategy</title>
    <tgroup cols="2">
     <colspec colwidth="2*" colname="name"/>
     <colspec colwidth="5*" colname="description"/>
     <thead>
      <row>
       <entry>Name</entry>
       <entry>Description</entry>
      </row>
     </thead>
     <tbody>
      <row>
       <entry>z39.50</entry><entry>Z39.50 resultset sort</entry>
      </row>
      <row>
       <entry>type7</entry><entry>Sorting embedded in RPN(Type-7)</entry>
      </row>
      <row>
       <entry>cql</entry><entry>CQL SORTBY</entry>
      </row>
      <row>
       <entry>sru11</entry><entry>SRU sortKeys parameter</entry>
      </row>
      <row>
       <entry>solr</entry><entry>Solr sort</entry>
      </row>
      <row>
       <entry>embed</entry><entry>type7 for Z39.50, cql for SRU,
        solr for Solr protocol</entry>
      </row>
     </tbody>
    </tgroup>
   </table>
  </sect1>
  <sect1 id="zoom.resultsets"><title>Result sets</title>
   <para>
    The result set object is a container for records returned from
    a target.
   </para>
   <synopsis>
     ZOOM_resultset ZOOM_connection_search(ZOOM_connection, ZOOM_query q);

     ZOOM_resultset ZOOM_connection_search_pqf(ZOOM_connection c,
                                               const char *q);
     void ZOOM_resultset_destroy(ZOOM_resultset r);
   </synopsis>
   <para>
    Function <function>ZOOM_connection_search</function> creates
    a result set given a connection and query.
    Destroy a result set by calling
    <function>ZOOM_resultset_destroy</function>.
    Simple clients may using PQF only may use function
    <function>ZOOM_connection_search_pqf</function> in which case
    creating query objects is not necessary.
   </para>
   <synopsis>
     void ZOOM_resultset_option_set(ZOOM_resultset r,
                                    const char *key, const char *val);

     const char *ZOOM_resultset_option_get(ZOOM_resultset r, const char *key);

     size_t ZOOM_resultset_size(ZOOM_resultset r);
   </synopsis>
   <para>
    Functions <function>ZOOM_resultset_options_set</function> and
    <function>ZOOM_resultset_get</function> sets and gets an option
    for a result set similar to <function>ZOOM_connection_option_get</function>
    and <function>ZOOM_connection_option_set</function>.
   </para>
   <para>
    The number of hits also called result-count is returned by
    function <function>ZOOM_resultset_size</function>.
   </para>
   <table id="zoom.resultset.options"
    frame="top"><title>ZOOM Result set Options</title>
    <tgroup cols="3">
     <colspec colwidth="4*" colname="name"></colspec>
     <colspec colwidth="7*" colname="description"></colspec>
     <colspec colwidth="2*" colname="default"></colspec>
     <thead>
      <row>
       <entry>Option</entry>
       <entry>Description</entry>
       <entry>Default</entry>
      </row>
     </thead>
     <tbody>
      <row><entry>
        start</entry><entry>Offset of first record to be
        retrieved from target. First record has offset 0 unlike the
        protocol specifications where first record has position 1.
        This option affects ZOOM_resultset_search and
        ZOOM_resultset_search_pqf and must be set before any of
        these functions are invoked. If a range of
        records must be fetched manually after search,
        function ZOOM_resultset_records should be used.
       </entry><entry>0</entry></row>
      <row><entry>
        count</entry><entry>Number of records to be retrieved.
        This option affects ZOOM_resultset_search and
        ZOOM_resultset_search_pqf and must be set before any of
        these functions are invoked.
       </entry><entry>0</entry></row>
      <row><entry>
        presentChunk</entry><entry>The number of records to be
        requested from the server in each chunk (present request). The
        value 0 means to request all the records in a single chunk.
        (The old <literal>step</literal>
        option is also supported for the benefit of old applications.)
       </entry><entry>0</entry></row>
      <row><entry>
        elementSetName</entry><entry>Element-Set name of records.
        Most targets should honor element set name <literal>B</literal>
        and <literal>F</literal> for brief and full respectively.
       </entry><entry>none</entry></row>
      <row><entry>
        preferredRecordSyntax</entry><entry>Preferred Syntax, such as
        <literal>USMARC</literal>, <literal>SUTRS</literal>, etc.
       </entry><entry>none</entry></row>
      <row><entry>
        schema</entry><entry>Schema for retrieval, such as
        <literal>Gils-schema</literal>, <literal>Geo-schema</literal>, etc.
       </entry><entry>none</entry></row>
      <row><entry>
        setname</entry><entry>Name of Result Set (Result Set ID).
        If this option isn't set, the ZOOM module will automatically
        allocate a result set name.
       </entry><entry>default</entry></row>
      <row><entry>
        rpnCharset</entry><entry>Character set for RPN terms.
        If this is set, ZOOM C will assume that the ZOOM application is
        running UTF-8. Terms in RPN queries are then converted to the
        rpnCharset. If this is unset, ZOOM C will not assume any encoding
        of RPN terms and no conversion is performed.
       </entry><entry>none</entry></row>
     </tbody>
    </tgroup>
   </table>
   <para>
    For servers that support Search Info report, the following
    options may be read using <function>ZOOM_resultset_get</function>.
    This detailed information is read after a successful search has
    completed.
   </para>
   <para>
    This information is a list of of items, where each item is
    information about a term or subquery. All items in the list
    are prefixed by
    <literal>SearchResult.</literal><replaceable>no</replaceable>
    where no presents the item number (0=first, 1=second).
    Read <literal>searchresult.size</literal> to determine the
    number of items.
   </para>
   <table id="zoom.search.info.report.options"
    frame="top"><title>Search Info Report Options</title>
    <tgroup cols="2">
     <colspec colwidth="4*" colname="name"></colspec>
     <colspec colwidth="7*" colname="description"></colspec>
     <thead>
      <row>
       <entry>Option</entry>
       <entry>Description</entry>
      </row>
     </thead>
     <tbody>
      <row>
       <entry>searchresult.size</entry>
       <entry>
        number of search result entries. This option is-nonexistant
        if no entries are returned by the server.
       </entry>
      </row>
      <row>
       <entry>searchresult.<replaceable>no</replaceable>.id</entry>
       <entry>sub query ID</entry>
      </row>
      <row>
       <entry>searchresult.<replaceable>no</replaceable>.count</entry>
       <entry>result count for item (number of hits)</entry>
      </row>
      <row>
       <entry>searchresult.<replaceable>no</replaceable>.subquery.term</entry>
       <entry>subquery term</entry>
      </row>
      <row>
       <entry>
        searchresult.<replaceable>no</replaceable>.interpretation.term
       </entry>
       <entry>interpretation term</entry>
      </row>
      <row>
       <entry>
        searchresult.<replaceable>no</replaceable>.recommendation.term
       </entry>
       <entry>recommendation term</entry>
      </row>
     </tbody>
    </tgroup>
   </table>
   <sect2 id="zoom.z3950.resultset.sort">
    <title>Z39.50 Result-set Sort</title>
    <synopsis>
     void ZOOM_resultset_sort(ZOOM_resultset r,
                              const char *sort_type, const char *sort_spec);

     int ZOOM_resultset_sort1(ZOOM_resultset r,
                              const char *sort_type, const char *sort_spec);
    </synopsis>
    <para>
     <function>ZOOM_resultset_sort</function> and
     <function>ZOOM_resultset_sort1</function> both sort an existing
     result-set. The sort_type parameter is not use. Set it to "yaz".
     The sort_spec is same notation as ZOOM_query_sortby and identical
     to that offered by yaz-client's
     <link linkend="sortspec">sort command</link>.
    </para>
    <para>
     These functions only work for Z39.50. Use the more generic utility
     <link linkend="zoom.query.sortby2">
      <function>ZOOM_query_sortby2</function></link>
     for other protocols (and even Z39.50).
    </para>
   </sect2>
   <sect2 id="zoom.z3950.resultset.behavior">
    <title>Z39.50 Protocol behavior</title>
    <para>
     The creation of a result set involves at least a SearchRequest
     - SearchResponse protocol handshake. Following that, if a sort
     criteria was specified as part of the query, a SortRequest -
     SortResponse handshake takes place. Note that it is necessary to
     perform sorting before any retrieval takes place, so no records will
     be returned from the target as part of the SearchResponse because these
     would be unsorted. Hence, piggyback is disabled when sort criteria
     are set. Following Search - and a possible sort - Retrieval takes
     place - as one or more Present Requests/Response pairs being
     transferred.
     </para>
    <para>
     The API allows for two different modes for retrieval. A high level
     mode which is somewhat more powerful and a low level one.
     The low level is enabled when searching on a Connection object
     for which the settings
     <literal>smallSetUpperBound</literal>,
     <literal>mediumSetPresentNumber</literal> and
     <literal>largeSetLowerBound</literal> are set. The low level mode
     thus allows you to precisely set how records are returned as part
     of a search response as offered by the Z39.50 protocol.
     Since the client may be retrieving records as part of the
     search response, this mode doesn't work well if sorting is used.
     </para>
    <para>
     The high-level mode allows you to fetch a range of records from
     the result set with a given start offset. When you use this mode
     the client will automatically use piggyback if that is possible
     with the target and perform one or more present requests as needed.
     Even if the target returns fewer records as part of a present response
     because of a record size limit, etc. the client will repeat sending
     present requests. As an example, if option <literal>start</literal>
     is 0 (default) and <literal>count</literal> is 4, and
     <literal>piggyback</literal> is 1 (default) and no sorting criteria
     is specified, then the client will attempt to retrieve the 4
     records as part the search response (using piggyback). On the other
     hand, if either <literal>start</literal> is positive or if
     a sorting criteria is set, or if <literal>piggyback</literal>
     is 0, then the client will not perform piggyback but send Present
     Requests instead.
    </para>
    <para>
     If either of the options <literal>mediumSetElementSetName</literal> and
     <literal>smallSetElementSetName</literal> are unset, the value
     of option <literal>elementSetName</literal> is used for piggyback
     searches. This means that for the high-level mode you only have
     to specify one elementSetName option rather than three.
     </para>
   </sect2>
   <sect2 id="zoom.sru.resultset.behavior">
    <title>SRU Protocol behavior</title>
    <para>
     Current version of &yaz; does not take advantage of a result set id
     returned by the SRU server. Future versions might do, however.
     Since, the ZOOM driver does not save result set IDs any
     present (retrieval) is transformed to a SRU SearchRetrieveRequest
     with same query but, possibly, different offsets.
    </para>
    <para>
     Option <literal>schema</literal> specifies SRU schema
     for retrieval. However, options <literal>elementSetName</literal> and
     <literal>preferredRecordSyntax</literal> are ignored.
    </para>
    <para>
     Options <literal>start</literal> and <literal>count</literal>
     are supported by SRU.
     The remaining options
     <literal>piggyback</literal>,
     <literal>smallSetUpperBound</literal>,
     <literal>largeSetLowerBound</literal>,
     <literal>mediumSetPresentNumber</literal>,
     <literal>mediumSetElementSetName</literal>,
      <literal>smallSetElementSetName</literal> are
     unsupported.
    </para>
    <para>
     SRU supports CQL queries, <emphasis>not</emphasis> PQF.
     If PQF is used, however, the PQF query is transferred anyway
     using non-standard element <literal>pQuery</literal> in
     SRU SearchRetrieveRequest.
    </para>
    <para>
     Solr queries has to be done in Solr query format.
    </para>
    <para>
     Unfortunately, SRU or Solr does not define a database setting. Hence,
     <literal>databaseName</literal> is unsupported and ignored.
     However, the path part in host parameter for functions
     <function>ZOOM_connecton_new</function> and
     <function>ZOOM_connection_connect</function> acts as a
     database (at least for the &yaz; SRU server).
    </para>
   </sect2>
  </sect1>
  <sect1 id="zoom.records">
   <title>Records</title>
   <para>
    A record object is a retrieval record on the client side -
    created from result sets.
   </para>
   <synopsis>
     void ZOOM_resultset_records(ZOOM_resultset r,
                                 ZOOM_record *recs,
                                 size_t start, size_t count);
     ZOOM_record ZOOM_resultset_record(ZOOM_resultset s, size_t pos);

