bump to 0.15_32

[bse.git] / INSTALL.pod

=head1 NAME

   BSE installation guide.


=head1 SYSTEM REQUIREMENTS

=over 4

=item *

perl 5.005_03 or later

=item *

mysql 3.22. or later

=item *

GnuPG or PGP 5 or 6

=item *

a web host:

=over 4

=item *

with telnet or ssh access

=item *

that runs your CGI scripts as your user

=item *

that runs your CGI scripts with the current working directory
set to the directory that contains the CGI script.

=back

=back

You will need at least the following Perl modules installed:

=over 4

=item *

DBI

=item *

DBD::mysql

=item *

Digest::MD5

=item *

Apache::Session

=item *

Storable (and this requires Log::Agent sometimes)

=item *

HTML::Parser

=item *

URI::Escape

=item *

HTML::Entities

=back

and their dependants.  If you use the CPAN shell to install these then
the dependants will be installed automatically.

All other modules are either supplied or standard with perl.

I assume you know how to use a text editor, and have a basic knowledge
of how directories work, and know enough perl to be able to edit
constants.

=head1 PLANNING

You need to know:

=over 4

=item *

the layout of directories on your web host: where CGI programs go
(cgi-bin), where's the root of the document tree (htdocs), and where
can you safely keep static data (datapath).  The names in parentheses
() will be used in this documentation.

=item *

how you want your site to look, presumably as one or more HTML files
and associated style sheets.

=item *

order processing information: the email address and public key of
the user who will be receiving the order emails.  The email address
and private key of the sender of the order emails (though this can be
generated during installation.

=back

=head1 EXTRACT FILES

Telnet or ssh to the web host.

Extract the archive into a working directory:

  mkdir work
  cd work
  tar xzf workpath/bse-0.04.tar.gz

The directories in the archive are:

  bse/                     Base directory
  bse/schema               Database schema definitions (and some test data)
  bse/site                 Laid out site
  bse/site/htdocs          Document root
  bse/site/cgi-bin         CGI programs
  bse/site/templates       Sample page templates
  bse/site/data            Static site data (currently just the stopwords list)
  bse/site/docs            Documentation
  bse/site/util            Utilities (currently just initial.pl)

If you are running your own host, or have sufficient control over
Apache, you may want to extract the archive in it's final directory
and simply create a new <VirtualHost..> that uses the extracted
directories.  If so, skip Copy Files.


=head1 COPY FILES

Replacing the names as decribed in L<PLANNING>

  # the base documents
  mkdir htdocs/admin
  cp -R workpath/site/htdocs htdocs/
  # page templates (you will need to modify these) and static data
  mkdir datadir/templates
  cp -R workpath/site/templates datadir/templates/
  cp -R workpath/site/data datadir/
  # cgi
  cp -R workpath/site/cgi-bin cgi-bin/


=head1 CONFIGURATION

Most configuration information is kept in Constants.pm, which is in
the cgi-bin/modules directory:

  vi cgi-bin/modules/Constants.pm

=head2 Database

$DB should be the name of your mysql database.

$UN and $PW should be your mysql login name and password.

=head2 Directory structure

If your directory structure matches that of the archive, this is
simple, set $BASEDIR to point to the site directory.  The other
variables are set based on that layout.

Otherwise, set:

=over 4

=item $TMPLDIR

to the directory you are keeping document templates in,
'datapath/templates/' if you followed L<COPY FILES>

=item $CONTENTBASE

to the directory you are keeping the site document files in, 'htdocs/'
if you followed L<COPY FILES>.

=item $IMAGEDIR

to the directory that image files are kept in.  This should be left as
$CONTENTBASE . 'images/'

=item $DATADIR

to the directory containing F<stopwords.txt>

=back

B<Warning:> these paths I<must> be kept as absolute directories.  They
must the directories as seen by the running CGI scripts.


=head2 Site name

You should set $URLBASE and $SECURLBASE to the URLs used to access
your site in normal and SSL (https).  If you currently don't have secure access setup, you can use the same non-secure URL for both.

B<Warning:> You I<must> set the $SECURLBASE to a secure URL and
regenerate the site before accepting orders on the site.

=head2 Level defaults

This is probably the most complex item to configure.

The %LEVEL_DEFAULTS hash describes how your site will look from the
administration interface.  It has little effect on how the site looks
from a user's perspective, except of course, that it sets defaults for
some items.

Each level of your site needs an entry in the hash, with the top level
being 0 to indicate the whole site.  This top-level should only have
the C<display> keyword defined.

