Initial revision

[catagits/fcgi2.git] / doc / fcgi-devel-kit.gut

FastCGI Developer's Kit
/fastcgi/words
fcgi-hd.gif
[FastCGI]
<center>FastCGI Developer's Kit</center>

<!--Copyright (c) 1996 Open Market, Inc.                                    -->
<!--See the file "LICENSE.TERMS" for information on usage and redistribution-->
<!--of this file, and for a DISCLAIMER OF ALL WARRANTIES.                   -->

<center>
Mark R. Brown<br>
Open Market, Inc.<br>
<p>

Document Version: 1.08<br>
11 June 1996<br>
</center>
<p>

<h5 align=center>
Copyright &copy; 1996 Open Market, Inc.  245 First Street, Cambridge,
  MA 02142 U.S.A.<br>
Tel: 617-621-9500 Fax: 617-621-1703 URL:
  <a href="http://www.openmarket.com/">http://www.openmarket.com/</a><br>
$Id: fcgi-devel-kit.gut,v 1.1 1997/09/16 15:36:26 stanleyg Exp $ <br>
</h5>
<hr>

<ul type=square>
  <li><a HREF = "#S1">1. Introduction</a>
  <li><a HREF = "#S2">2. Getting started</a>
  <li><a HREF = "#S3">3. Writing applications</a>
  <ul type=square>
    <li><a HREF = "#S3.1">3.1 Using the <tt>fcgi_stdio</tt> library</a>
    <li><a HREF = "#S3.2">3.2 Using the <tt>fcgiapp</tt> library</a>
    <li><a HREF = "#S3.3">3.3 Using Perl and Tcl</a>
    <li><a HREF = "#S3.4">3.4 Using Java</a>
  </ul>
  <li><a HREF = "#S4">4. Running applications</a>
  <ul type=square>
    <li><a HREF = "#S4.1">4.1 Using a Web server that supports FastCGI</a>
    <li><a HREF = "#S4.2">4.2 Using <tt>cgi-fcgi</tt> with any Web server</a>
  </ul>
  <li><a HREF = "#S5">5. Known problems</a>
  <li><a HREF = "#S6">6. Getting support</a>
</ul>

<hr>



<h3><a NAME = "S1">1. Introduction</a></h3>

FastCGI is an open extension to CGI that provides high performance
for all Internet applications without the penalties of Web server
APIs.<p>

FastCGI is designed to be layered on top of existing Web server
APIs.  For instance, the <tt>mod_fastcgi</tt> Apache module adds
FastCGI support to the Apache server.  FastCGI can also be used,
with reduced functionality and reduced performance, on any Web server
that supports CGI.<p>

This FastCGI Developer's Kit is designed to make developing FastCGI
applications easy.  The kit currently supports FastCGI
applications written in C/C++, Perl, Tcl, and Java.<p>

This document:

<ul type=square>
  <li>Describes how to configure and build the
      kit for your development platform.

  <li>Tells how to write applications using the
      libraries in the kit.

  <li>Tells how to run applications using Web servers that support
      FastCGI or using any Web server and <tt>cgi-fcgi</tt>.
</ul>

The kit includes a <a href="fastcgi-whitepaper/fastcgi.htm">technical white
paper</a>, <tt>doc/fastcgi-whitepaper/fastcgi.htm</tt>.
You should read at least the first three sections of the
technical white paper before starting to write FastCGI applications.
The <a href="fcgi-perf.htm">performance paper</a> will help you understand
how application design affects performance with FastCGI.<p>

The <a href ="fcgi-spec.html">FastCGI Specification</a>,
<tt>doc/fcgi-spec.html</tt>, defines the interface between a FastCGI
application and a Web server that supports FastCGI.  The software in
the kit implements the specification.  You don't need to
read the specification in order to write applications.<p>

Additional information is provided in the 
<a href ="http://www.fastcgi.com/words/FAQ.htm">FAQ</a> document, which
contains frequently asked questions about application development 
using FastCGI, as well as some general information.<p>

Experience with CGI programming will be extremely valuable in writing FastCGI
applications.  If you don't have enough experience with CGI programming,
you should read one of the popular books on the topic or study the
<a href = "http://hoohoo.ncsa.uiuc.edu/cgi/">NCSA CGI page</a>.
For a more formal treatment of CGI/1.1 see the
<a href = "http://ds.internic.net/internet-drafts/draft-robinson-www-interface-01.txt">Internet Draft CGI 1.1 Specification</a>.<p>



