NAME
    Apache2::SSI - Apache2 Server Side Include

SYNOPSIS
    Outside of Apache:

        use Apache2::SSI;
        my $ssi = Apache2::SSI->new(
            ## If running outside of Apache
            document_root => '/path/to/base/directory'
            ## Default error message to display when ssi failed to parse
            ## Default to [an error occurred while processing this directive]
            errmsg => '[Oops]'
        );
        my $fh = IO::File->new( "</some/file.html" ) || die( "$!\n" );
        $fh->binmode( ':utf8' );
        my $size = -s( $fh );
        my $html;
        $fh->read( $html, $size );
        $fh->close;
        if( !defined( my $result = $ssi->parse( $html ) ) )
        {
            $ssi->throw;
        };
        print( $result );

    Inside Apache, in the VirtualHost configuration, for example:

        PerlModule Apache2::SSI
        PerlOptions +GlobalRequest
        PerlSetupEnv On
        <Directory "/home/joe/www">
            Options All +Includes +ExecCGI -Indexes -MultiViews
            AllowOverride All
            SetHandler modperl
            # You can choose to set this as a response handler or a output filter, whichever works.
            # PerlResponseHandler Apache2::SSI
            PerlOutputFilterHandler Apache2::SSI
            # If you do not set this to On, path info will not work, example:
            # /path/to/file.html/path/info
            # See: <https://httpd.apache.org/docs/current/en/mod/core.html#acceptpathinfo>
            AcceptPathInfo On
            # To enable no-caching (see no_cache() in Apache2::RequestUtil:
            PerlSetVar Apache2_SSI_NO_CACHE On
            # This is required for exec cgi to work:
            # <https://httpd.apache.org/docs/current/en/mod/mod_include.html#element.exec>
            <Files ~ "\.pl$">
                SetHandler perl-script
                AcceptPathInfo On
                PerlResponseHandler ModPerl::PerlRun
                ## Even better for stable cgi scripts:
                ## PerlResponseHandler ModPerl::Registry
                ## Change this in mod_perl1 PerlSendHeader On to the following:
                ## <https://perl.apache.org/docs/2.0/user/porting/compat.html#C_PerlSendHeader_>
                PerlOptions +ParseHeaders
            </Files>
            <Files ~ "\.cgi$">
                SetHandler cgi-script
                AcceptPathInfo On
            </Files>
            # To enable debugging output in the Apache error log
            # PerlSetVar Apache2_SSI_DEBUG 3
            # To set the default echo message
            # PerlSetVar Apache2_SSI_Echomsg 
            # To Set the default error message
            # PerlSetVar Apache2_SSI_Errmsg "Oops, something went wrong"
            # To Set the default size format: bytes or abbrev
            # PerlSetVar Apache2_SSI_Sizefmt "bytes"
            # To Set the default date time format
            # PerlSetVar Apache2_SSI_Timefmt ""
            # To enable legacy mode:
            # PerlSetVar Apache2_SSI_Expression "legacy"
            # To enable trunk mode:
            # PerlSetVar Apache2_SSI_Expression "trunk"
        </Directory>

VERSION
        v0.2.0

DESCRIPTION
    Apache2::SSI implements Apache Server Side Include
    <https://httpd.apache.org/docs/current/en/howto/ssi.html>, a.k.a. SSI,
    within and outside of Apache2/mod_perl2 framework.

    Apache2::SSI is inspired from the original work of Apache::SSI with the
    main difference that Apache2::SSI works well when called from within
    Apache mod_perl2 as well as when called outside of Apache if you want to
    simulate SSI <https://httpd.apache.org/docs/current/en/howto/ssi.html>.

    Apache2::SSI also implements all of Apache SSI features, including
    functions, encoding and decoding and old style variables such as
    "${QUERY_STRING}" as well as modern style such as "v('QUERY_STRING')"
    and variants such as "%{REQUEST_URI}".

    See below details in this documentation and in the section on "SSI
    Directives"

    Under Apache mod_perl, you would implement it like this in your
    "apache2.conf" or "httpd.conf"

        <Files *.phtml>
            SetHandler modperl
            PerlOutputFilterHandler Apache2::SSI
        </Files>

    This would enable Apache2::SSI for files whose extension is ".phtml".
    You can also limit this by location, such as:

        <Location /some/web/path>
            <Files *.html>
                SetHandler modperl
                PerlOutputFilterHandler Apache2::SSI
            </Files>
        </Location>

    In the example above, we enable it in files with extensions ".phtml",
    but you can, of course, enable it for all html by setting extension
    ".html" or whatever extension you use for your html files.

    As pointed out by Ken Williams, the original author of Apache::SSI, the
    benefit for using Apache2::SSI is:

    1. You want to subclass Apache2::SSI and have granular control on how to
    render ssi
    2. You want to "parse the output of other mod_perl handlers, or send the
    SSI output through another handler"
    3. You want to imitate SSI without activating them or without using
    Apache (such as in command line) or within your perl/cgi script

  INSTALLATION
        perl Makefile.PL
        make
        make test
        sudo make install

    This will detect if you have Apache installed and run the Apache
    mod_perl2 tests by starting a separate instance of Apache on a
    non-standard port like 8123 under your username just for the purpose of
    testing. This is all handled automatically by Apache::Test

    If you do not have Apache or mod_perl installed, it will still install,
    but obviously not start an instance of Apache/mod_perl, nor perform any
    of the Apache mod_perl tests.