     const char *ZOOM_record_get(ZOOM_record rec, const char *type,
                                 size_t *len);

     int ZOOM_record_error(ZOOM_record rec, const char **msg,
                           const char **addinfo, const char **diagset);

     ZOOM_record ZOOM_record_clone(ZOOM_record rec);

     void ZOOM_record_destroy(ZOOM_record rec);
   </synopsis>
   <para>
    References to temporary records are returned by functions
    <function>ZOOM_resultset_records</function> or
    <function>ZOOM_resultset_record</function>.
    </para>
   <para>
    If a persistent reference to a record is desired
    <function>ZOOM_record_clone</function> should be used.
    It returns a record reference that should be destroyed
    by a call to <function>ZOOM_record_destroy</function>.
   </para>
   <para>
    A single record is returned by function
    <function>ZOOM_resultset_record</function> that takes a
    position as argument. First record has position zero.
    If no record could be obtained <literal>NULL</literal> is returned.
   </para>
   <para>
    Error information for a record can be checked with
    <function>ZOOM_record_error</function> which returns non-zero
    (error code) if record is in error, called <emphasis>Surrogate
     Diagnostics</emphasis> in Z39.50.
   </para>
   <para>
    Function <function>ZOOM_resultset_records</function> retrieves
    a number of records from a result set. Parameter <literal>start</literal>
    and <literal>count</literal> specifies the range of records to
    be returned. Upon completion array
    <literal>recs[0], ..recs[count-1]</literal>
    holds record objects for the records. The array of records
     <literal>recs</literal> should be allocated prior the call
    <function>ZOOM_resultset_records</function>. Note that for those
    records that couldn't be retrieved from the target
    <literal>recs[ ..]</literal> is set to <literal>NULL</literal>.
   </para>
   <para id="zoom.record.get">
    In order to extract information about a single record,
    <function>ZOOM_record_get</function> is provided. The
    function returns a pointer to certain record information. The
    nature (type) of the pointer depends on the parameter,
    <parameter>type</parameter>.
   </para>
   <para>
    The <parameter>type</parameter> is a string of the format:
   </para>
   <para>
    <replaceable>format</replaceable>[;charset=<replaceable>from</replaceable>[/<replaceable>opacfrom</replaceable>][,<replaceable>to</replaceable>]][;format=<replaceable>v</replaceable>][;base64=<replaceable>xpath</replaceable>]
   </para>
   <para>
    If <literal>charset</literal> is given, then <replaceable>from</replaceable>
    specifies the character set of the record in its original form
    (as returned by the server), <replaceable>to</replaceable> specifies
    the output (returned) character set encoding.
    If <replaceable>to</replaceable> is omitted, then UTF-8 is assumed.
    If charset is not given, then no character set conversion takes place.
    OPAC records may be returned in a different
    set from the bibliographic MARC record. If this is this the case,
    <replaceable>opacfrom</replaceable> should be set to the character set
    of the OPAC record part.
   </para>

   <para>
    The <literal>format</literal> is generic but can only be used to
    specify XML indentation when the value <replaceable>v</replaceable>
    is 1 (<literal>format=1</literal>).
   </para>
   <para>
    The <literal>base64</literal> allows a full record to be extracted
    from base64-encoded string in an XML document.
   </para>
   <note>
     <para>
       Specifying the OPAC record character set requires YAZ 4.1.5 or later.
     </para>
     <para>
       Specifying the base64 parameter requires YAZ 4.2.35 or later.
     </para>
   </note>
   <para>
    The format argument controls whether record data should be XML
    pretty-printed (post process operation).
    It is enabled only if format value <replaceable>v</replaceable> is
    <literal>1</literal> and the record content is XML well-formed.
   </para>
   <para>
    In addition, for certain types, the length
    <literal>len</literal> passed will be set to the size in bytes of
    the returned information.
    </para>
   <para>
    The following are the supported values for <replaceable>form</replaceable>.
    <variablelist>
     <varlistentry><term><literal>database</literal></term>
      <listitem><para>Database of record is returned
        as a C null-terminated string. Return type
        <literal>const char *</literal>.
       </para></listitem>
     </varlistentry>
     <varlistentry><term><literal>syntax</literal></term>
      <listitem><para>The transfer syntax of the record is returned
        as a C null-terminated string containing the symbolic name of
        the record syntax, e.g. <literal>Usmarc</literal>. Return type
        is
        <literal>const char *</literal>.
       </para></listitem>
     </varlistentry>
     <varlistentry><term><literal>schema</literal></term>
      <listitem><para>The schema of the record is returned
        as a C null-terminated string. Return type is
        <literal>const char *</literal>.
       </para></listitem>
     </varlistentry>
     <varlistentry><term><literal>render</literal></term>
      <listitem><para>The record is returned in a display friendly
        format. Upon completion buffer is returned
        (type <literal>const char *</literal>) and length is stored in
        <literal>*len</literal>.
       </para></listitem>
     </varlistentry>
     <varlistentry><term><literal>raw</literal></term>
      <listitem><para>The record is returned in the internal
        YAZ specific format. For GRS-1, Explain, and others, the
        raw data is returned as type
        <literal>Z_External *</literal> which is just the type for
        the member <literal>retrievalRecord</literal> in
        type <literal>NamePlusRecord</literal>.
        For SUTRS and octet aligned record (including all MARCs) the
        octet buffer is returned and the length of the buffer.
       </para></listitem>
     </varlistentry>
     <varlistentry><term><literal>xml</literal></term>
      <listitem><para>The record is returned in XML if possible.
        SRU, Solr and Z39.50 records with transfer syntax XML are
        returned verbatim. MARC records are returned in
        <ulink url="&url.marcxml;">
         MARCXML
         </ulink>
        (converted from ISO2709 to MARCXML by YAZ).
        OPAC records are also converted to XML and the
        bibliographic record is converted to MARCXML (when possible).
        GRS-1 records are not supported for this form.
        Upon completion, the XML buffer is returned
        (type <literal>const char *</literal>) and length is stored in
        <literal>*len</literal>.
       </para></listitem>
     </varlistentry>
     <varlistentry><term><literal>opac</literal></term>
      <listitem><para>OPAC information for record is returned in XML
        if an OPAC record is present at the position given. If no
        OPAC record is present, a NULL pointer is returned.
       </para></listitem>
     </varlistentry>
     <varlistentry><term><literal>txml</literal></term>
      <listitem><para>The record is returned in TurboMARC if possible.
        SRU and Z39.50 records with transfer syntax XML are
        returned verbatim. MARC records are returned in
        <link linkend="tools.turbomarc">
         TurboMARC
        </link>
        (converted from ISO2709 to TurboMARC by YAZ).
        Upon completion, the XML buffer is returned
        (type <literal>const char *</literal>) and length is stored in
        <literal>*len</literal>.
       </para></listitem>
     </varlistentry>
     <varlistentry><term><literal>json</literal></term>
      <listitem><para>Like xml, but MARC records are converted to
        <ulink url="&url.marc_in_json;">MARC-in-JSON</ulink>.
       </para></listitem>
     </varlistentry>

    </variablelist>
   </para>
   <para>
    Most
    <ulink url="&url.marc21;">MARC21</ulink>
    records uses the
    <ulink url="&url.marc8;">MARC-8</ulink>
    character set encoding.
    An application that wishes to display in Latin-1 would use
    <screen>
     render; charset=marc8,iso-8859-1
    </screen>
   </para>
   <sect2 id="zoom.z3950.record.behavior">
    <title>Z39.50 Protocol behavior</title>
    <para>
     The functions <function>ZOOM_resultset_record</function> and
     <function>ZOOM_resultset_records</function> inspects the client-side
     record cache. Records not found in cache are fetched using
     Present.
     The functions may block (and perform network I/O)  - even though option
     <literal>async</literal> is 1, because they return records objects.
     (and there's no way to return records objects without retrieving them!).
     </para>
    <para>
     There is a trick, however, in the usage of function
     <function>ZOOM_resultset_records</function> that allows for
     delayed retrieval (and makes it non-blocking). By using
     a null pointer for <parameter>recs</parameter> you're indicating
     you're not interested in getting records objects
     <emphasis>now</emphasis>.
    </para>
   </sect2>
   <sect2 id="zoom.sru.record.behavior">
    <title>SRU/Solr Protocol behavior</title>
    <para>
     The ZOOM driver for SRU/Solr treats records returned by a SRU/Solr server
     as if they where Z39.50 records with transfer syntax XML and
     no element set name or database name.
    </para>
   </sect2>
  </sect1>
  <sect1 id="zoom.facets"><title>Facets</title>
   <para>
    Facet operations is not part of the official ZOOM specification, but
    is an Index Data extension for YAZ-based Z39.50 targets,
    <ulink url="&url.solr;">Solr</ulink> and SRU 2.0 targets.