<h3><a NAME = "S2">2. Getting started</a></h3>

The kit is a compressed tar (tar.Z) file,
distributed via the
<a href = "http://www.fastcgi.com/applibs/">www.fastcgi.com/applibs</a>
Web page.  Unpacking the tar file creates a new directory
<tt>fcgi-devel-kit</tt>.<p>

Open the kit's index page, <tt>fcgi-devel-kit/index.html</tt>, using
the "Open File" command in your Web browser.  The index page gives you
an overview of the kit structure and helps you navigate the kit. The
index page also contains links that run some example applications, but
the applications won't work when index.html is opened using the "Open
File" command because they aren't aren't being accessed through a Web
server.<p>

In order to use the kit in earnest you'll need a Web server that you
control, a Web server running with your user ID.  The Web server will
be starting FastCGI applications that you will need to debug; this
will be a lot more convenient for you if these processes run with your
user ID.  It is best to have a Web server that supports FastCGI.
<a href = "#S4">Section 4</a> discusses Web server issues.<p>

If you can, keep the kit on a file system accessible from your
personal workstation, do your builds on your workstation, and run your
Web server on your workstation.  If that's not possible, arrange a
configuration such that the kit is accessible from the machine that's
going to run your Web server, and build the kit and your applications
on a machine that's configured exactly the same way (same processor
architecture, operating system, etc.) as the machine that's going to
run your Web server.<p>

To build the kit you execute this sequence of commands
in the <tt>fcgi-devel-kit</tt> directory:<p>

<pre>
    % ./configure
    % make
</pre>

We've built and exercised the kit on these platforms
(listed in alphabetical order):<p>

<ul type=square>
    <li>
        BSD/OS 1.1 (Intel Pentium), gcc
    <li>
        Digital UNIX V3.2 148 (Alpha), gcc/cc
    <li>
        Hewlett-Packard HP-UX A.09.05 C and B.10.01 A (PA-RISC), gcc/cc
    <li>
        IBM AIX 1 4 (RS/6000), gcc
    <li>
        Silicon Graphics IRIX 5.3 11091812 (MIPS), gcc
    <li>
        Sun Solaris 2.4 and 2.5 (SPARC), gcc/cc
    <li>
        Sun SunOS 4.1.4 (SPARC), gcc
</ul>

Once you've built the kit, follow the directions in
<a href = "#S4">Section 4</a> to bring up your Web server
and run the example applications.<p>



<h3><a NAME = "S3">3. Writing applications</a></h3>


<h4><a NAME = "S3.1">3.1 Using the <tt>fcgi_stdio</tt> library</a>
</h4>

The <tt>fcgi_stdio</tt> library provides
the easiest transition for C CGI programs and C CGI
programmers to FastCGI.  Using this library your application
can run using either CGI or FastCGI, with the same binary
for both situations.<p>

To introduce the <tt>fcgi_stdio</tt> library
we give a pair of examples: a tiny CGI program and the translation of this
program to FastCGI.  These two example programs are included in the kit.<p>

The CGI program is <tt>examples/tiny-cgi.c</tt>:<p>

<pre>
    #include &lt;stdio.h&gt;
    #include &lt;stdlib.h&gt;

    void main(void)
    {
        int count = 0;
        printf("Content-type: text/html\r\n"
               "\r\n"
               "&lt;title&gt;CGI Hello!&lt;/title&gt;"
               "&lt;h1&gt;CGI Hello!&lt;/h1&gt;"
               "Request number %d running on host &lt;i&gt;%s&lt;/i&gt;\n",
               ++count, getenv("SERVER_NAME"));
    }
</pre>

The key features of this tiny CGI program are:<p>

<ul type = square>
    <li>
        The program sends data to the Web server by writing to
        <tt>stdout</tt>, using <tt>printf</tt> in this example.  The CGI
        program first sends a <tt>Content-type</tt> header, then a
        small HTML document.  The program includes <tt>stdio.h</tt> in
        order to get access to the <tt>printf</tt> function.<p>
    <li>
        The program obtains parameters provided by the Web server by
        reading environment variables.  The CGI program reads the
        <tt>SERVER_NAME</tt> variable using <tt>getenv</tt> and
        includes the value in the HTML document.  The program includes
        <tt>stdlib.h</tt> in order to get access to the
        <tt>getenv</tt> function.<p>
</ul>