    It tries hard to find the Apache configuration file. You can help it by
    providing command line modifiers, such as:

        perl Makefile.PL -apxs /usr/bin/apxs

    or, even specify the Apache configuration file:

        perl Makefile.PL -apxs /usr/bin/apxs -httpd_conf /home/john/etc/apache2/apache2.conf

    To run only some tests, for example:

        make test TEST_FILES="./t/31.file.t"

    If you are on a Linux type system, you can install "apxs" by issuing on
    the command line:

        apt install apache2-dev

    You can check if you have it installed with the following command:

        dpkg -l | grep apache

    See ExtUtils::MakeMaker for more information.

METHODS
  new
    This instantiate an object that is used to access other key methods. It
    takes the following parameters:

    *apache_filter*
        This is the Apache2::Filter object object that is provided if
        running under mod_perl.

    *apache_request*
        This is the Apache2::RequestRec object that is provided if running
        under mod_perl.

        it can be retrieved from "request" in Apache2::RequestUtil or via
        "r" in Apache2::Filter

        You can get this Apache2::RequestRec object by requiring
        Apache2::RequestUtil and calling its class method "request" in
        Apache2::RequestUtil such as Apache2::RequestUtil->request and
        assuming you have set "PerlOptions +GlobalRequest" in your Apache
        Virtual Host configuration.

        Note that there is a main request object and subprocess request
        object, so to find out which one you are dealing with, use
        "is_initial_req" in Apache2::RequestUtil, such as:

            use Apache2::RequestUtil (); # extends Apache2::RequestRec objects
            my $r = $r->is_initial_req ? $r : $r->main;

    *debug*
        Sets the debug level. Starting from 3, this will output on the
        STDERR or in Apache error log a lot of debugging output.

    *document_root*
        This is only necessary to be provided if this is not running under
        Apache mod_perl. Without this value, Apache2::SSI has no way to
        guess the document root and will not be able to function properly
        and will return an "error".

    *document_uri*
        This is only necessary to be provided if this is not running under
        Apache mod_perl. This must be the uri of the document being served,
        such as "/my/path/index.html". So, if you are using this outside of
        the rim of Apache mod_perl and your file resides, for example, at
        "/home/john/www/my/path/index.html" and your document root is
        "/home/john/www", then the document uri would be
        "/my/path/index.html"

    *errmsg*
        The error message to be returned when a ssi directive fails. By
        default, it is "[an error occurred while processing this directive]"

    *html*
        The html data to be parsed. You do not have to provide that value
        now. You can provide it to "parse" as its first argument when you
        call it.

    *legacy*
        Takes a boolean value suchas 1 or 0 to indicate whether the Apache2
        expression supported accepts legacy style.

        Legacy Apache expression typically allows for perl style variable
        "${REQUEST_URI}" versus the modern style of "%{REQUEST_URI}" and
        just an equal sign to imply a regular expression such as:

            $HTTP_COOKIES = /lang\%22\%3A\%22([a-zA-Z]+\-[a-zA-Z]+)\%22\%7D;?/

        Modern expression equivalent would be:

            %{HTTP_COOKIES} =~ /lang\%22\%3A\%22([a-zA-Z]+\-[a-zA-Z]+)\%22\%7D;?/

        See Regexp::Common::Apache2 for more information.

        See also the property *trunk* to enable experimental expressions.

    *remote_ip*
        This is used when you want to artificially set the remote ip
        address, i.e. the address of the visitor accessing the page. This is
        used essentially by the SSI directive:

            my $ssi = Apache2::SSI->new( remote_ip => '192.168.2.10' ) ||
                die( Apache2::SSI->error );

            <!--#if expr="-R '192.168.2.0/24' || -R '127.0.0.1/24'" -->
            Remote ip is part of my private network
            <!--#else -->
            Go away!
            <!--#endif -->

    *sizefmt*
        The default way to format a file size. By default, this is "abbrev",
        which means a human readable format such as "2.5M" for 2.5
        megabytes. Other possible value is "bytes" which would have the
        "fsize" ssi directive return the size in bytes.

        See Apache2 documentation
        <https://httpd.apache.org/docs/current/en/howto/ssi.html> for more
        information on this.

    *timefmt*
        The default way to format a date time. By default, this uses the
        display according to your locale, such as "ja_JP" (for Japan) or
        "en_GB" for the United Kingdoms. The time zone can be specified in
        the format, or it will be set to the local time zone, whatever it
        is.

        See Apache2 documentation
        <https://httpd.apache.org/docs/current/en/howto/ssi.html> for more
        information on this.

    *trunk*
        This takes a boolean value such as 0 or 1 and when enabled this
        allows the support for Apache2 experimental expressions.

        See Regexp::Common::Apache2 for more information.

        Also, see the property *legacy* to enable legacy Apache2
        expressions.

  handler
    This is a key method expected by mod_perl. Depending on how this module
    is used, it will redirect either to "apache_filter_handler" or to
    "apache_response_handler"

  ap2perl_expr
    This method is used to convert Apache2 expressions into perl equivalents
    to be then eval'ed.

    It takes an hash reference provided by "parse" in Apache2::Expression,
    an array reference to store the output recursively and an optional hash
    reference of parameters.

    It parse recursively the structure provided in the hash reference to
    provide the perl equivalent for each Apache2 expression component.

    It returns the array reference provided used as the content buffer. This
    array is used by "parse_expr" and then joined using a single space to
    form a string of perl expression to be eval'ed.

  apache_filter
    Set or get the Apache2::Filter object.

    When running under Apache mod_perl this is set automatically from the
    special "handler" method.

  apache_filter_handler
    This method is called from "handler" to handle the Apache response when
    this module Apache2::SSI is used as a filter handler.

    See also "apache_response_handler"

  apache_request
    Sets or gets the Apache2::RequestRec object. As explained in the "new"
    method, you can get this Apache object by requiring the package
    Apache2::RequestUtil and calling "request" in Apache2::RequestUtil such
    as "Apache2::RequestUtil-"request> assuming you have set "PerlOptions
    +GlobalRequest" in your Apache Virtual Host configuration.