    Facets may be requestd by the
     <link linkend="zoom.facets.option">facets</link> option before a
    search is sent.
    For inspection of the returned facets, the following functions are
    available:
   </para>
   <synopsis>
    ZOOM_facet_field *ZOOM_resultset_facets(ZOOM_resultset r);

    ZOOM_facet_field ZOOM_resultset_get_facet_field(ZOOM_resultset r,
                                                    const char *facet_name);

    ZOOM_facet_field ZOOM_resultset_get_facet_field_by_index(ZOOM_resultset r,
                                                             int pos);

    size_t ZOOM_resultset_facets_size(ZOOM_resultset r);

    const char *ZOOM_facet_field_name(ZOOM_facet_field facet_field);

    size_t ZOOM_facet_field_term_count(ZOOM_facet_field facet_field);

    const char *ZOOM_facet_field_get_term(ZOOM_facet_field facet_field,
                                          size_t idx, int *freq);
   </synopsis>
   <para>
    References to temporary structures are returned by all functions.
    They are only valid as long the Result set is valid.
    <function>ZOOM_resultset_get_facet_field</function> or
    <function>ZOOM_resultset_get_facet_field_by_index</function>.
    <function>ZOOM_resultset_facets</function>.
    <function>ZOOM_facet_field_name</function>.
    <function>ZOOM_facet_field_get_term</function>.
    </para>
   <para id="zoom.resultset.get_facet_field">
    A single Facet field  is returned by function
    <function>ZOOM_resultset_get_facet_field</function> or
    <function>ZOOM_resultset_get_facet_field_by_index</function> that takes
    a  result set and facet name or positive index respectively. First
    facet has position zero. If no facet could be obtained (invalid name
    or index out of bounds) <literal>NULL</literal> is returned.
   </para>
   <para id="zoom.resultset.facets">
    An array of facets field can be returned by
    <function>ZOOM_resultset_facets</function>. The length of the array is
    given by <function>ZOOM_resultset_facets_size</function>. The array is
    zero-based and last entry will be at
    <function>ZOOM_resultset_facets_size(result_set)</function>-1.
   </para>
   <para id="zoom.resultset.facets_names">
    It is possible to interate over facets by name, by calling
    <function>ZOOM_resultset_facets_names</function>.
    This will return an const array of char * where each string can be used
    as parameter for <function>ZOOM_resultset_get_facet_field</function>.
   </para>
   <para>
   Function <function>ZOOM_facet_field_name</function> gets the request
    facet name from a returned facet field.
   </para>
   <para>
    Function <function>ZOOM_facet_field_get_term</function> returns the
    idx'th term and term count for a facet field.
    Idx must between 0 and
    <function>ZOOM_facet_field_term_count</function>-1, otherwise the
    returned reference will be <literal>NULL</literal>. On a valid idx, the
    value of the freq reference will be the term count.
    The <literal>freq</literal> parameter must be valid pointer to integer.
   </para>
   </sect1>
  <sect1 id="zoom.scan"><title>Scan</title>
   <para>
    This section describes an interface for Scan. Scan is not an
    official part of the ZOOM model yet. The result of a scan operation
    is the <literal>ZOOM_scanset</literal> which is a set of terms
    returned by a target.
   </para>

   <para>
    The Scan interface is supported for both Z39.50, SRU and Solr.
   </para>

   <synopsis>
    ZOOM_scanset ZOOM_connection_scan(ZOOM_connection c,
                                      const char *startpqf);

    ZOOM_scanset ZOOM_connection_scan1(ZOOM_connection c,
                                       ZOOM_query q);

    size_t ZOOM_scanset_size(ZOOM_scanset scan);

    const char *ZOOM_scanset_term(ZOOM_scanset scan, size_t pos,
                                  size_t *occ, size_t *len);

    const char *ZOOM_scanset_display_term(ZOOM_scanset scan, size_t pos,
                                          size_t *occ, size_t *len);

    void ZOOM_scanset_destroy(ZOOM_scanset scan);

    const char *ZOOM_scanset_option_get(ZOOM_scanset scan,
                                        const char *key);

    void ZOOM_scanset_option_set(ZOOM_scanset scan, const char *key,
                                 const char *val);
    </synopsis>
   <para>
    The scan set is created by function
    <function>ZOOM_connection_scan</function> which performs a scan
    operation on the connection using the specified
    <parameter>startpqf</parameter>.
    If the operation was successful, the size of the scan set can be
    retrieved by a call to <function>ZOOM_scanset_size</function>.
    Like result sets, the items are numbered 0,..size-1.
    To obtain information about a particular scan term, call function
    <function>ZOOM_scanset_term</function>. This function takes
    a scan set offset <literal>pos</literal> and returns a pointer
    to a <emphasis>raw term</emphasis> or <literal>NULL</literal> if
    non-present.
    If present, the <literal>occ</literal> and <literal>len</literal>
    are set to the number of occurrences and the length
    of the actual term respectively.
    <function>ZOOM_scanset_display_term</function> is similar to
    <function>ZOOM_scanset_term</function> except that it returns
    the <emphasis>display term</emphasis> rather than the raw term.
    In a few cases, the term is different from display term. Always
    use the display term for display and the raw term for subsequent
    scan operations (to get more terms, next scan result, etc).
   </para>
   <para>
    A scan set may be freed by a call to function
    <function>ZOOM_scanset_destroy</function>.
    Functions <function>ZOOM_scanset_option_get</function> and
    <function>ZOOM_scanset_option_set</function> retrieves and sets
    an option respectively.
   </para>
   <para>
    The <parameter>startpqf</parameter> is a subset of PQF, namely
    the Attributes+Term part. Multiple <literal>@attr</literal> can
    be used. For example to scan in title (complete) phrases:
    <literallayout>
     @attr 1=4 @attr 6=2 "science o"
    </literallayout>
   </para>
   <para>
    The <function>ZOOM_connecton_scan1</function> is a newer and
    more generic alternative to <function>ZOOM_connection_scan</function>
    which allows to use both CQL and PQF for Scan.
   </para>
   <table frame="top" id="zoom.scanset.options">
    <title>ZOOM Scan Set Options</title>
    <tgroup cols="3">
     <colspec colwidth="4*" colname="name"></colspec>
     <colspec colwidth="7*" colname="description"></colspec>
     <colspec colwidth="2*" colname="default"></colspec>
     <thead>
      <row>
       <entry>Option</entry>
       <entry>Description</entry>
       <entry>Default</entry>
      </row>
     </thead>
     <tbody>
      <row><entry>
        number</entry><entry>Number of Scan Terms requested in next scan.
        After scan it holds the actual number of terms returned.
       </entry><entry>20</entry></row>
      <row><entry>
        position</entry><entry>Preferred Position of term in response
        in next scan; actual position after completion of scan.
       </entry><entry>1</entry></row>
      <row><entry>
        stepSize</entry><entry>Step Size
       </entry><entry>0</entry></row>
      <row><entry>
        scanStatus</entry><entry>An integer indicating the Scan Status
        of last scan.
       </entry><entry>0</entry></row>
      <row><entry>
        rpnCharset</entry><entry>Character set for RPN terms.
        If this is set, ZOOM C will assume that the ZOOM application is
        running UTF-8. Terms in RPN queries are then converted to the
        rpnCharset. If this is unset, ZOOM C will not assume any encoding
        of RPN terms and no conversion is performed.
       </entry><entry>none</entry></row>
     </tbody>
    </tgroup>
   </table>
  </sect1>
  <sect1 id="zoom.extendedservices">
   <title>Extended Services</title>
   <para>
    ZOOM offers an interface to a subset of the Z39.50 extended services
    as well as a few privately defined ones:
   </para>
   <itemizedlist>
    <listitem>
     <para>
      Z39.50 Item Order (ILL).
      See <xref linkend="zoom.item.order"/>.
     </para>
    </listitem>
    <listitem>
     <para>
      Record Update. This allows a client to insert, modify or delete
      records.
      See <xref linkend="zoom.record.update"/>.
     </para>
    </listitem>
    <listitem>
     <para>
      Database Create. This a non-standard feature. Allows a client
      to create a database.
      See <xref linkend="zoom.database.create"/>.
     </para>
    </listitem>
    <listitem>
     <para>
      Database Drop. This a non-standard feature. Allows a client
      to delete/drop a database.
      See <xref linkend="zoom.database.drop"/>.
     </para>
     </listitem>
    <listitem>
     <para>
      Commit operation. This a non-standard feature. Allows a client
      to commit operations.
      See <xref linkend="zoom.commit"/>.
     </para>
    </listitem>
    <!-- all the ILL PDU options should go here too -->
   </itemizedlist>
   <para>
    To create an extended service operation a <literal>ZOOM_package</literal>
    must be created. The operation is a five step operation. The
    package is created, package is configured by means of options,
    the package is send, result is inspected (by means of options),
    the package is destroyed.
   </para>
   <synopsis>
    ZOOM_package ZOOM_connection_package(ZOOM_connection c,
                                         ZOOM_options options);

    const char *ZOOM_package_option_get(ZOOM_package p,
                                        const char *key);
    void ZOOM_package_option_set(ZOOM_package p, const char *key,
                                 const char *val);
    void ZOOM_package_send(ZOOM_package p, const char *type);

    void ZOOM_package_destroy(ZOOM_package p);
   </synopsis>
   <para>
    The <function>ZOOM_connection_package</function> creates a
    package for the connection given using the options specified.
   </para>
   <para>
    Functions <function>ZOOM_package_option_get</function> and
    <function>ZOOM_package_option_set</function> gets and sets
    options.
   </para>
   <para>
    <function>ZOOM_package_send</function> sends
    the package the via connection specified in
    <function>ZOOM_connection_package</function>.
    The <parameter>type</parameter> specifies the actual extended service
    package type to be sent.
   </para>
   <table frame="top" id="zoom.extendedservices.options">
    <title>Extended Service Common Options</title>
    <tgroup cols="3">
     <colspec colwidth="4*" colname="name"></colspec>
     <colspec colwidth="7*" colname="description"></colspec>
     <colspec colwidth="3*" colname="default"></colspec>
     <thead>
      <row>
       <entry>Option</entry>
       <entry>Description</entry>
       <entry>Default</entry>
      </row>
     </thead>
     <tbody>
      <row>
       <entry>package-name</entry>
       <entry>Extended Service Request package name. Must be specified
       as part of a request</entry>
       <entry>none</entry>
      </row>
      <row>
       <entry>user-id</entry>
       <entry>User ID of Extended Service Package. Is a request option</entry>
       <entry>none</entry>
      </row>
      <row>
       <entry>function</entry>
       <entry>
        Function of package - one of <literal>create</literal>,
        <literal>delete</literal>, <literal>modify</literal>. Is
        a request option.
       </entry>
       <entry><literal>create</literal></entry>
      </row>
      <row>
       <entry>waitAction</entry>
       <entry>
        Wait action for package. Possible values:
        <literal>wait</literal>, <literal>waitIfPossible</literal>,
        <literal>dontWait</literal> or <literal>dontReturnPackage</literal>.
       </entry>
       <entry><literal>waitIfPossible</literal></entry>
      </row>
      <row>
       <entry>targetReference</entry>
       <entry>
        Target Reference. This is part of the response as returned
        by the server. Read it after a successful operation.
       </entry>
       <entry><literal>none</literal></entry>
      </row>
     </tbody>
    </tgroup>
   </table>
   <sect2 id="zoom.item.order">
    <title>Item Order</title>
    <para>
     For Item Order, type must be set to <literal>itemorder</literal> in
     <function>ZOOM_package_send</function>.
    </para>