The <tt>count</tt> variable is degenerate in this example;
the CGI program runs a single request, so the request number
is always one.  This variable will be more interesting
in the FastCGI example.<p>

<a NAME = "S3.1.1">The</a>
corresponding FastCGI program is <tt>examples/tiny-fcgi.c</tt>:<p>

<pre>
    #include "fcgi_stdio.h"
    #include &lt;stdlib.h&gt;

    void main(void)
    {
        int count = 0;
        while(FCGI_Accept() &gt;= 0)
            printf("Content-type: text/html\r\n"
                   "\r\n"
                   "&lt;title&gt;FastCGI Hello!&lt;/title&gt;"
                   "&lt;h1&gt;FastCGI Hello!&lt;/h1&gt;"
                   "Request number %d running on host &lt;i&gt;%s&lt;/i&gt;\n",
                    ++count, getenv("SERVER_NAME"));
    }
</pre>

The key features of this tiny FastCGI program are:<p>

<ul type = square>
    <li>
        The program is structured as a loop that begins by calling the
        function <tt>FCGI_Accept</tt>.  The <tt>FCGI_Accept</tt>
        function blocks until a new request arrives for the program to
        execute.  The program includes <tt>fcgi_stdio.h</tt> in order
        to get access to the <tt>FCGI_Accept</tt> function.<p>
    <li>
        Within the loop, <tt>FCGI_Accept</tt> creates a
        CGI-compatible world.  <tt>printf</tt> and <tt>getenv</tt>
        operate just as in the CGI program.  <tt>stdin</tt> and
        <tt>stderr</tt>, not used by this tiny program, also
        operate just as in a CGI program.<p>
</ul>

The <tt>count</tt> variable increments each time through the loop, so
the program displays a new request number each time.  You can use
the reload button in your browser to demonstrate this, once you've
got the program built and running.<p>


<h4>Building the program</h4>

If you can build <tt>examples/tiny-cgi.c</tt>,
it will be straightforward for you to build
<tt>examples/tiny-fcgi.c</tt>.  You need to:<p>

<ul type = square>
    <li>
        Add the directory containing the
        <tt>fcgi_stdio.h</tt> header to the compiler's include search
        path.  The kit calls this directory <tt>include</tt>.<p>
    <li>
        Add the library <tt>libfcgi.a</tt> to the linker's command
        line so that it will be searched when linking.  The
        <tt>libfcgi.a</tt> library implements the functions defined in
        <tt>fcgi_stdio.h</tt>.  The kit calls the directory containing
        this library <tt>libfcgi</tt>.<p>
    <li>
        Determine whether or not the linker on your
        platform searches the Berkeley socket
        library by default, and if not, add linker directives to
        force this search.<p>
</ul>

See <tt>examples/Makefile</tt> (created by <tt>configure</tt>)
for a Makefile that builds both
programs.  Autoconf handles the platform-dependent linking issues; to
see how, examine <tt>configure.in</tt> and
<tt>examples/Makefile.in</tt>.<p>


<h4>Running the program</h4>

<a href = "#S4">Section 4</a> is all about
how to run FastCGI applications.<p>

You can use CGI to run application binaries built with the
<tt>fcgi_stdio</tt> library.  The <tt>FCGI_Accept</tt>
function tests its environment to determine how the application was
invoked.  If it was invoked as a CGI program, the first
call to FCGI_Accept is essentially a no-op and the second call
returns <tt>-1</tt>.  In effect, the request loop disappears.<p>

Of course, when a FastCGI application is run using CGI it does not
get the benefits of FastCGI.  For instance, the application exits
after servicing a single request, so it cannot maintain
cached information.<p>


<h4>Implementation details</h4>

<tt>fcgi_stdio.h</tt> works by first including <tt>stdio.h</tt>, then
defining macros to replace essentially all of the types and procedures
defined in <tt>stdio.h</tt>.  (<tt>stdio.h</tt> defines a few
procedures that have nothing to do with <tt>FILE *</tt>, such as
<tt>sprintf</tt> and <tt>sscanf</tt>; <tt>fcgi_stdio.h</tt> doesn't
replace these.)  For instance, <tt>FILE</tt> becomes
<tt>FCGI_FILE</tt> and <tt>printf</tt> becomes <tt>FCGI_printf</tt>.
You'll only see these new names if you read <tt>fcgi_stdio.h</tt> or
examine your C source code after preprocessing.<p>