    When running under Apache mod_perl this is set automatically from the
    special "handler" method, such as:

        my $r = $f->r; # $f is the Apache2::Filter object provided by Apache

  apache_response_handler
    This method is called from "handler" to handle the Apache response when
    this module Apache2::SSI is used as a response handler.

    See also "apache_filter_handler"

  clone
    Create a clone of the object and return it.

  decode_base64
    Decode base64 data provided. When running under Apache mod_perl, this
    uses "decode" in APR::Base64 module, otherwise it uses "decode" in
    MIME::Base64

    If the decoded data contain utf8 data, this will decoded the utf8 data
    using "decode" in Encode

    If an error occurred during decoding, it will return undef and set an
    "error" object accordingly.

  decode_entities
    Decode html data containing entities. This uses "decode_entities" in
    HTML::Entities

    If an error occurred during decoding, it will return undef and set an
    "error" object accordingly.

    Example:

        $ssi->decode_entities( 'Tous les &Atilde;&ordf;tres humains naissent libres et &Atilde;&copy;gaux en dignit&Atilde;&copy; et en droits.' );
        # Tous les ��tres humains naissent libres et ��gaux en dignit�� et en droits.

  decode_uri
    Decode uri encoded data. This uses "uri_unescape" in URI::Escape.

    Not to be confused with x-www-form-urlencoded data. For that see
    "decode_url"

    If an error occurred during decoding, it will return undef and set an
    "error" object accordingly.

    Example:

        $ssi->decode_uri( 'https%3A%2F%2Fwww.example.com%2F' );
        # https://www.example.com/

  decode_url
    Decode x-www-form-urlencoded encoded data. When using Apache mod_perl,
    this uses "decode" in APR::Request and "decode" in Encode, otherwise it
    uses "url_decode_utf8" in URL::Encode (its XS version) to achieve the
    same result.

    If an error occurred during decoding, it will return undef and set an
    "error" object accordingly.

    Example:

        $ssi->decode_url( 'Tous+les+%C3%83%C2%AAtres+humains+naissent+libres+et+%C3%83%C2%A9gaux+en+dignit%C3%83%C2%A9+et+en+droits.' );
        # Tous les ��tres humains naissent libres et ��gaux en dignit�� et en droits.

  document_filename
    This is an alias for "filename" in Apache2::SSI::URI

  document_directory
    Returns an Apache2::SSI::URI object of the current directory of the
    "document_uri" provided.

  document_path
    Sets or gets the uri path to the document. This is the same as
    "document_uri", except it is striped from "query_string" and
    "path_info".

  document_root
    Sets or gets the document root.

    Wen running under Apache mod_perl, this value will be available
    automatically, using "document_root" in Apache2::RequestRec method.

    If it runs outside of Apache, this will use the value provided upon
    instantiating the object and passing the *document_root* parameter. If
    this is not set, it will return the value of the environment variable
    "DOCUMENT_ROOT".

  document_uri
    Sets or gets the document uri, which is the uri of the document being
    processed.

    For example:

        /index.html

    Under Apache, this will get the environment variable "DOCUMENT_URI" or
    calls the "uri" in Apache2::RequestRec method.

    Outside of Apache, this will rely on a value being provided upon
    instantiating an object, or the environment variable "DOCUMENT_URI" be
    present.

    The value should be an absolute uri.

  echomsg
    The default message to be returned for the "echo" command when the
    variable called is not defined.

    Example:

        $ssi->echomsg( '[Value Undefined]' );
        ## or in the document itself
        <!--#config echomsg="[Value Undefined]" -->
        <!--#echo var="NON_EXISTING" encoding="none" -->

    would produce:

        [Value Undefined]

  encode_base64
    Encode data provided into base64. When running under Apache mod_perl,
    this uses "encode" in APR::Base64 module, otherwise it uses "encode" in
    MIME::Base64

    If the data have the perl internal utf8 flag on as checked with
    "is_utf8" in Encode, this will encode the data into utf8 using "encode"
    in Encode before encoding it into base64.

    Please note that the base64 encoded resulting data is all on one line,
    similar to what Apache would do. The data is NOT broken into lines of 76
    characters.

    If an error occurred during encoding, it will return undef and set an
    "error" object accordingly.

  encode_entities
    Encode data into html entities. This uses "encode_entities" in
    HTML::Entities

    If an error occurred during encoding, it will return undef and set an
    "error" object accordingly.

    Example:

        $ssi->encode_entities( 'Tous les ��tres humains naissent libres et ��gaux en dignit�� et en droits.' );
        # Tous les &Atilde;&ordf;tres humains naissent libres et &Atilde;&copy;gaux en dignit&Atilde;&copy; et en droits.

  encode_uri
    Encode uri data. This uses "uri_escape_utf8" in URI::Escape.

    Not to be confused with x-www-form-urlencoded data. For that see
    "encode_url"

    If an error occurred during encoding, it will return undef and set an
    "error" object accordingly.

    Example:

        $ssi->encode_uri( 'https://www.example.com/' );
        # https%3A%2F%2Fwww.example.com%2F

  encode_url
    Encode data provided into an x-www-form-urlencoded string. When using
    Apache mod_perl, this uses "encode" in APR::Request, otherwise it uses
    "url_encode_utf8" in URL::Encode (its XS version)

    If an error occurred during decoding, it will return undef and set an
    "error" object accordingly.