Each level except for level 0 should have the following keywords defined:

=over 4

=item *

C<display> defines how the articles at this level is described when
adding new articles.

=item *

C<template> is the default template name used for that level.  It
should exist in the levels directory under your template directory.

=item *

C<threshold> is the default threshold used when a new article is
created at that level.  Thresholds are used to control whether child
article are rendered inline, or as summaries.

=back

=head2 Link Titles

If you set $LINK_TITLES to non-zero, then the links created for the
C<article link> and C<url article> tags will include the title as a
translated suffix.  This allows some search engines to index on the
URL as well as the content of the document. Without this suffix the
url contains no useful indexable information.

For this to work under Apache you will need the following in a
.htaccess file in the htdocs/a directory:

   RewriteEngine On
   RewriteRule ^([0-9]+\.html)/[0-9a-zA-Z_]*$ ./$1 [T=text/html]

=head2 Search options

In geneal the search engine will generate a search index for all
listed articles, and exclude unlisted articles.

You have some control over the indexing.  Set @SEARCH_INCLUDE to the
ids of sections that should be indexed, even if not listed.  Set
@SEARCH_EXCLUDE to the ids of sections that should not be indexed.

@SEARCH_EXCLUDE overrides @SEARCH_INCLUDE.

Set $SEARCH_LEVEL to the lowest-level (highest number) of the articles
to be indexed.  Lower level articles will still be indexed, but
searches for them will find their parent article (or I<their> parent
article, if it still isn't a low enough level.)

$SEARCH_ALL is used as the name of the entry in the drop-down list of
sections generated by the <:list:> tag on the search template.

=head2 Deletion control

Some articles are critical to the operation of your site.  If the
article's id is in @NO_DELETE it cannot be deleted.

=head2 The Shop

This is probably the second hardest item to configure.

It isn't necessary to configure this section until you need to use the
shop.

$SHOP_CRYPTO should be set to the supplied class that uses your
installed encryption software.  Note that currently only Squirrel::GPG
has been tested in production.  (Patches welcome on the other modules.)

$SHOP_SIGNING_ID should be the user id of the key to use for signing
orders.

$SHOP_GPG, $SHOP_PGPE and $SHOP_PGP are the locations of the specified
executables, for GunPG, PGP5 and PGP6 respectively.  They only need to
be set if the executable isn't in the PATH when the shop.pl runs.

$SHOP_SENDMAIL needs to be set to the name of sendmail or a compatible
program.

$SHOP_FROM will be used as the sender of the order emails.

Processed, encrypted orders will be $SHOP_TO_NAME and $SHOP_TO_EMAIL.
There must be a public key in the keyring, which has been signed by
the private key $SHOP_SIGNING_ID (or whatever your default private
signing key is.)

You can set $SHOP_EMAIL_ORDER to 0 to prevent the order being emailed
as above.  A confirmation email is still sent to the person who made
the order.  This is only really for testing, since only the encrypted
email contains the credit card number and expiry date.

=head2 Maintenance Tools

$DATA_EMAIL is the email address of a person that the 
I<Dump database to email> page will send the MySQL database dump to.

$MYSQLDUMP is the location of the MySQL C<mysqldump> tool.  If this
isn't in the PATH you will need to add this here.

=head1 DATABASE SETUP

=head2 Schema load

You need to install the database schema.  Presumably you have already
created a user and database for the new site (or your host has.)

To install the schema and the base data:

  mysql -u youruserl -p yourdatabase <bse.sql

You will be asked for your password.

=head2 Base data load

Change directory to the util directory in your workpath.

If you moved the files you will need to edit initial.pl to change the line:

  use lib '../cgi-bin/modules';

so that it will specify the correct modules directory.

You may want to simply set PERL5LIB to your new modules director instead:

  PERL5LIB=cgi-bin/modules
  export PERL5LIB

Once you've done that, you can run initial.pl:

  perl initial.pl

B<Warning:> Do not run this after your site is up and has articles you
want to keep.  This will delete all existing articles from your site.


=head1 TEMPLATE SETUP

See the site/docs/templates.pod (or .html).


=head1 SECURITY

Since your web host runs your CGI programs as you, you can make
Constants.pm readble only by you.

You will need to protect the htdocs/admin and cgi-bin/admin
directories with .htaccess files.  For example:

  AuthType Basic
  AuthName "Administrator Only"
  AuthUserFile "somepath/users.dat"
  require valid-user

The AuthUserFile needs to point to a file accessible by the user that
the web server runs as.  So it won't work if you use your home
directory.