Here are some
consequences of this implementation technique:<p>

<ul type = square>
    <li>
        On some platforms the implementation will break if you include
        <tt>stdio.h</tt> after including <tt>fcgi_stdio.h</tt>,
        because <tt>stdio.h</tt> often defines macros for functions such
        as <tt>getc</tt> and <tt>putc</tt>.  Fortunately, on most
        platforms <tt>stdio.h</tt> is protected against multiple
        includes by lines near the top of the file that look like
    <pre>
    #ifndef _STDIO_H
    #define _STDIO_H
    </pre>
        The specific symbol used for multiple-include protection,
        <tt>_STDIO_H</tt> in this example, varies from platform to
        platform.  As long as your platform protects <tt>stdio.h</tt>
        against multiple includes, you can forget about this issue.<p>
    <li>
        If your application passes <tt>FILE *</tt> to functions
        implemented in libraries for which you have source code, then
        you'll want to recompile these
        libraries with <tt>fcgi_stdio.h</tt> included.
        Most C compilers provide a command-line option for including
        headers in a program being compiled; using such a compiler feature
        allows you to rebuild your libraries without making source changes.
        For instance the gcc command line
    <pre>
    gcc -include /usr/local/include/fcgi_stdio.h wonderlib.c
    </pre>
        causes gcc to include <tt>fcgi_stdio.h</tt> before it even
        begins to read the module <tt>wonderlib.c</tt>.<p>
    <li>
        If your application passes <tt>FILE *</tt> to functions
        implemented in libraries
        for which you do not have source code, then
        you'll need to include the headers for these libraries
        <i>before</i> you include <tt>fcgi_stdio.h</tt>.
        You can't pass the <tt>stdin</tt>,
        <tt>stdout</tt>, or <tt>stderr</tt> streams produced by
        <tt>FCGI_Accept</tt> to any functions implemented by these
        libraries.  You can pass a stream on a Unix file to a library
        function by following this pattern:
    <pre>
    FILE *myStream = fopen(path, "r");
    answer = MungeStream(FCGI_ToFile(myStream));
    </pre>
        Here <tt>MungeStream</tt> is a library function that you can't
        recompile and <tt>FCGI_ToFile</tt> is a macro that converts
        from <tt>FCGI_FILE *</tt> to <tt>FILE *</tt>.  The macro
        <tt>FCGI_ToFile</tt> is defined in <tt>fcgi_stdio.h</tt>.<p>
</ul>


<h4>Converting CGI programs</h4>

The main task in converting a CGI program into a FastCGI program is
separating the code that needs to execute once, initializing the
program, from the code that needs to run for each request.
In our tiny example, initializing the <tt>count</tt> variable
is outside the loop, while incrementing the <tt>count</tt> variable
goes inside.<p>

Retained application state may be an issue.  You must ensure that
any application state created in processing one request has no
unintended effects on later requests.  FastCGI offers the possibility
of significant application performance improvements, through caching;
it is up to you to make the caches work
correctly.<p>

Storage leaks may be an issue.  Many CGI programs don't worry about
storage leaks because the programs don't run for long enough for
bloating to be a problem.  When converting to FastCGI, you
can either use a tool such as
<a href ="http://www.pure.com/"><i>Purify</i></a> from Pure Software
to discover and fix storage
leaks, or you can run a C garbage collector such as
<a href ="http://www.geodesic.com/"><i>Great Circle</i></a>
from Geodesic Systems.<p>


<h4>Limitations</h4>

Currently there are some limits to the compatibility provided
by the <tt>fcgi_stdio</tt> library:<p>

<ul type = square>
    <li>
        The library does not provide FastCGI versions of
        the functions <tt>fscanf</tt> and <tt>scanf</tt>.  If you wish
        to apply <tt>fscanf</tt> or <tt>scanf</tt> to <tt>stdin</tt>
        of a FastCGI program, the workaround is to read lines or other
        natural units into memory and then call <tt>sscanf</tt>.  If
        you wish to apply <tt>fscanf</tt> to a stream on a Unix file,
        the workaround is to follow the pattern:
    <pre>
    FILE *myStream = fopen(path, "r");
    count = fscanf(FCGI_ToFile(myStream), format, ...);
    </pre>
</ul>