    Example:

        $ssi->encode_url( 'Tous les ��tres humains naissent libres et ��gaux en dignit�� et en droits.' );
        # Tous+les+%C3%83%C2%AAtres+humains+naissent+libres+et+%C3%83%C2%A9gaux+en+dignit%C3%83%C2%A9+et+en+droits.

  env
    Sets or gets the value for an environment variable. Or, if no
    environment variable name is provided, it returns the entire hash
    reference. This method is intended to be used by users of this module,
    not by developers wanting to inherit from it.

    Note that the environment variable hash is unique for each new object,
    so it works like "subprocess_env" in Apache2::RequestRec, meaning each
    process has its set of environment variable.

    When a value is set for an environment variable that has an equivalent
    name, it will call the method as well with the new value provided. This
    is done to ensure data consistency and also additional processing if
    necessary.

    For example, let assume you set the environment variable "REQUEST_URI"
    or "DOCUMENT_URI" like this:

        $ssi->env( REQUEST_URI => '/some/path/to/file.html?q=something&l=ja_JP' );

    This will, in turn, call "request_uri", which is an alias for
    document_uri and this method will get the uri, path info and query
    string from the value provided and set those values accordingly, so they
    can be available when parsing.

  errmsg
    Sets or gets the error message to be displayed in lieu of a faulty ssi
    directive. This is the same behaviour as in Apache.

  error
    Retrieve the error object set. This is a Module::Generic::Error object.

    This module does not die nor "croak", but instead returns undef when an
    error occurs and set the error object.

    It is up to you to check the return value of the method calls. If you do
    not, you will miss important information. If you really want your script
    to die, it is up to you to interrupt it:

        if( !defined( $ssi->parse( $some_html_data ) ) )
        {
            die( $ssi->error );
        }

    or maybe more simply, when you are sure you will not get a false, but
    defined value:

        $ssi->parse( $some_html_data ) || die( $ssi->error );

    This example is dangerous, because "parse" might return an empty string
    which will be construed as a false value and will trigger the die
    statement, even though no error had occurred.

  filename
    This is an alias for "filename" in Apache2::SSI::URI

  find_file
    Provided with a file path, and this will resolve any variable used and
    attempt to look it up as a file if the argument *file* is provided with
    a file path as a value, or as a URI if the argument "virtual" is
    provided as an argument.

    This will call "lookup_file" or "lookup_uri" depending on whether it is
    dealing with a file or an uri.

    It returns a Apache2::SSI::URI object which is stringifyable and contain
    the file path.

  finfo
    Returns a Apache2::SSI::Finfo object. This provides access to "stat" in
    perlfunc information as method, taking advantage of APR::Finfo when
    running under Apache, and File::stat-like interface otherwise. See
    Apache2::SSI::Finfo for more information.

  html
    Sets or gets the html data to be processed.

  lookup_file
    Provided with a file path and this will look up the file.

    When using Apache, this will call "lookup_file" in Apache2::SubRequest.
    Outside of Apache, this will mimick Apache's lookup_file method by
    searching the file relative to the directory of the current document
    being served, i.e. the "document_uri".

    As per Apache SSI documentation, you cannot specify a path starting with
    "/" or "../"

    It returns a Apache2::SSI::File object.

  lookup_uri
    Provided with an uri, and this will loo it up and return a
    Apache2::SSI::URI object.

    Under Apache mod_perl, this uses "lookup_uri" in Apache2::SubRequest to
    achieve that. Outside of Apache it will attempt to lookup the uri
    relative to the document root if it is an absolute uri or to the current
    document uri.

    It returns a Apache2::SSI::URI object.

  mod_perl
    Returns true when running under mod_perl, false otherwise.

  parse
    Provided with html data and if none is provided will use the data
    specified with the method "html", this method will parse the html and
    process the ssi directives.

    It returns the html string with the ssi result.

  parse_config
    Provided with an hash reference of parameters and this sets three of the
    object parameters that can also be set during object instantiation:

    *echomsg*
        The value is a message that is sent back to the client if the echo
        element attempts to echo an undefined variable.

        This overrides any default value set for the parameter *echomsg*
        upon object instantiation.

    *errmsg*
        This is the default error message to be used as the result for a
        faulty ssi directive.

        See the "echomsg" method.

    *sizefmt*
        This is the format to be used to format the files size. Value can be
        either "bytes" or "abbrev"

        See also the "sizefmt" method.

    *timefmt*
        This is the format to be used to format the dates and times. The
        value is a date formatting based on "strftime" in POSIX

        See also the "timefmt" method.

  parse_echo
    Provided with an hash reference of parameter and this process the "echo"
    ssi directive and returns its output as a string.

    For example:

        Query string passed: <!--#echo var="QUERY_STRING" -->

    There are a number of standard environment variable accessible under SSI
    on top of other environment variables set. See "SSI Directives" section
    below.

  parse_echo_date_gmt
    Returns the current date with time zone set to gmt and based on the
    provided format or the format available for the current locale such as
    "ja_JP" or "en_GB".

  parse_echo_date_local
    Returns the current date with time zone set to the local time zone
    whatever that may be and on the provided format or the format available
    for the current locale such as "ja_JP" or "en_GB".

    Example:

        <!--#echo var="DATE_LOCAL" -->

  parse_echo_document_name
    Returns the document name. Under Apache, this returns the environment
    variable "DOCUMENT_NAME", if set, or the base name of the value returned
    by "filename" in Apache2::RequestRec

    Outside of Apache, this returns the environment variable
    "DOCUMENT_NAME", if set, or the base name of the value for
    "document_uri"

    Example:

        <!--#echo var="DOCUMENT_NAME" -->

    If the uri were "/some/where/file.html", this would return only
    "file.html"

  parse_echo_document_uri
    Returns the value of "document_uri"

    Example:

        <!--#echo var="DOCUMENT_URI" -->

    The document uri would include, if any, any path info and query string.

  parse_echo_last_modified
    This returns document last modified date. Under Apache, there is a
    standard environment variable called "LAST_MODIFIED" (see the section on
    "SSI Directives"), and if somehow absent, it will return instead the
    formatted last modification datetime for the file returned with
    "filename" in Apache2::RequestRec. The formatting of that date follows
    whatever format provided with "timefmt" or by default the datetime
    format for the current locale (e.g. "ja_JP").