    <table frame="top" id="zoom.item.order.options">
     <title>Item Order Options</title>
     <tgroup cols="3">
      <colspec colwidth="4*" colname="name"></colspec>
      <colspec colwidth="7*" colname="description"></colspec>
      <colspec colwidth="3*" colname="default"></colspec>
      <thead>
       <row>
        <entry>Option</entry>
        <entry>Description</entry>
        <entry>Default</entry>
       </row>
      </thead>
      <tbody>
       <row>
        <entry>contact-name</entry>
        <entry>ILL contact name</entry>
        <entry>none</entry>
       </row>
       <row>
        <entry>contact-phone</entry>
        <entry>ILL contact phone</entry>
        <entry>none</entry>
       </row>
       <row>
        <entry>contact-email</entry>
        <entry>ILL contact email</entry>
        <entry>none</entry>
       </row>
       <row>
        <entry>itemorder-item</entry>
        <entry>Position for item (record) requested. An integer</entry>
        <entry>1</entry>
       </row>
      </tbody>
     </tgroup>
    </table>
   </sect2>
   <sect2 id="zoom.record.update">
    <title>Record Update</title>
    <para>
     For Record Update, type must be set to <literal>update</literal> in
     <function>ZOOM_package_send</function>.
    </para>
    <table frame="top" id="zoom.record.update.options">
     <title>Record Update Options</title>
     <tgroup cols="3">
      <colspec colwidth="4*" colname="name"></colspec>
      <colspec colwidth="7*" colname="description"></colspec>
      <colspec colwidth="3*" colname="default"></colspec>
      <thead>
       <row>
        <entry>Option</entry>
        <entry>Description</entry>
        <entry>Default</entry>
       </row>
      </thead>
      <tbody>
       <row>
        <entry>action</entry>
        <entry>
         The update action. One of
         <literal>specialUpdate</literal>,
         <literal>recordInsert</literal>,
         <literal>recordReplace</literal>,
         <literal>recordDelete</literal>,
         <literal>elementUpdate</literal>.
        </entry>
        <entry><literal>specialUpdate (recordInsert for updateVersion=1 which does not support specialUpdate)</literal></entry>
       </row>
       <row>
        <entry>recordIdOpaque</entry>
        <entry>Opaque Record ID</entry>
        <entry>none</entry>
       </row>
       <row>
        <entry>recordIdNumber</entry>
        <entry>Record ID number</entry>
        <entry>none</entry>
       </row>
       <row>
        <entry>record</entry>
        <entry>The record itself</entry>
        <entry>none</entry>
       </row>
       <row>
        <entry>recordOpaque</entry>
        <entry>Specifies an opaque record which is
          encoded as an ASN.1 ANY type with the OID as tiven by option
          <literal>syntax</literal> (see below).
          Option <literal>recordOpaque</literal> is an alternative
          to record - and <literal>record</literal> option (above) is
          ignored if recordOpaque is set. This option is only available in
          YAZ 3.0.35 and later and is meant to facilitate Updates with
          servers from OCLC.
        </entry>
        <entry>none</entry>
       </row>
       <row>
        <entry>syntax</entry>
        <entry>The record syntax (transfer syntax). Is a string that
         is a known record syntax.
        </entry>
        <entry>no syntax</entry>
       </row>
       <row>
        <entry>databaseName</entry>
        <entry>Database from connection object</entry>
        <entry>Default</entry>
       </row>
       <row>
        <entry>correlationInfo.note</entry>
        <entry>Correlation Info Note (string)</entry>
        <entry>none</entry>
       </row>
       <row>
        <entry>correlationInfo.id</entry>
        <entry>Correlation Info ID (integer)</entry>
        <entry>none</entry>
       </row>
       <row>
        <entry>elementSetName</entry>
        <entry>Element Set for Record</entry>
        <entry>none</entry>
       </row>
       <row>
        <entry>updateVersion</entry>
        <entry>Record Update version which holds one of the values
         1, 2 or 3. Each version has a distinct OID:
         1.2.840.10003.9.5
         (<ulink url="&url.z39.50.extupdate1;">first version</ulink>) ,
         1.2.840.10003.9.5.1
         (second version) and
         1.2.840.10003.9.5.1.1
         (<ulink url="&url.z39.50.extupdate3;">third and
          newest version</ulink>).
        </entry>
        <entry>3</entry>
       </row>
      </tbody>
     </tgroup>
    </table>

   </sect2>

   <sect2 id="zoom.database.create"><title>Database Create</title>
    <para>
     For Database Create, type must be set to <literal>create</literal> in
     <function>ZOOM_package_send</function>.
    </para>

    <table frame="top" id="zoom.database.create.options">
     <title>Database Create Options</title>
     <tgroup cols="3">
      <colspec colwidth="4*" colname="name"></colspec>
      <colspec colwidth="7*" colname="description"></colspec>
      <colspec colwidth="3*" colname="default"></colspec>
      <thead>
       <row>
        <entry>Option</entry>
        <entry>Description</entry>
        <entry>Default</entry>
       </row>
      </thead>
      <tbody>
       <row>
        <entry>databaseName</entry>
        <entry>Database from connection object</entry>
        <entry>Default</entry>
       </row>
      </tbody>
     </tgroup>
    </table>
   </sect2>
   <sect2 id="zoom.database.drop">
    <title>Database Drop</title>
    <para>
     For Database Drop, type must be set to <literal>drop</literal> in
     <function>ZOOM_package_send</function>.
    </para>
    <table frame="top" id="zoom.database.drop.options">
     <title>Database Drop Options</title>
     <tgroup cols="3">
      <colspec colwidth="4*" colname="name"></colspec>
      <colspec colwidth="7*" colname="description"></colspec>
      <colspec colwidth="3*" colname="default"></colspec>
      <thead>
       <row>
        <entry>Option</entry>
        <entry>Description</entry>
        <entry>Default</entry>
       </row>
      </thead>
      <tbody>
       <row>
        <entry>databaseName</entry>
        <entry>Database from connection object</entry>
        <entry>Default</entry>
       </row>
      </tbody>
     </tgroup>
    </table>
   </sect2>
   <sect2 id="zoom.commit">
    <title>Commit Operation</title>
    <para>
     For Commit, type must be set to <literal>commit</literal> in
     <function>ZOOM_package_send</function>.
    </para>
   </sect2>
   <sect2 id="zoom.extended.services.behavior">
    <title>Protocol behavior</title>
    <para>
     All the extended services are Z39.50-only.
    </para>
    <note>
     <para>
      The database create, drop and commit services are privately defined
      operations.
      Refer to <filename>esadmin.asn</filename> in YAZ for the ASN.1
      definitions.
     </para>
    </note>
   </sect2>
  </sect1>
  <sect1 id="zoom.options">
   <title>Options</title>
   <para>
    Most &zoom; objects provide a way to specify options to change behavior.
    From an implementation point of view a set of options is just like
    an associative array / hash.
   </para>
   <synopsis>
     ZOOM_options ZOOM_options_create(void);

     ZOOM_options ZOOM_options_create_with_parent(ZOOM_options parent);

     void ZOOM_options_destroy(ZOOM_options opt);
   </synopsis>
   <synopsis>
     const char *ZOOM_options_get(ZOOM_options opt, const char *name);

     void ZOOM_options_set(ZOOM_options opt, const char *name,
                           const char *v);
   </synopsis>
   <synopsis>
     typedef const char *(*ZOOM_options_callback)
                            (void *handle, const char *name);

     ZOOM_options_callback
             ZOOM_options_set_callback(ZOOM_options opt,
                                       ZOOM_options_callback c,
                                       void *handle);
   </synopsis>
  </sect1>
  <sect1 id="zoom.queryconversions">
   <title>Query conversions</title>
   <synopsis>
    int ZOOM_query_cql2rpn(ZOOM_query s, const char *cql_str,
                           ZOOM_connection conn);

    int ZOOM_query_ccl2rpn(ZOOM_query s, const char *ccl_str,
                           const char *config,
                           int *ccl_error, const char **error_string,
                           int *error_pos);
   </synopsis>
   <para>
    <function>ZOOM_query_cql2rpn</function> translates the CQL string,
    client-side, into RPN which may be passed to the server.
    This is useful for server's that don't themselves
    support CQL, for which <function>ZOOM_query_cql</function> is useless.
    `conn' is used  only as a place to stash diagnostics if compilation
    fails; if this information is not needed, a null pointer may be used.
    The CQL conversion is driven by option <literal>cqlfile</literal> from
    connection conn. This specifies a conversion file (eg pqf.properties)
    which <emphasis>must</emphasis> be present.
   </para>
   <para>
    <function>ZOOM_query_ccl2rpn</function> translates the CCL string,
    client-side, into RPN which may be passed to the server.
    The conversion is driven by the specification given by
    <literal>config</literal>. Upon completion 0 is returned on success; -1
    is returned on on failure. Om failure <literal>error_string</literal> and
    <literal>error_pos</literal> holds error message and position of
    first error in original CCL string.
   </para>
  </sect1>
  <sect1 id="zoom.events"><title>Events</title>
   <para>
    If you're developing non-blocking applications, you have to deal
    with events.
   </para>
   <synopsis>
    int ZOOM_event(int no, ZOOM_connection *cs);
   </synopsis>
   <para>
    The <function>ZOOM_event</function> executes pending events for
    a number of connections. Supply the number of connections in
    <literal>no</literal> and an array of connections in
    <literal>cs</literal> (<literal>cs[0] ... cs[no-1]</literal>).
    A pending event could be a sending a search, receiving a response,
    etc.
    When an event has occurred for one of the connections, this function
    returns a positive integer <literal>n</literal> denoting that an event
    occurred for connection <literal>cs[n-1]</literal>.
    When no events are pending for the connections, a value of zero is
    returned.
    To ensure that all outstanding requests are performed call this function
    repeatedly until zero is returned.
   </para>
   <para>
    If <function>ZOOM_event</function> returns and returns non-zero, the
    last event that occurred can be expected.
   </para>
   <synopsis>
    int ZOOM_connection_last_event(ZOOM_connection cs);
   </synopsis>
   <para>
    <function>ZOOM_connection_last_event</function> returns an event type
    (integer) for the last event.
   </para>