<h4>Reference documentation</h4>

The <a href = "FCGI_Accept.3"><tt>FCGI_Accept</tt> manpage</a>,
<tt>doc/FCGI_Accept.3</tt>, describes the function in the traditional
format.<p>

The <a href = "FCGI_Finish.3"><tt>FCGI_Finish</tt></a>
(<tt>doc/FCGI_Finish.3</tt>),
<a href = "FCGI_SetExitStatus.3"><tt>FCGI_SetExitStatus</tt></a>
(<tt>doc/FCGI_SetExitStatus.3</tt>), and
<a href = "FCGI_StartFilterData.3"><tt>FCGI_StartFilterData</tt></a>
(<tt>doc/FCGI_StartFilterData.3</tt>)
manpages document capabilities of the <tt>fcgi-stdio</tt>
library that are not illustrated above.<p>


<h4><a NAME = "S3.2">3.2 Using the <tt>fcgiapp</tt> library</a></h4>

The <tt>fcgiapp</tt> library is a second C library for FastCGI.  It
does not provide the high degree of source code compatibility provided
by <tt>fcgi_stdio</tt>; in return, it does not make such heavy use of
<tt>#define</tt>.  <tt>fcgi_stdio</tt> is implemented as
a thin layer on top of <tt>fcgiapp</tt>.<p>

Applications built using the <tt>fcgiapp</tt> library cannot run
as CGI programs; that feature is provided at the <tt>fcgi_stdio</tt>
level.<p>

Functions defined in <tt>fcgiapp</tt> are named using the prefix
<tt>FCGX_</tt> rather than <tt>FCGI_</tt>.  For instance,
<tt>FCGX_Accept</tt> is the <tt>fcgiapp</tt> version of
<tt>FCGI_Accept</tt>.<p>

Documentation of the <tt>fcgiapp</tt> library takes the form
of extensive comments in the header file <tt>include/fcgiapp.h</tt>.
The sample programs <tt>examples/tiny-fcgi2.c</tt> and
<tt>examples/echo2.c</tt> illustrate how to use
<tt>fcgiapp</tt>.<p>


<h4><a NAME = "S3.3">3.3 Using Perl and Tcl</a></h4>

A major advantage of the FastCGI approach to high-performance Web
applications is its language-neutrality.  CGI scripts written in
popular languages such as Perl and Tcl can be evolved into
high-performance FastCGI applications.<p>

We have produced FastCGI-integrated Perl and Tcl
interpreters.  Doing so was easy, since Perl and Tcl
are conventional C applications and <tt>fcgi_stdio</tt> was
designed for converting conventional C applications.  Essentially no source
code changes were required in these programs; a small amount
of code was added in order to make <tt>FCGI_Accept</tt> and other
FastCGI primitives available in these languages.  And because
these interpreters were developed using <tt>fcgi_stdio</tt>, they
run standard Perl and Tcl applications (e.g. CGI scripts) as well
as FastCGI applications.<p>

See the
<a href = "http://www.fastcgi.com/applibs/">www.fastcgi.com/applibs</a>
Web page for downloadable Perl and Tcl binaries for selected platforms.
Because many users of Perl and Tcl run extended versions of these languages,
the kit includes separate companion documents describing
<a href ="fcgi-perl.htm">how to build FastCGI-integrated Perl</a>
and <a href ="fcgi-tcl.htm">how to build FastCGI-integrated Tcl</a>.<p>

Here are the Perl and Tcl versions of <tt>tiny-fcgi</tt>:<p>
<pre>
#!./perl
use FCGI;
$count = 0;
while(FCGI::accept() >= 0) {
    print("Content-type: text/html\r\n\r\n",
          "&lt;title&gt;FastCGI Hello! (Perl)&lt;/title&gt;\n",
          "&lt;h1&gt;FastCGI Hello! (Perl)&lt;/h1&gt;\n";
          "Request number ",  ++$count,
          " running on host &lt;i&gt";$env(SERVER_NAME)&lt;/i&gt;");
}
</pre>

<pre>
#!./tclsh
set count 0 
while {[FCGI_Accept] &gt;= 0 } {
    incr count
    puts -nonewline "Content-type: text/html\r\n\r\n"
    puts "&lt;title&gt;FastCGI Hello! (Tcl)&lt;/title&gt;"
    puts "&lt;h1&gt;FastCGI Hello! (Tcl)&lt;/h1&gt;"
    puts "Request number $count running on host &lt;i&gt;$env(SERVER_NAME)&lt;/i&gt;"
}
</pre>

Converting a Perl or Tcl CGI application to FastCGI is not fundamentally
different from converting a C CGI application to FastCGI.  You separate
the portion of the application that performs one-time
initialization from the portion that performs per-request
processing.  You put the per-request processing into a loop
controlled by <tt>FCGI::accept</tt> (Perl) or <tt>FCGI_Accept</tt>
(Tcl).