    Outside of Apache, the similar result is achieved by returning the value
    of the environment variable "LAST_MODIFIED" if available, or the
    formatted datetime of the document uri as set with "document_uri"

    Example:

        <!--#echo var="LAST_MODIFIED" -->

  parse_eval_expr
    Provided with a string representing an Apache2 expression and this will
    parse it, transform it into a perl equivalent and return its value.

    It does the parsing using "parse" in Apache2::Expression called from
    "parse_expr"

    If the expression contains regular expression with capture groups, the
    value of capture groups will be stored and will be usable in later
    expressions, such as:

        <!--#config errmsg="[Include error]" -->
        <!--#if expr="%{HTTP_COOKIE} =~ /lang\%22\%3A\%22([a-zA-Z]+\-[a-zA-Z]+)\%22\%7D;?/"-->
            <!--#set var="CONTENT_LANGUAGE" value="%{tolower:$1}"-->
        <!--#elif expr="-z %{CONTENT_LANGUAGE}"-->
            <!--#set var="CONTENT_LANGUAGE" value="en"-->
        <!--#endif-->
        <!DOCTYPE html>
        <html lang="<!--#echo encoding="none" var="CONTENT_LANGUAGE" -->">

  parse_exec
    Provided with an hash reference of parameters and this process the
    "exec" ssi directives.

    Example:

        <!--#exec cgi="/uri/path/to/progr.cgi" -->

    or

        <!--#exec cmd="/some/system/file/path.sh" -->

  parse_expr
    It takes a string representing an Apache2 expression and calls "parse"
    in Apache2::Expression to break it down, and then calls "ap2perl_expr"
    to transform it into a perl expression that is then eval'ed by
    "parse_eval_expr".

    It returns the perl representation of the Apache2 expression.

    To make this work, certain Apache2 standard functions used such as
    "base64" or "md5" are converted to use this package function
    equivalents. See the "parse_func_*" methods for more information.

  parse_elif
    Parse the "elif" condition.

    Example:

        <!--#if expr=1 -->
         Hi, should print
        <!--#elif expr=1 -->
         Shouldn't print
        <!--#else -->
         Shouldn't print
        <!--#endif -->

  parse_else
    Parse the "else" condition.

    See "parse_elif" above for example.

  parse_endif
    Parse the "endif" condition.

    See "parse_elif" above for example.

  parse_flastmod
    Process the ssi directive "flastmod"

    Provided with an hash reference of parameters and this will return the
    formatted date time of the file last modification time.

  parse_fsize
    Provided with an hash reference of parameters and this will return the
    formatted file size.

    The output is affected by the value of "sizefmt". If its value is
    "bytes", it will return the raw size in bytes, and if its value is
    "abbrev", it will return its value formated in kilo, mega or giga units.

    Example

        <!--#config sizefmt="abbrev" -->
        This file size is <!--#fsize file="/some/filesystem/path/to/archive.tar.gz" -->

    would return:

    This file size is 12.7M

    Or:

        <!--#config sizefmt="bytes" -->
        This file size is <!--#fsize virtual="/some/filesystem/path/to/archive.tar.gz" -->

    would return:

    This file size is 13,316,917 bytes

    The size value before formatting is a Module::Generic::Number and the
    output is formatted using Number::Format by calling "format" in
    Module::Generic::Number

  parse_func_base64
    Returns the arguments provided into a base64 string.

    If the arguments are utf8 data with perl internal flag on, as checked
    with "is_utf8" in Encode, this will encode the data into utf8 with
    "encode" in Encode before encoding it into base64.

    Example:

        <!--#set var="payload" value='{"sub":"1234567890","name":"John Doe","iat":1609047546}' encoding="base64" -->
        <!--#if expr="$payload == 'eyJzdWIiOiIxMjM0NTY3ODkwIiwibmFtZSI6IkpvaG4gRG9lIiwiaWF0IjoxNjA5MDQ3NTQ2fQo='" -->
        Payload matches
        <!--#else -->
        Sorry, this failed
        <!--#endif -->

  parse_func_env
    Return first match of note, reqenv, and osenv

    Example:

        <!--#if expr="env( $QUERY_STRING ) == /\bl=ja_JP/" -->
        Showing Japanese data
        <!--#else -->
        Defaulting to English
        <!--#endif -->

  parse_func_escape
    Escape special characters in %hex encoding.

    Example:

        <!--#set var="website" value="https://www.example.com/" -->
        Please go to <a href="<!--#echo var='website' encoding='escape' -->"><!--#echo var="website" --></a>

  parse_func_http
    Get HTTP request header; header names may be added to the Vary header.

    Example:

        <!--#if expr="http('X-API-ID') == 1234567" -->
        You're good to go.
        <!--#endif -->

    However, outside of an Apache environment this will return the value of
    the environment variable in the following order:

    X-API-ID (i.e. the name as-is)
    HTTP_X_API_ID (i.e. adding "HTTP_" and replace "-" for "_")
    X_API_ID (i.e. same as above, but without the "HTTP_" prefix)

    If none is found, it returns an empty string.

    For an equivalent function for response headers, see "parse_func_resp"

  parse_func_ldap
    Escape characters as required by LDAP distinguished name escaping
    (RFC4514) and LDAP filter escaping (RFC4515).