   <table frame="top" id="zoom.event.ids">
    <title>ZOOM Event IDs</title>
    <tgroup cols="2">
     <colspec colwidth="4*" colname="name"></colspec>
     <colspec colwidth="7*" colname="description"></colspec>
     <thead>
      <row>
       <entry>Event</entry>
       <entry>Description</entry>
      </row>
     </thead>
     <tbody>
      <row>
       <entry>ZOOM_EVENT_NONE</entry>
       <entry>No event has occurred</entry>
      </row>
      <row>
       <entry>ZOOM_EVENT_CONNECT</entry>
       <entry>TCP/IP connect has initiated</entry>
      </row>
      <row>
       <entry>ZOOM_EVENT_SEND_DATA</entry>
       <entry>Data has been transmitted (sending)</entry>
      </row>
      <row>
       <entry>ZOOM_EVENT_RECV_DATA</entry>
       <entry>Data has been received)</entry>
      </row>
      <row>
       <entry>ZOOM_EVENT_TIMEOUT</entry>
       <entry>Timeout</entry>
      </row>
      <row>
       <entry>ZOOM_EVENT_UNKNOWN</entry>
       <entry>Unknown event</entry>
      </row>
      <row>
       <entry>ZOOM_EVENT_SEND_APDU</entry>
       <entry>An APDU has been transmitted (sending)</entry>
      </row>
      <row>
       <entry>ZOOM_EVENT_RECV_APDU</entry>
       <entry>An APDU has been received</entry>
      </row>
      <row>
       <entry>ZOOM_EVENT_RECV_RECORD</entry>
       <entry>A result-set record has been received</entry>
      </row>
      <row>
       <entry>ZOOM_EVENT_RECV_SEARCH</entry>
       <entry>A search result been received</entry>
      </row>
     </tbody>
    </tgroup>
   </table>
  </sect1>
 </chapter>
 <chapter id="server">
  <title>Generic server</title>
  <sect1 id="server.introduction"><title>Introduction</title>
   <para>
    If you aren't into documentation, a good way to learn how the
    back end interface works is to look at the <filename>backend.h</filename>
    file. Then, look at the small dummy-server in
    <filename>ztest/ztest.c</filename>. The <filename>backend.h</filename>
    file also makes a good reference, once you've chewed your way through
    the prose of this file.
   </para>
   <para>
    If you have a database system that you would like to make available by
    means of Z39.50 or SRU, &yaz; basically offers your two options. You
    can use the APIs provided by the &asn;, &odr;, and &comstack;
    modules to
    create and decode PDUs, and exchange them with a client.
    Using this low-level interface gives you access to all fields and
    options of the protocol, and you can construct your server as close
    to your existing database as you like.
    It is also a fairly involved process, requiring
    you to set up an event-handling mechanism, protocol state machine,
    etc. To simplify server implementation, we have implemented a compact
    and simple, but reasonably full-functioned server-frontend that will
    handle most of the protocol mechanics, while leaving you to
    concentrate on your database interface.
   </para>
   <note>
    <para>
     The backend interface was designed in anticipation of a specific
     integration task, while still attempting to achieve some degree of
     generality. We realize fully that there are points where the
     interface can be improved significantly. If you have specific
     functions or parameters that you think could be useful, send us a
     mail (or better, sign on to the mailing list referred to in the
     top-level README file). We will try to fit good suggestions into future
     releases, to the extent that it can be done without requiring
     too many structural changes in existing applications.
    </para>
   </note>
   <note>
    <para>
     The &yaz; server does not support XCQL.
     </para>
   </note>
  </sect1>
  <sect1 id="server.frontend">
   <title>The Database Frontend</title>
   <para>
    We refer to this software as a generic database frontend. Your
    database system is the <emphasis>backend database</emphasis>, and the
    interface between the two is called the <emphasis>backend API</emphasis>.
    The backend API consists of a small number of function handlers and
    structure definitions. You are required to provide the
    <function>main()</function> routine for the server (which can be
    quite simple), as well as a set of handlers to match each of the
    prototypes.
    The interface functions that you write can use any mechanism you like
    to communicate with your database system: You might link the whole
    thing together with your database application and access it by
    function calls; you might use IPC to talk to a database server
    somewhere; or you might link with third-party software that handles
    the communication for you (like a commercial database client library).
    At any rate, the handlers will perform the tasks of:
   </para>
   <itemizedlist>
    <listitem><para>
     Initialization.
    </para></listitem>
    <listitem><para>
     Searching.
    </para></listitem>
    <listitem><para>
     Fetching records.
    </para></listitem>
    <listitem><para>
     Scanning the database index (optional - if you wish to implement SCAN).
    </para></listitem>
    <listitem><para>
     Extended Services (optional).
    </para></listitem>
    <listitem><para>
     Result-Set Delete (optional).
    </para></listitem>
    <listitem><para>
     Result-Set Sort (optional).
    </para></listitem>
    <listitem><para>
     Return Explain for SRU (optional).
    </para></listitem>
   </itemizedlist>
   <para>
    (more functions will be added in time to support as much of
    Z39.50-1995 as possible).
   </para>
  </sect1>
  <sect1 id="server.backend">
   <title>The Backend API</title>
   <para>
    The header file that you need to use the interface are in the
    <filename>include/yaz</filename> directory. It's called
    <filename>backend.h</filename>. It will include other files from
    the <filename>include/yaz</filename> directory, so you'll
    probably want to use the -I option of your compiler to tell it
    where to find the files. When you run
    <literal>make</literal> in the top-level &yaz; directory,
    everything you need to create your server is to link with the
    <filename>lib/libyaz.la</filename> library.
   </para>
  </sect1>
  <sect1 id="server.main">
   <title>Your main() Routine</title>
   <para>
    As mentioned, your <function>main()</function> routine can be quite brief.
    If you want to initialize global parameters, or read global configuration
    tables, this is the place to do it. At the end of the routine, you should
    call the function
   </para>
   <synopsis>
int statserv_main(int argc, char **argv,
                  bend_initresult *(*bend_init)(bend_initrequest *r),
                  void (*bend_close)(void *handle));
   </synopsis>
   <para>
    The third and fourth arguments are pointers to handlers. Handler
    <function>bend_init</function> is called whenever the server receives
    an Initialize Request, so it serves as a Z39.50 session initializer. The
    <function>bend_close</function> handler is called when the session is
    closed.
   </para>
   <para>
    <function>statserv_main</function> will establish listening sockets
    according to the parameters given. When connection requests are received,
    the event handler will typically <function>fork()</function> and
    create a sub-process to handle a new connection.
    Alternatively the server may be setup to create threads for each
    connection.
    If you do use global variables and forking, you should be aware, then,
    that these cannot be shared between associations, unless you explicitly
    disable forking by command line parameters.
   </para>
   <para>
    The server provides a mechanism for controlling some of its behavior
    without using command-line options. The function
   </para>
   <synopsis>
    statserv_options_block *statserv_getcontrol(void);
   </synopsis>
   <para>
    will return a pointer to a <literal>struct statserv_options_block</literal>
    describing the current default settings of the server. The structure
    contains these elements:
    <variablelist>
     <varlistentry>
      <term><literal>int dynamic</literal></term>
      <listitem><para>
       A boolean value, which determines whether the server
       will fork on each incoming request (TRUE), or not (FALSE). Default is
       TRUE. This flag is only read by UNIX-based servers (WIN32 based servers
       doesn't fork).
     </para></listitem>
     </varlistentry>
     <varlistentry>
      <term><literal>int threads</literal></term>
      <listitem><para>
       A boolean value, which determines whether the server
       will create a thread on each incoming request (TRUE), or not (FALSE).
       Default is FALSE. This flag is only read by UNIX-based servers
       that offer POSIX Threads support.
       WIN32-based servers always operate in threaded mode.
     </para></listitem>
     </varlistentry>
     <varlistentry>
      <term><literal>int inetd</literal></term>
      <listitem><para>
       A boolean value, which determines whether the server
       will operates under a UNIX INET daemon (inetd). Default is FALSE.
     </para></listitem>
     </varlistentry>
     <varlistentry>
      <term><literal>char logfile[ODR_MAXNAME+1]</literal></term>
      <listitem><para>File for diagnostic output (&quot;&quot;: stderr).
     </para></listitem>
     </varlistentry>
     <varlistentry>
      <term><literal>char apdufile[ODR_MAXNAME+1]</literal></term>
      <listitem><para>
       Name of file for logging incoming and outgoing APDUs
       (&quot;&quot;: don't log APDUs, &quot;-&quot;:
       <literal>stderr</literal>).
     </para></listitem>
     </varlistentry>
     <varlistentry>
      <term><literal>char default_listen[1024]</literal></term>
      <listitem><para>Same form as the command-line specification of
      listener address. &quot;&quot;: no default listener address.
      Default is to listen at &quot;tcp:@:9999&quot;. You can only
      specify one default listener address in this fashion.
     </para></listitem>
     </varlistentry>
     <varlistentry>
      <term><literal>enum oid_proto default_proto;</literal></term>
      <listitem><para>Either <literal>PROTO_Z3950</literal> or
      <literal>PROTO_SR</literal>.
      Default is <literal>PROTO_Z39_50</literal>.
     </para></listitem>
     </varlistentry>
     <varlistentry>
      <term><literal>int idle_timeout;</literal></term>
      <listitem><para>Maximum session idle-time, in minutes. Zero indicates
      no (infinite) timeout. Default is 15 minutes.
     </para></listitem>
     </varlistentry>
     <varlistentry>
      <term><literal>int maxrecordsize;</literal></term>
      <listitem><para>Maximum permissible record (message) size. Default
      is 64 MB. This amount of memory will only be allocated if a
      client requests a very large amount of records in one operation
      (or a big record).
      Set it to a lower number if you are worried about resource
      consumption on your host system.
     </para></listitem>
     </varlistentry>
     <varlistentry>
      <term><literal>char configname[ODR_MAXNAME+1]</literal></term>
      <listitem><para>Passed to the backend when a new connection is received.
     </para></listitem>
     </varlistentry>
     <varlistentry>
      <term><literal>char setuid[ODR_MAXNAME+1]</literal></term>
      <listitem><para>Set user id to the user specified, after binding
      the listener addresses.
     </para></listitem>
     </varlistentry>
     <varlistentry>
      <term>
       <literal>void (*bend_start)(struct statserv_options_block *p)</literal>
      </term>
      <listitem><para>Pointer to function which is called after the
      command line options have been parsed - but before the server
      starts listening.
      For forked UNIX servers this handler is called in the mother
      process; for threaded servers this handler is called in the
      main thread.
      The default value of this pointer is NULL in which case it
      isn't invoked by the frontend server.
      When the server operates as an NT service this handler is called
      whenever the service is started.
      </para></listitem>
     </varlistentry>
     <varlistentry>
      <term>
       <literal>void (*bend_stop)(struct statserv_options_block *p)</literal>
      </term>
      <listitem><para>Pointer to function which is called whenever the server
      has stopped listening for incoming connections. This function pointer
      has a default value of NULL in which case it isn't called.
      When the server operates as an NT service this handler is called
      whenever the service is stopped.
      </para></listitem>
     </varlistentry>
     <varlistentry>
      <term><literal>void *handle</literal></term>
      <listitem><para>User defined pointer (default value NULL).
      This is a per-server handle that can be used to specify "user-data".
      Do not confuse this with the session-handle as returned by bend_init.
      </para></listitem>
     </varlistentry>
    </variablelist>
   </para>
   <para>
    The pointer returned by <literal>statserv_getcontrol</literal> points to
    a static area. You are allowed to change the contents of the structure,
    but the changes will not take effect before you call
   </para>
   <synopsis>
void statserv_setcontrol(statserv_options_block *block);
   </synopsis>
   <note>
    <para>
     that you should generally update this structure before calling
     <function>statserv_main()</function>.
    </para>
   </note>
  </sect1>
  <sect1 id="server.backendfunctions">
   <title>The Backend Functions</title>
   <para>
    For each service of the protocol, the backend interface declares one or
    two functions. You are required to provide implementations of the
    functions representing the services that you wish to implement.
   </para>
   <sect2 id="server.init">
    <title>Init</title>
    <synopsis>
bend_initresult (*bend_init)(bend_initrequest *r);
    </synopsis>
    <para>
     This handler is called once for each new connection request, after
     a new process/thread has been created, and an Initialize Request has
     been received from the client. The pointer to the
     <function>bend_init</function> handler is passed in the call to
     <function>statserv_start</function>.
    </para>
    <para>
     This handler is also called when operating in SRU mode - when
     a connection has been made (even though SRU does not offer
     this service).
    </para>
    <para>
     Unlike previous versions of YAZ, the <function>bend_init</function> also
     serves as a handler that defines the Z39.50 services that the backend
     wish to support. Pointers to <emphasis>all</emphasis> service handlers,
     including search - and fetch must be specified here in this handler.
    </para>
    <para>
     The request - and result structures are defined as
    </para>
    <synopsis>
typedef struct bend_initrequest
{
    /** \brief user/name/password to be read */
    Z_IdAuthentication *auth;
    /** \brief encoding stream (for results) */
    ODR stream;
    /** \brief printing stream */
    ODR print;
    /** \brief decoding stream (use stream for results) */
    ODR decode;
    /** \brief reference ID */
    Z_ReferenceId *referenceId;
    /** \brief peer address of client */
    char *peer_name;