<h4><a NAME = "S3.4">3.4 Using Java</a></h4>

Java is not just for browser-based applets.  It is already suitable for
writing some Web server applications, and its range of applicability will
only grow as Java compilers and other Java tools improve.  Java's
modules, garbage collection, and threads are especially valuable
for writing long-lived application servers.<p>

The <tt>FCGIInterface</tt> class provides facilities for Java
applications analogous to what <tt>fcgi_stdio</tt> provides for C
applications.  Using this library your Java application can run using
either CGI or FastCGI.<p>

The kit includes separate companion document on
<a href ="fcgi-java.htm">using FastCGI with Java</a>.  The
source code for FastCGI classes is contained in
directory <tt>java/src</tt>
and the compiled code in <tt>java/classes</tt>.<p>

Here is the Java version of <tt>tiny-fcgi</tt>:<p>

<pre>
import FCGIInterface;

class TinyFCGI {        
    public static void main (String args[]) {           
        int count = 0;
        while(new FCGIInterface().FCGIaccept()>= 0) {
            count ++;
            System.out.println("Content-type: text/html\r\n\r\n");
            System.out.println(
                    "&lt;title&gt;FastCGI Hello! (Java)&lt;/title&gt;");
            System.out.println("&lt;h1&gt;FastCGI Hello! (Java)&lt;/h1&gt;");
            System.out.println(
                    "request number " + count + " running on host &lt;i&gt;" +
                    System.getProperty("SERVER_NAME") + "&lt;/i&gt;");
        }
    }
}
</pre>
<p>


<h3><a NAME = "S4">4. Running applications</a></h3>


<h3><a NAME = "S4.1">4.1 Using a Web server that supports FastCGI</a></h3>

For a current listing of Web servers that support FastCGI,
see the <a HREF = "http://www.fastcgi.com/servers">www.fastcgi.com/servers</a>
Web page.<p>

All of the Web servers that support FastCGI perform management of
FastCGI applications.  You don't need to start and stop FastCGI
applications; the Web server takes care of this.  If an application
process should crash, the Web server restarts it.<p>

Web servers support FastCGI via new configuration directives.
Since these directives are server-specific, get more information
from the documentation that accompanies each server.<p>

The directory
<a HREF = "../examples/conf"><tt>examples/conf</tt></a> contains config
files designed to run the example programs included in the FastCGI
Developer's Kit.  Each config file contains specific installation
instructions.<p>

The more advanced example programs take advantage of special features
of the Open Market Secure WebServer, such as anonymous ticketing
and support for the Authorizer role.  If you don't have this server,
download a
<a HREF= "http://www.openmarket.com/store/eval/swsg.htm">free
evaluation copy</a> to run the examples.<p>


<h3><a NAME = "S4.2">4.2 Using <tt>cgi-fcgi</tt> with any Web server</a></h3>

The program <tt>cgi-fcgi</tt> allows you to run FastCGI applications
using any Web server that supports CGI.<p>

Here is how <tt>cgi-fcgi</tt> works.  <tt>cgi-fcgi</tt> is a
standard CGI program that uses Unix domain or TCP/IP sockets
to communicate with
a FastCGI application.  <tt>cgi-fcgi</tt> takes the path name
or host/port name of a
listening socket as a parameter and <tt>connect</tt>s to the FastCGI
application listening on that socket.  <tt>cgi-fcgi</tt> then forwards the CGI
environment variables and <tt>stdin</tt> data to the FastCGI
application, and forwards the <tt>stdout</tt> and <tt>stderr</tt> data
from the FastCGI application to the Web server.  When the FastCGI
application signals the end of its response, <tt>cgi-fcgi</tt> flushes
its buffers and exits.<p>

Obviously, having <tt>cgi-fcgi</tt> is not as good as having
a server with integrated FastCGI support:<p>

<ul>
  <li>Communication is slower than with a Web server that avoids the
      fork/exec overhead on every FastCGI request.