    See Apache documentation
    <https://httpd.apache.org/docs/trunk/en/expr.html#page-header> for more
    information

    Example:

        <!--#set var="phrase" value="%{ldap:'Tous les ��tres humains naissent libres (et ��gaux) en dignit�� et\ en\ droits.\n'}" -->
        # Tous les ��tres humains naissent libres \28et ��gaux\29 en dignit�� et\5c en\5c droits.\5cn

  parse_func_md5
    Hash the string using MD5, then encode the hash with hexadecimal
    encoding.

    If the arguments are utf8 data with perl internal flag on, as checked
    with "is_utf8" in Encode, this will encode the data into utf8 with
    "encode" in Encode before encoding it with md5.

    Example:

        <!--#if expr="md5( $hash_data ) == '2f50e645b6ef04b5cfb76aed6de343eb'" -->
        You're good to go.
        <!--#endif -->

  parse_func_note
    Lookup request note

        <!--#set var="CUSTOMER_ID" value="1234567" -->
        <!--#if expr="note('CUSTOMER_ID') == 1234567" -->
        Showing special message
        <!--#endif -->

    This uses Apache2::SSI::Notes to enable notes to be shared on and off
    Apache2/mod_perl2 environment. Thus, you could set a note from a
    command-line perl script, and then access it under Apache2/mod_perl2 or
    just your regular script running under a web server.

    For example:

    In your perl script outside of Apache:

        # Basic parameters to make Apache2::SSI happy
        my $ssi = Apache2::SSI->new( document_root => '/home/john/www', document_uri => '/' ) ||
            die( Apache2::SSI->error );
        $ssi->notes( API_VERSION => 2 );

    Then, in your perl script running under the web server, be it
    Apache2/mod_perl2 or not:

        my $ssi = Apache2::SSI->new || die( Apache2::SSI->error );
        my $api_version = $ssi->notes( 'API_VERSION' );

    To enable shareability of notes on and off Apache, this makes uses of
    shared memory segments. See Apache2::SSI::Notes for more information on
    the notes api and perlipc for more information on shared memory
    segments.

    Just keep in mind that the notes are never removed even when Apache
    shuts down, so it is your responsibility to remove them if you do not
    want them anymore. For example:

        use Apache2::SSI::Notes;
        my $notes = Apache2::SSI::Notes->new;
        $notes->remove;

    be aware that shared notes might note be available for your platform.
    Check Apache2::SSI::Notes for more information and also perlport on
    shared memory segments.

  parse_func_osenv
    Lookup operating system environment variable

        <!--#if expr="env('LANG') =~ /en(_(GB|US))/" -->
        Showing English language
        <!--#endif -->

  parse_func_replace
    replace(string, "from", "to") replaces all occurrences of "from" in the
    string with "to".

    Example:

        <!--#if expr="replace( 'John is in Tokyo', 'John', 'Jack' ) == 'Jack is in Tokyo'" -->
        This worked!
        <!--#else -->
        Nope, it failed.
        <!--#endif -->

  parse_func_req
    See "parse_func_http"

  parse_func_reqenv
    Lookup request environment variable (as a shortcut, v can also be used
    to access variables).

    This is only different from "parse_func_env" under Apache.

    See "parse_func_env"

    Example:

        <!--#if expr="reqenv('ProcessId') == '$$'" -->
        This worked!
        <!--#else -->
        Nope, it failed.
        <!--#endif -->

    Or using the Apache SSI "v" shortcut:

        <!--#if expr="v('ProcessId') == '$$'" -->

  parse_func_req_novary
    Same as "parse_func_req", but header names will not be added to the Vary
    header.

  parse_func_resp
    Get HTTP response header.

    Example:

        <!--#if expr="resp('X-ProcessId') == '$$'" -->
        This worked!
        <!--#else -->
        Nope, it failed.
        <!--#endif -->

    An important note here:

    First, there is obviously no response header available for perl scripts
    running outside of Apache2/mod_perl2 framework.

    If the script runs under mod_perl, not all response header will be
    available depending on whether you are using Apache2::SSI in your Apache
    configuration as an output filter handler ("PerlOutputFilterHandler") or
    a response handler ("PerlResponseHandler").

    If it is running as an output filter handler, then some headers, such as
    "Content-Type" will not be available, unless they have been set by a
    script in a previous phase. Only basic headers will be available. For
    more information, check the Apache/mod_perl2 documentation on each
    phase.

  parse_func_sha1
    Hash the string using SHA1, then encode the hash with hexadecimal
    encoding.

    Example:

        <!--#if expr="sha1('Tous les ��tres humains naissent libres et ��gaux en dignit�� et en droits.') == '8c244078c64a51e8924ecf646df968094a818d59'" -->
        This worked!
        <!--#else -->
        Nope, it failed.
        <!--#endif -->

  parse_func_tolower
    Convert string to lower case.

    Example:

        <!--#if expr="tolower('Tous les ��tres humains naissent libres et ��gaux en dignit�� et en droits.') == 'tous les ��tres humains naissent libres et ��gaux en dignit�� et en droits.'" -->
        This worked!
        <!--#else -->
        Nope, it failed.
        <!--#endif -->

  parse_func_toupper
    Convert string to upper case.

    Example:

        <!--#if expr="toupper('Tous les ��tres humains naissent libres et ��gaux en dignit�� et en droits.') == 'TOUS LES ��TRES HUMAINS NAISSENT LIBRES ET ��GAUX EN DIGNIT�� ET EN DROITS.'" -->
        This worked!
        <!--#else -->
        Nope, it failed.
        <!--#endif -->

  parse_func_unbase64
    Decode base64 encoded string, return truncated string if 0x00 is found.