    /** \brief character set and language negotiation

    see include/yaz/z-charneg.h
    */
    Z_CharSetandLanguageNegotiation *charneg_request;

    /** \brief character negotiation response */
    Z_External *charneg_response;

    /** \brief character set (encoding) for query terms

    This is NULL by default. It should be set to the native character
    set that the backend assumes for query terms */
    char *query_charset;

    /** \brief whehter query_charset also applies to recors

    Is 0 (No) by default. Set to 1 (yes) if records is in the same
    character set as queries. If in doubt, use 0 (No).
    */
    int records_in_same_charset;

    char *implementation_id;
    char *implementation_name;
    char *implementation_version;

    /** \brief Z39.50 sort handler */
    int (*bend_sort)(void *handle, bend_sort_rr *rr);
    /** \brief SRU/Z39.50 search handler */
    int (*bend_search)(void *handle, bend_search_rr *rr);
    /** \brief SRU/Z39.50 fetch handler */
    int (*bend_fetch)(void *handle, bend_fetch_rr *rr);
    /** \brief SRU/Z39.50 present handler */
    int (*bend_present)(void *handle, bend_present_rr *rr);
    /** \brief Z39.50 extended services handler */
    int (*bend_esrequest) (void *handle, bend_esrequest_rr *rr);
    /** \brief Z39.50 delete result set handler */
    int (*bend_delete)(void *handle, bend_delete_rr *rr);
    /** \brief Z39.50 scan handler */
    int (*bend_scan)(void *handle, bend_scan_rr *rr);
    /** \brief Z39.50 segment facility handler */
    int (*bend_segment)(void *handle, bend_segment_rr *rr);
    /** \brief SRU explain handler */
    int (*bend_explain)(void *handle, bend_explain_rr *rr);
    /** \brief SRU scan handler */
    int (*bend_srw_scan)(void *handle, bend_scan_rr *rr);
    /** \brief SRU record update handler */
    int (*bend_srw_update)(void *handle, bend_update_rr *rr);

    /** \brief whether named result sets are supported (0=disable, 1=enable) */
    int named_result_sets;
} bend_initrequest;

typedef struct bend_initresult
{
    int errcode;               /* 0==OK */
    char *errstring;           /* system error string or NULL */
    void *handle;              /* private handle to the backend module */
} bend_initresult;
    </synopsis>
    <para>
     In general, the server frontend expects that the
     <literal>bend_*result</literal> pointer that you return is valid at
     least until the next call to a <literal>bend_* function</literal>.
     This applies to all of the functions described herein. The parameter
     structure passed to you in the call belongs to the server frontend, and
     you should not make assumptions about its contents after the current
     function call has completed. In other words, if you want to retain any
     of the contents of a request structure, you should copy them.
    </para>
    <para>
     The <literal>errcode</literal> should be zero if the initialization of
     the backend went well. Any other value will be interpreted as an error.
     The <literal>errstring</literal> isn't used in the current version, but
     one option would be to stick it in the initResponse as a VisibleString.
     The <literal>handle</literal> is the most important parameter. It should
     be set to some value that uniquely identifies the current session to
     the backend implementation. It is used by the frontend server in any
     future calls to a backend function.
     The typical use is to set it to point to a dynamically allocated state
     structure that is private to your backend module.
    </para>
    <para>
     The <literal>auth</literal> member holds the authentication information
     part of the Z39.50 Initialize Request. Interpret this if your serves
     requires authentication.
    </para>
    <para>
     The members <literal>peer_name</literal>,
     <literal>implementation_id</literal>,
     <literal>implementation_name</literal> and
     <literal>implementation_version</literal> holds
     DNS of client, ID of implementor, name
     of client (Z39.50) implementation - and version.
    </para>
    <para>
     The <literal>bend_</literal> - members are set to NULL when
     <function>bend_init</function> is called. Modify the pointers by
     setting them to point to backend functions.
    </para>
   </sect2>
   <sect2 id="server.search.retrieve">
    <title>Search and Retrieve</title>
    <para>
     We now describe the handlers that are required to support search -
     and retrieve. You must support two functions - one for search - and one
     for fetch (retrieval of one record). If desirable you can provide a
     third handler which is called when a present request is received which
     allows you to optimize retrieval of multiple-records.
    </para>
    <synopsis>
int (*bend_search) (void *handle, bend_search_rr *rr);

typedef struct {
    char *setname;             /* name to give to this set */
    int replace_set;           /* replace set, if it already exists */
    int num_bases;             /* number of databases in list */
    char **basenames;          /* databases to search */
    Z_ReferenceId *referenceId;/* reference ID */
    Z_Query *query;            /* query structure */
    ODR stream;                /* encode stream */
    ODR decode;                /* decode stream */
    ODR print;                 /* print stream */

    bend_request request;
    bend_association association;
    int *fd;
    int hits;                  /* number of hits */
    int errcode;               /* 0==OK */
    char *errstring;           /* system error string or NULL */
    Z_OtherInformation *search_info; /* additional search info */
    char *srw_sortKeys;        /* holds SRU/SRW sortKeys info */
    char *srw_setname;         /* holds SRU/SRW generated resultsetID */
    int *srw_setnameIdleTime;  /* holds SRU/SRW life-time */
    int estimated_hit_count;   /* if hit count is estimated */
    int partial_resultset;     /* if result set is partial */
} bend_search_rr;
    </synopsis>
    <para>
     The <function>bend_search</function> handler is a fairly close
     approximation of a protocol Z39.50 Search Request - and Response PDUs
     The <literal>setname</literal> is the resultSetName from the protocol.
     You are required to establish a mapping between the set name and whatever
     your backend database likes to use.
     Similarly, the <literal>replace_set</literal> is a boolean value
     corresponding to the resultSetIndicator field in the protocol.
     <literal>num_bases/basenames</literal> is a length of/array of character
     pointers to the database names provided by the client.
     The <literal>query</literal> is the full query structure as defined in
     the protocol ASN.1 specification.
     It can be either of the possible query types, and it's up to you to
     determine if you can handle the provided query type.
     Rather than reproduce the C interface here, we'll refer you to the
     structure definitions in the file
     <filename>include/yaz/z-core.h</filename>. If you want to look at the
     attributeSetId OID of the RPN query, you can either match it against
     your own internal tables, or you can use the <link linkend="tools.oid">
     OID tools</link>.
    </para>
    <para>
     The structure contains a number of hits, and an
     <literal>errcode/errstring</literal> pair. If an error occurs
     during the search, or if you're unhappy with the request, you should
     set the errcode to a value from the BIB-1 diagnostic set. The value
     will then be returned to the user in a nonsurrogate diagnostic record
     in the response. The <literal>errstring</literal>, if provided, will
     go in the addinfo field. Look at the protocol definition for the
     defined error codes, and the suggested uses of the addinfo field.
    </para>
    <para>
     The <function>bend_search</function> handler is also called when
     the frontend server receives a SRU SearchRetrieveRequest.
     For SRU, a CQL query is usually provided by the client.
     The CQL query is available as part of <literal>Z_Query</literal>
     structure (note that CQL is now part of Z39.50 via an external).
     To support CQL in existing implementations that only do Type-1,
     we refer to the CQL-to-PQF tool described
     <link linkend="cql.to.pqf">here</link>.
    </para>
    <para>
     To maintain backwards compatibility, the frontend server
     of yaz always assume that error codes are BIB-1 diagnostics.
     For SRU operation, a Bib-1 diagnostic code is mapped to
     SRU diagnostic.
    </para>
    <synopsis>
int (*bend_fetch) (void *handle, bend_fetch_rr *rr);