  <li><tt>cgi-fcgi</tt> does not perform application management,
      so you need to provide this yourself.

  <li><tt>cgi-fcgi</tt> supports only the Responder role.
</ul>

But <tt>cgi-fcgi</tt> does allow you to
develop applications that retain state in memory between connections,
which often provides a major performance boost over normal CGI.  And
all the applications you develop using <tt>cgi-fcgi</tt>
will work with Web servers that have integrated support for
FastCGI.<p>

The file <tt>examples/tiny-fcgi.cgi</tt> demonstrates
a way to use <tt>cgi-fcgi</tt> to run a typical application,
in this case the <tt>examples/tiny-fcgi</tt> application:

<pre>
    #!../cgi-fcgi/cgi-fcgi -f
    -connect sockets/tiny-fcgi tiny-fcgi
</pre>

On most Unix platforms, executing this command-interpreter file runs
<tt>cgi-fcgi</tt> with arguments <tt>-f</tt> and
<tt>examples/tiny-fcgi.cgi</tt>.  (Beware: On some
Unix platforms, including HP-UX, the first line of a command-interpreter file
cannot contain more than 32 characters, including the newline;
you may need to install the <tt>cgi-fcgi</tt> application in a standard
place like <tt>/usr/local/bin</tt> or create a symbolic link
to the <tt>cgi-fcgi</tt> application in the directory containing
your application.)
The <tt>cgi-fcgi</tt> program reads the command-interpreter file and
connects to the FastCGI application whose listening socket is
<tt>examples/sockets/tiny-fcgi</tt>.<p>

Continuing the example, if
<tt>cgi-fcgi</tt>'s connection attempt fails, it
creates a new process
running the program <tt>examples/tiny-fcgi</tt> and listening on
socket <tt>examples/sockets/tiny-fcgi</tt>.  Then <tt>cgi-fcgi</tt>
retries the connection attempt, which now should succeed.<p>

The <tt>cgi-fcgi</tt> program has two other modes of operation.
In one mode it connects to applications but does not start them;
in the other it starts applications but does not connect
to them.  These modes are required when using TCP/IP.  The
<a href = "cgi-fcgi.1"><tt>cgi-fcgi</tt> manpage</a>,
<tt>doc/cgi-fcgi.1</tt>, tells the full story.<p>

To run the example applications using <tt>cgi-fcgi</tt>, start your
Web server and give it the directory <tt>fcgi-devel-kit</tt> as the
root of its URL space.  If the machine running your server is called
<tt>bowser</tt> and your server is running on port <tt>8888</tt>,
you'd then open the URL <tt>http://bowser:8888/index.html</tt> to
reach the kit's index page.  Now the links on the index page that run
example applications via <tt>cgi-fcgi</tt> should be active.<p>



<h3><a NAME = "S5">5. Known problems</a></h3>

On Digital UNIX 3.0 there's a problem with Unix domain listening
sockets on NFS file systems.  The symptom when using cgi-fcgi is an
exit status of 38 (<tt>ENOTSOCK</tt>: socket operation on non-socket),
but cgi-fcgi may dump core in this case when compiled optimized.
Work-around: Store your Unix domain listening
sockets on a non NFS file system, upgrade to Digital UNIX 3.2, or use
TCP sockets.<p>

On AIX there's a problem with shared listening sockets.
The symptoms can include application core dumps and kernel panic.
Work-around: Run a single FastCGI application server per listening
socket.<p>



<h3><a NAME = "S6">6. Getting support</a></h3>

The mailing list <tt>fastcgi-developers@openmarket.com</tt>
is used for discussions of issues in developing FastCGI applications.
Topics include announcement of FastCGI-capable Web servers or
changes to such servers, announcement of new application libraries
or changes to such libraries, announcement of known bugs, discussion
of design trade-offs in FastCGI application programming, and discussion
of development plans and experiences.  To join the
list, send a message to
<a href = "mailto:fastcgi-developers-request@openmarket.com">fastcgi-developers-request@openmarket.com</a>
with a message body consisting of the word "subscribe"
(leaving off the quotes).<p>

Mail sent to this list is archived and available on the
World-Wide Web at
<pre>
    http://www.fastcgi.com/mail/
</pre>

Open Market Secure WebServer customers get FastCGI
support where they get server support.<p>



<hr>

<address>
&#169 1996, Open Market, Inc. / mbrown@openmarket.com
</address>