    Example:

        <!--#if expr="unbase64('VG91cyBsZXMgw6p0cmVzIGh1bWFpbnMgbmFpc3NlbnQgbGlicmVzIGV0IMOpZ2F1eCBlbiBkaWduaXTDqSBldCBlbiBkcm9pdHMu') == 'Tous les ��tres humains naissent libres et ��gaux en dignit�� et en droits.'" -->
        This worked!
        <!--#else -->
        Nope, it failed.
        <!--#endif -->

  parse_func_unescape
    Unescape %hex encoded string, leaving encoded slashes alone; return
    empty string if %00 is found.

    Example:

        <!--#if expr="unescape('https%3A%2F%2Fwww.example.com%2F') == 'https://www.example.com/'" -->
        This worked!
        <!--#else -->
        Nope, it failed.
        <!--#endif -->

  parse_if
    Parse the "if" condition.

    See "parse_elif" above for example.

  parse_include
    Provided with an hash reference of parameters and this process the ssi
    directive "include", which is arguably the most used.

    It will try to resolve the file to include by calling "find_file" with
    the same arguments this is called with.

    Under Apache, if the previous look up succeeded, it calls "run" in
    Apache2::SubRequest

    Outside of Apache, it reads the entire file, utf8 decode it and return
    it.

  parse_perl
    Provided with an hash reference of parameters and this parse some perl
    command and returns the output as a string.

    Example:

        <!--#perl sub="sub{ print 'Hello!' }" -->

    or

        <!--#perl sub="package::subroutine" -->

  parse_printenv
    This returns a list of environment variables sorted and their values.

  parse_set
    Provided with an hash reference of parameters and this process the ssi
    directive "set".

    Possible parameters are:

    *decoding*
        The decoding of the variable before it is set. This can be "none",
        "url", "urlencoded", "base64" or "entity"

    *encoding*
        This instruct to encode the variable value before display. It can
        the same possible value as for decoding.

    *value*
        The string value for the variable to be set.

    *var*
        The variable name

    Example:

        <!--#set var="debug" value="2" -->
        <!--#set decoding="entity" var="HUMAN_RIGHT" value="Tous les &Atilde;&ordf;tres humains naissent libres et &Atilde;&copy;gaux en dignit&Atilde;&copy; et en droits." encoding="urlencoded" -->

    See the Apache SSI documentation
    <https://httpd.apache.org/docs/current/en/mod/mod_include.html> for more
    information.

  parse_ssi
    Provided with the html data as a string and this will parse its embedded
    ssi directives and return its output as a string.

    If it fails, it sets an "error" and returns an empty string.

  path_info
    Sets or gets the path info for the current uri.

    Example:

        my $string = $ssi->path_info;
        $ssi->path_info( '/my/path/info' );

    The path info value is also set automatically when "document_uri" is
    called, such as:

        $ssi->document_uri( '/some/path/to/file.html/my/path/info?q=something&l=ja_JP' );

    This will also set automatically the "PATH_INFO" environment variable.

  query_string
    Set or gets the query string for the current uri.

    Example:

        my $string = $ssi->query_string;
        $ssi->query_string( 'q=something&l=ja_JP' );

    or, using the URI module:

        $ssi->query_string( $uri->query );

    The query string value is set automatically when you provide an
    document_uri upon instantiation or after:

        $ssi->document_uri( '/some/path/to/file.html?q=something&l=ja_JP' );

    This will also set automatically the "QUERY_STRING" environment
    variable.

  remote_ip
    Sets or gets the remote ip address of the visitor.

    Under Apache mod_perl, this will call "remote_ip" in Apache2::Connection
    for version 2.2 or lower and will call "useragent_ip" in
    Apache2::Connection for version above 2.2, and otherwise this will get
    the value from the environment variable "REMOTE_ADDR"

    This value can also be overriden by being provided during object
    instantiation.

        # Pretend the ssi directives are accessed from this ip
        $ssi->remote_ip( '192.168.2.20' );

    This is useful when one wants to check how the rendering will be when
    accessed from certain ip addresses.

    This is used primarily when there is an expression such as

        <!--#if expr="-R '192.168.1.0/24' -->
        Visitor is part of my private network
        <!--#endif -->

    or

        <!--#if expr="v('REMOTE_ADDR') -R '192.168.1.0/24' -->
        <!--#include file="/home/john/special_hidden_login_feature.html" -->
        <!--#endif -->

    Apache2::Connection also has a "remote_addr" in Apache2::Connection
    method, but this returns a APR::SockAddr object that is used to get the
    binary version of the ip. However you can also get the string version
    like this:

        use APR::SockAddr ();
        my $ip = $r->connection->remote_addr->ip_get();

    Versions above 2.2 make a distinction between ip from direct connection,
    or the real ip behind a proxy, i.e. "useragent_ip" in
    Apache2::Connection

  request_uri
    This is an alias for "document_uri"

  server_version
    Returns the server version as a version object can caches that value.

    Under mod_perl2, it uses "get_server_description" in Apache2::ServerUtil
    and outside of mod_perl, it tries to find "apxs" using File::Which and
    in last resort, tries to find the "apache2" or "httpd" binary to get its
    version information.

  sizefmt
    Sets or gets the formatting for file sizes. Value can be either "bytes"
    or "abbrev"

  timefmt
    Sets or gets the formatting for date and time values. The format takes
    the same values as "strftime" in POSIX

Encoding
    At present time, the html data are treated as utf8 data and decoded and
    encoded back as such.

    If there is a need to broaden support for other charsets, let me know.