typedef struct bend_fetch_rr {
    char *setname;             /* set name */
    int number;                /* record number */
    Z_ReferenceId *referenceId;/* reference ID */
    Odr_oid *request_format;        /* format, transfer syntax (OID) */
    Z_RecordComposition *comp; /* Formatting instructions */
    ODR stream;                /* encoding stream - memory source if req */
    ODR print;                 /* printing stream */

    char *basename;            /* name of database that provided record */
    int len;                   /* length of record or -1 if structured */
    char *record;              /* record */
    int last_in_set;           /* is it?  */
    Odr_oid *output_format;        /* response format/syntax (OID) */
    int errcode;               /* 0==success */
    char *errstring;           /* system error string or NULL */
    int surrogate_flag;        /* surrogate diagnostic */
    char *schema;              /* string record schema input/output */
} bend_fetch_rr;
    </synopsis>
    <para>
     The frontend server calls the <function>bend_fetch</function> handler
     when it needs database records to fulfill a Z39.50 Search Request, a
     Z39.50 Present Request or a SRU SearchRetrieveRequest.
     The <literal>setname</literal> is simply the name of the result set
     that holds the reference to the desired record.
     The <literal>number</literal> is the offset into the set (with 1
     being the first record in the set). The <literal>format</literal> field
     is the record format requested by the client (See
     <xref linkend="tools.oid"/>).
     A value of NULL for <literal>format</literal> indicates that the
     client did not request a specific format.
     The <literal>stream</literal> argument is an &odr; stream which
     should be used for allocating space for structured data records.
     The stream will be reset when all records have been assembled, and
     the response package has been transmitted.
     For unstructured data, the backend is responsible for maintaining a
     static or dynamic buffer for the record between calls.
    </para>
    <para>
     If a SRU SearchRetrieveRequest is received by the frontend server,
     the <literal>referenceId</literal> is NULL and the
     <literal>format</literal> (transfer syntax) is the OID for XML.
     The schema for SRU is stored in both the
     <literal>Z_RecordComposition</literal>
     structure and <literal>schema</literal> (simple string).
    </para>
    <para>
     In the structure, the <literal>basename</literal> is the name of the
     database that holds the
     record. <literal>len</literal> is the length of the record returned, in
     bytes, and <literal>record</literal> is a pointer to the record.
     <literal>last_in_set</literal> should be nonzero only if the record
     returned is the last one in the given result set.
     <literal>errcode</literal> and <literal>errstring</literal>, if
     given, will be interpreted as a global error pertaining to the
     set, and will be returned in a non-surrogate-diagnostic.
     If you wish to return the error as a surrogate-diagnostic
     (local error) you can do this by setting
     <literal>surrogate_flag</literal> to 1 also.
    </para>
    <para>
     If the <literal>len</literal> field has the value -1, then
     <literal>record</literal> is assumed to point to a constructed data
     type. The <literal>format</literal> field will be used to determine
     which encoder should be used to serialize the data.
    </para>
    <note>
     <para>
      If your backend generates structured records, it should use
      <function>odr_malloc()</function> on the provided stream for allocating
      data: This allows the frontend server to keep track of the record sizes.
     </para>
    </note>
    <para>
     The <literal>format</literal> field is mapped to an object identifier
     in the direct reference of the resulting EXTERNAL representation
     of the record.
    </para>
    <note>
     <para>
      The current version of &yaz; only supports the direct reference mode.
     </para>
    </note>
    <synopsis>
int (*bend_present) (void *handle, bend_present_rr *rr);

typedef struct {
    char *setname;             /* set name */
    int start;
    int number;                /* record number */
    Odr_oid *format;           /* format, transfer syntax (OID) */
    Z_ReferenceId *referenceId;/* reference ID */
    Z_RecordComposition *comp; /* Formatting instructions */
    ODR stream;                /* encoding stream - memory source if required */
    ODR print;                 /* printing stream */
    bend_request request;
    bend_association association;

    int hits;                  /* number of hits */
    int errcode;               /* 0==OK */
    char *errstring;           /* system error string or NULL */
} bend_present_rr;
    </synopsis>
    <para>
     The <function>bend_present</function> handler is called when
     the server receives a Z39.50 Present Request.
     The <literal>setname</literal>,
     <literal>start</literal> and <literal>number</literal> is the
     name of the result set - start position - and number of records to
     be retrieved respectively. <literal>format</literal> and
     <literal>comp</literal> is the preferred transfer syntax and element
     specifications of the present request.
    </para>
    <para>
     Note that this is handler serves as a supplement for
     <function>bend_fetch</function> and need not to be defined in order to
     support search - and retrieve.
    </para>
   </sect2>
   <sect2 id="server.delete">
    <title>Delete</title>
    <para>
     For back-ends that supports delete of a result set only one handler
     must be defined.
    </para>
    <synopsis>
int (*bend_delete)(void *handle, bend_delete_rr *rr);

typedef struct bend_delete_rr {
    int function;
    int num_setnames;
    char **setnames;
    Z_ReferenceId *referenceId;
    int delete_status;      /* status for the whole operation */
    int *statuses;          /* status each set - indexed as setnames */
    ODR stream;
    ODR print;
} bend_delete_rr;
    </synopsis>
    <note>
     <para>
      The delete set function definition is rather primitive, mostly because
      we have had no practical need for it as of yet. If someone wants
      to provide a full delete service, we'd be happy to add the
      extra parameters that are required. Are there clients out there
      that will actually delete sets they no longer need?
     </para>
    </note>
   </sect2>
   <sect2 id="server.scan">
    <title>Scan</title>
    <para>
     For servers that wish to offer the scan service one handler
     must be defined.
    </para>
    <synopsis>
int (*bend_scan)(void *handle, bend_scan_rr *rr);

typedef enum {
    BEND_SCAN_SUCCESS,  /* ok */
    BEND_SCAN_PARTIAL   /* not all entries could be found */
} bend_scan_status;

typedef struct bend_scan_rr {
    int num_bases;      /* number of elements in databaselist */
    char **basenames;   /* databases to search */
    Odr_oid *attributeset;
    Z_ReferenceId *referenceId; /* reference ID */
    Z_AttributesPlusTerm *term;
    ODR stream;         /* encoding stream - memory source if required */
    ODR print;          /* printing stream */

    int *step_size;     /* step size */
    int term_position;  /* desired index of term in result list/returned */
    int num_entries;    /* number of entries requested/returned */

    /* scan term entries. The called handler does not have
       to allocate this. Size of entries is num_entries (see above) */
    struct scan_entry *entries;
    bend_scan_status status;
    int errcode;
    char *errstring;
    char *scanClause;   /* CQL scan clause */
    char *setname;      /* Scan in result set (NULL if omitted) */
} bend_scan_rr;
    </synopsis>
   <para>
    This backend server handles both Z39.50 scan
    and SRU scan. In order for a handler to distinguish between SRU (CQL) scan
    Z39.50 Scan , it must check for a non-NULL value of
    <literal>scanClause</literal>.
   </para>
   <note>
    <para>
     if designed today, it would be a choice using a union or similar,
     but that would break binary compatibility with existing servers.
    </para>
    </note>
   </sect2>
  </sect1>
  <sect1 id="server.invocation">
   <title>Application Invocation</title>
   <para>
    The finished application has the following
    invocation syntax (by way of <function>statserv_main()</function>):
   </para>
   &gfs-synopsis;
   <para>
    The options are:
    &gfs-options;
   </para>
   <para>
    A listener specification consists of a transport mode followed by a
    colon (:) followed by a listener address. The transport mode is
    either <literal>tcp</literal>, <literal>unix:</literal> or
    <literal>ssl</literal>.
   </para>
   <para>
    For TCP and SSL, an address has the form
   </para>
   <synopsis>
    hostname | IP-number [: portnumber]
   </synopsis>
   <para>
    The port number defaults to 210 (standard Z39.50 port).
   </para>
   <para>
    For UNIX, the address is the filename of socket.
   </para>
   <para>
    For TCP/IP and SSL, the special hostnames <literal>@</literal>,
    maps to <literal>IN6ADDR_ANY_INIT</literal> with
    IPV4 binding as well (bindv6only=0),
    The special hostname <literal>@4</literal> binds to
    <literal>INADDR_ANY</literal> (IPV4 only listener).
    The special hostname <literal>@6</literal> binds to
    <literal>IN6ADDR_ANY_INIT</literal> with bindv6only=1 (IPV6 only listener).
   </para>
   <example id="server.example.running.unix">
    <title>Running the GFS on Unix</title>
    <para>
     Assuming the server application <replaceable>appname</replaceable> is
     started as root, the following will make it listen on port 210.
     The server will change identity to <literal>nobody</literal>
     and write its log to <filename>/var/log/app.log</filename>.
     <screen>
      application -l /var/log/app.log -u nobody tcp:@:210
     </screen>
    </para>
    <para>
     The server will accept Z39.50 requests and offer SRU service on port 210.
    </para>
   </example>
   <example id="server.example.apache.sru">
    <title>Setting up Apache as SRU Frontend</title>
    <para>
     If you use <ulink url="&url.apache;">Apache</ulink>
     as your public web server and want to offer HTTP port 80
     access to the YAZ server on 210, you can use the
     <ulink url="&url.apache.directive.proxypass;">
      <literal>ProxyPass</literal></ulink>
     directive.
     If you have virtual host
     <literal>srw.mydomain</literal> you can use the following directives
     in Apache's httpd.conf:
     <screen>
      &lt;VirtualHost *>
       ErrorLog /home/srw/logs/error_log
       TransferLog /home/srw/logs/access_log
       ProxyPass / http://srw.mydomain:210/
      &lt;/VirtualHost>
     </screen>
    </para>
    <para>
     The above for the Apache 1.3 series.
    </para>
   </example>
   <example id="server.example.local.access">
    <title>Running a server with local access only</title>
    <para>
     Servers that is only being accessed from the local host should listen
     on UNIX file socket rather than a Internet socket. To listen on
     <filename>/tmp/mysocket</filename> start the server as follows:
     <screen>
      application unix:/tmp/mysocket
     </screen>
    </para>
   </example>
  </sect1>
  <sect1 id="server.vhosts">
   <title>GFS Configuration and Virtual Hosts</title>
   &gfs-virtual;
  </sect1>
 </chapter>
 <chapter id="asn">
  <title>The Z39.50 ASN.1 Module</title>
  <sect1 id="asn.introduction">
   <title>Introduction</title>
   <para>
    The &asn; module provides you with a set of C struct definitions for the
    various PDUs of the Z39.50 protocol, as well as for the complex types
    appearing within the PDUs. For the primitive data types, the C
    representation often takes the form of an ordinary C language type,
    such as <literal>Odr_int</literal> which is equivalent to an integral
    C integer. For ASN.1 constructs that have no direct
    representation in C, such as general octet strings and bit strings,
    the &odr; module (see section <link linkend="odr">The ODR Module</link>)
    provides auxiliary definitions.
   </para>
   <para>
    The &asn; module is located in sub directory <filename>z39.50</filename>.
    There you'll find C files that implements encoders and decoders for the
    Z39.50 types. You'll also find the protocol definitions:
    <filename>z3950v3.asn</filename>, <filename>esupdate.asn</filename>,
    and others.
   </para>
  </sect1>
  <sect1 id="asn.preparing">
   <title>Preparing PDUs</title>
   <para>
    A structure representing a complex ASN.1 type doesn't in itself contain the
    members of that type. Instead, the structure contains
    <emphasis>pointers</emphasis> to the members of the type.
    This is necessary, in part, to allow a mechanism for specifying which
    of the optional structure (SEQUENCE) members are present, and which
    are not. It follows that you will need to somehow provide space for
    the individual members of the structure, and set the pointers to
    refer to the members.
   </para>
   <para>
    The conversion routines don't care how you allocate and maintain your
    C structures - they just follow the pointers that you provide.
    Depending on the complexity of your application, and your personal
    taste, there are at least three different approaches that you may take
    when you allocate the structures.
   </para>
   <para>
    You can use static or automatic local variables in the function that
    prepares the PDU. This is a simple approach, and it provides the most
    efficient form of memory management. While it works well for flat
    PDUs like the InitReqest, it will generally not be sufficient for say,
    the generation of an arbitrarily complex RPN query structure.
   </para>
   <para>
    You can individually create the structure and its members using the
    <function>malloc(2)</function> function. If you want to ensure that
    the data is freed when it is no longer needed, you will have to
    define a function that individually releases each member of a
    structure before freeing the structure itself.
   </para>
   <para>
    You can use the <function>odr_malloc()</function> function (see
    <xref linkend="odr.use"/> for details). When you use
    <function>odr_malloc()</function>, you can release all of the
    allocated data in a single operation, independent of any pointers and
    relations between the data. <function>odr_malloc()</function> is based on a
    &quot;nibble-memory&quot;
    scheme, in which large portions of memory are allocated, and then
    gradually handed out with each call to <function>odr_malloc()</function>.
    The next time you call <function>odr_reset()</function>, all of the
    memory allocated since the last call is recycled for future use (actually,
    it is placed on a free-list).
   </para>
   <para>
    You can combine all of the methods described here. This will often be
    the most practical approach. For instance, you might use
    <function>odr_malloc()</function> to allocate an entire structure and
    some of its elements, while you leave other elements pointing to global
    or per-session default variables.
   </para>
   <para>
    The &asn; module provides an important aid in creating new PDUs. For
    each of the PDU types (say, <function>Z_InitRequest</function>), a
    function is provided that allocates and initializes an instance of
    that PDU type for you. In the case of the InitRequest, the function is
    simply named <function>zget_InitRequest()</function>, and it sets up
    reasonable default value for all of the mandatory members. The optional
    members are generally initialized to null pointers. This last aspect
    is very important: it ensures that if the PDU definitions are
    extended after you finish your implementation (to accommodate
    new versions of the protocol, say), you won't get into trouble with
    uninitialized pointers in your structures. The functions use
    <function>odr_malloc()</function> to
    allocate the PDUs and its members, so you can free everything again with a
    single call to <function>odr_reset()</function>. We strongly recommend
    that you use the <literal>zget_*</literal>
    functions whenever you are preparing a PDU (in a C++ API, the
    <literal>zget_</literal>
    functions would probably be promoted to constructors for the
    individual types).
   </para>
   <para>
   The prototype for the individual PDU types generally look like this:
   </para>
   <synopsis>
    Z_&lt;type> *zget_&lt;type>(ODR o);
   </synopsis>
   <para>
    eg.:
   </para>
   <synopsis>
    Z_InitRequest *zget_InitRequest(ODR o);
   </synopsis>
   <para>
   The &odr; handle should generally be your encoding stream, but it
    needn't be.
   </para>
   <para>
    As well as the individual PDU functions, a function
    <function>zget_APDU()</function> is provided, which allocates
    a top-level Z-APDU of the type requested:
   </para>
   <synopsis>
    Z_APDU *zget_APDU(ODR o, int which);
   </synopsis>
   <para>
    The <varname>which</varname> parameter is (of course) the discriminator
    belonging to the <varname>Z_APDU</varname> <literal>CHOICE</literal> type.
    All of the interface described here is provided by the &asn; module, and
    you access it through the <filename>proto.h</filename> header file.
   </para>
  </sect1>
  <sect1 id="asn.external">
   <title>EXTERNAL Data</title>
   <para>
    In order to achieve extensibility and adaptability to different
    application domains, the new version of the protocol defines many
    structures outside of the main ASN.1 specification, referencing them
    through ASN.1 EXTERNAL constructs. To simplify the construction and
    access to the externally referenced data, the &asn; module defines a
    specialized version of the EXTERNAL construct, called
    <literal>Z_External</literal>.It is defined thus:
   </para>
   <screen>
typedef struct Z_External
{
    Odr_oid *direct_reference;
    int *indirect_reference;
    char *descriptor;
    enum
    {
        /* Generic types */
        Z_External_single = 0,
        Z_External_octet,
        Z_External_arbitrary,

        /* Specific types */
        Z_External_SUTRS,
        Z_External_explainRecord,
        Z_External_resourceReport1,
        Z_External_resourceReport2

    ...

    } which;
    union
    {
        /* Generic types */
        Odr_any *single_ASN1_type;
        Odr_oct *octet_aligned;
        Odr_bitmask *arbitrary;

        /* Specific types */
        Z_SUTRS *sutrs;
        Z_ExplainRecord *explainRecord;
        Z_ResourceReport1 *resourceReport1;
        Z_ResourceReport2 *resourceReport2;

        ...

    } u;
} Z_External;
   </screen>
   <para>
    When decoding, the &asn; module will attempt to determine which
    syntax describes the data by looking at the reference fields
    (currently only the direct-reference). For ASN.1 structured data, you
    need only consult the <literal>which</literal> field to determine the
    type of data. You can the access  the data directly through the union.
    When constructing data for encoding, you set the union pointer to point
    to the data, and set the <literal>which</literal> field accordingly.
    Remember also to set the direct (or indirect) reference to the correct
    OID for the data type.
    For non-ASN.1 data such as MARC records, use the
    <literal>octet_aligned</literal> arm of the union.
   </para>
   <para>
    Some servers return ASN.1 structured data values (eg. database
    records) as BER-encoded records placed in the
    <literal>octet-aligned</literal> branch of the EXTERNAL CHOICE.
    The ASN-module will <emphasis>not</emphasis> automatically decode
    these records. To help you decode the records in the application, the
    function
   </para>
   <screen>
   Z_ext_typeent *z_ext_gettypebyref(const oid *oid);
   </screen>
   <para>
    Can be used to retrieve information about the known, external data
    types. The function return a pointer to a static area, or NULL, if no
    match for the given direct reference is found. The
    <literal>Z_ext_typeent</literal>
    is defined as:
   </para>
   <screen>
typedef struct Z_ext_typeent
{
    int oid[OID_SIZE]; /* the direct-reference OID. */
    int what;          /* discriminator value for the external CHOICE */
    Odr_fun fun;       /* decoder function */
} Z_ext_typeent;
   </screen>
   <para>
    The <literal>what</literal> member contains the
    <literal>Z_External</literal> union discriminator value for the
    given type: For the SUTRS record syntax, the value would be
    <literal>Z_External_sutrs</literal>.
    The <literal>fun</literal> member contains a pointer to the
    function which encodes/decodes the given type. Again, for the SUTRS
    record syntax, the value of <literal>fun</literal> would be
    <literal>z_SUTRS</literal> (a function pointer).
   </para>
   <para>
    If you receive an EXTERNAL which contains an octet-string value that
    you suspect of being an ASN.1-structured data value, you can use
    <literal>z_ext_gettypebyref</literal> to look for the provided
    direct-reference.
    If the return value is different from NULL, you can use the provided
    function to decode the BER string (see <xref linkend="odr.use"/>
    ).
   </para>
   <para>
    If you want to <emphasis>send</emphasis> EXTERNALs containing
    ASN.1-structured values in the occtet-aligned branch of the CHOICE, this
    is possible too. However, on the encoding phase, it requires a somewhat
    involved juggling around of the various buffers involved.
   </para>
   <para>
    If you need to add new, externally defined data types, you must update
    the struct above, in the source file <filename>prt-ext.h</filename>, as
    well as the encoder/decoder in the file <filename>prt-ext.c</filename>.
    When changing the latter, remember to update both the
    <literal>arm</literal> arrary and the list
    <literal>type_table</literal>, which drives the CHOICE biasing that
    is necessary to tell the different, structured types apart
    on decoding.
   </para>
   <note>
    <para>
     Eventually, the EXTERNAL processing will most likely
     automatically insert the correct OIDs or indirect-refs. First,
     however, we need to determine how application-context management
     (specifically the presentation-context-list) should fit into the
     various modules.
    </para>
   </note>
  </sect1>
  <sect1 id="asn.pdu">
   <title>PDU Contents Table</title>
   <para>
    We include, for reference, a listing of the fields of each top-level
    PDU, as well as their default settings.
   </para>
   <table frame="top" id="asn.default.initialize.request">
    <title>Default settings for PDU Initialize Request</title>
    <tgro