SSI Directives
    This is taken from Apache documentation and summarised here for
    convenience and clarity to the perl community.

  config
        <!--#config errmsg="Error occurred" sizefmt="abbrev" timefmt="%B %Y" -->
        <!--#config errmsg="Oopsie" -->
        <!--#config sizefmt="bytes" -->
        # Thursday 24 December 2020
        <!--#config timefmt="%A $d %B %Y" -->

  echo
         <!--#set var="HTMl_TITLE" value="Un sujet int��ressant" -->
         <!--#echo var="HTMl_TITLE" encoding="entity" -->

    Encoding can be either "entity", "url" or "none"

  exec
        # pwd is "print working directory" in shell
        <!--#exec cmd="pwd" -->
        <!--#exec cgi="/uri/path/to/prog.cgi" -->

  include
        # Filesystem file path
        <!--#include file="/home/john/var/quote_of_the_day.txt" -->
        # Relative to the document root
        <!--#include virtual="/footer.html" -->

  flastmod
         <!--#flastmod file="/home/john/var/quote_of_the_day.txt" -->
         <!--#flastmod virtual="/copyright.html" -->

  fsize
        <!--#fsize file="/download/software-v1.2.tgz" -->
        <!--#fsize virtual="/images/logo.jpg" -->

  printenv
        <!--#printenv -->

  set
        <!--#set var="debug" value="2" -->

  if, elif, endif and else
        <!--#if expr="$debug > 1" -->
        I will print a lot of debugging
        <!--#else -->
        Debugging output will be reasonable
        <!--#endif -->

    or with new version of Apache SSI:

        No such file or directory.
        <!--#if expr="v('HTTP_REFERER') != ''" -->
        Please let the admin of the <a href="<!--#echo encoding="url" var="HTTP_REFERER" -->"referring site</a> know about their dead link.
        <!--#endif -->

  functions
    Apache SSI supports the following functions, as of Apache version 2.4.

    See Apache documentation
    <https://httpd.apache.org/docs/current/en/expr.html#page-header> for
    detailed description of what they do.

    You can also refer to the methods "parse_func_*" documented above, which
    implement those Apache functions.

    *base64*
    *env*
    *escape*
    *http*
    *ldap*
    *md5*
    *note*
    *osenv*
    *replace*
    *req*
    *reqenv*
    *req_novary*
    *resp*
    *sha1*
    *tolower*
    *toupper*
    *unbase64*
    *unescape*

  variables
    On top of all environment variables available, Apache makes the
    following ones also accessible:

    DATE_GMT
    DATE_LOCAL
    DOCUMENT_ARGS
    DOCUMENT_NAME
    DOCUMENT_PATH_INFO
    DOCUMENT_URI
    LAST_MODIFIED
    QUERY_STRING_UNESCAPED
    USER_NAME

    See Apache documentation
    <https://httpd.apache.org/docs/current/en/mod/mod_include.html#page-head
    er> and this page too
    <https://httpd.apache.org/docs/current/en/expr.html#page-header> for
    more information.

  expressions
    There is reasonable, but limited support for Apache expressions. For
    example, the followings are supported

    In the examples below, we use the variable "QUERY_STRING", but you can
    use any other variable of course.

    The regular expression are the ones PCRE <http://www.pcre.org/>
    compliant, so your perl regular expressions should work.

        <!--#if expr="$QUERY_STRING = 'something'" -->
        <!--#if expr="v('QUERY_STRING') = 'something'" -->
        <!--#if expr="%{QUERY_STRING} = 'something'" -->
        <!--#if expr="$QUERY_STRING = /^something/" -->
        <!--#if expr="$QUERY_STRING == /^something/" -->
        # works also with eq, ne, lt, le, gt and ge
        <!--#if expr="9 gt 3" -->
        <!--#if expr="9 -gt 3" -->
        # Other operators work too, namely == != < <= > >= =~ !~
        <!--#if expr="9 > 3" -->
        <!--#if expr="9 !> 3" -->
        <!--#if expr="9 !gt 3" -->
        # Checks the remote ip is part of this subnet
        <!--#if expr="-R 192.168.2.0/24" -->
        <!--#if expr="192.168.2.10 -R 192.168.2.0/24" -->
        <!--#if expr="192.168.2.10 -ipmatch 192.168.2.0/24" -->
        # Checks if variable is non-empty
        <!--#if expr="-n $some_variable" -->
        # Checks if variable is empty
        <!--#if expr="-z $some_variable" -->
        # Checks if the visitor can access the uri /restricted/uri
        <!--#if expr="-A /restricted/uri" -->

    For subnet checks, this uses Net::Subnet

    Expressions that would not work outside of Apache, i.e. it will return
    an empty string:

        <!--#expr="%{HTTP:X-example-header} in { 'foo', 'bar', 'baz' }" -->

    See Apache documentation <http://httpd.apache.org/docs/2.4/en/expr.html>
    for more information.

CREDITS
    Credits to Ken Williams for his implementation of Apache::SSI from which
    I borrowed some code.

AUTHOR
    Jacques Deguest <jack@deguest.jp>

    CPAN ID: jdeguest

    <https://gitlab.com/jackdeguest/Apache2-SSI>

SEE ALSO
    Apache2::SSI::File, Apache2::SSI::Finfo, Apache2::SSI::Notes,
    Apache2::SSI::URI, Apache2::SSI::SharedMem and Apache2::SSI::SemStat

    mod_include, mod_perl(3), Apache::SSI,
    <https://httpd.apache.org/docs/current/en/mod/mod_include.html>,
    <https://httpd.apache.org/docs/current/en/howto/ssi.html>,
    <https://httpd.apache.org/docs/current/en/expr.html>
    <https://perl.apache.org/docs/2.0/user/handlers/filters.html#C_PerlOutpu
    tFilterHandler_>

COPYRIGHT & LICENSE
    Copyright (c) 2020-2021 DEGUEST Pte. Ltd.

    You can use, copy, modify and redistribute this package and associated
    files under the same terms as Perl itself.