Apache::Session(3pm)

1Session(3)            User Contributed Perl Documentation           Session(3)
2
3
4

NAME

6       Apache::Session - A persistence framework for session data
7

SYNOPSIS

9         use Apache::Session::MySQL;
10
11         my %session;
12
13         #make a fresh session for a first-time visitor
14         tie %session, 'Apache::Session::MySQL';
15
16         #stick some stuff in it
17         $session{visa_number} = "1234 5678 9876 5432";
18
19         #get the session id for later use
20         my $id = $session{_session_id};
21
22         #...time passes...
23
24         #get the session data back out again during some other request
25         my %session;
26         tie %session, 'Apache::Session::MySQL', $id;
27
28         &validate($session{visa_number});
29
30         #delete a session from the object store permanently
31         tied(%session)->delete;
32

DESCRIPTION

34       Apache::Session is a persistence framework which is particularly useful
35       for tracking session data between httpd requests.  Apache::Session is
36       designed to work with Apache and mod_perl, but it should work under CGI
37       and other web servers, and it also works outside of a web server alto‐
38       gether.
39
40       Apache::Session consists of five components: the interface, the object
41       store, the lock manager, the ID generator, and the serializer.  The
42       interface is defined in Session.pm, which is meant to be easily sub‐
43       classed.  The object store can be the filesystem, a Berkeley DB, a
44       MySQL DB, an Oracle DB, a Postgres DB, Sybase, or Informix. Locking is
45       done by lock files, semaphores, or the locking capabilities of the var‐
46       ious databases.  Serialization is done via Storable, and optionally
47       ASCII-fied via MIME or pack().  ID numbers are generated via MD5.  The
48       reader is encouraged to extend these capabilities to meet his own
49       requirements.
50
51       A derived class of Apache::Session is used to tie together the three
52       components.  The derived class inherits the interface from Apache::Ses‐
53       sion, and specifies which store and locker classes to use.
54       Apache::Session::MySQL, for instance, uses the MySQL storage class and
55       also the MySQL locking class. You can easily plug in your own object
56       store or locker class.
57

INTERFACE

59       The interface to Apache::Session is very simple: tie a hash to the
60       desired class and use the hash as normal.  The constructor takes two
61       optional arguments.  The first argument is the desired session ID num‐
62       ber, or undef for a new session.  The second argument is a hash of
63       options that will be passed to the object store and locker classes.
64
65       tieing the session
66
67       Get a new session using DBI:
68
69        tie %session, 'Apache::Session::MySQL', undef,
70           { DataSource => 'dbi:mysql:sessions' };
71
72       Restore an old session from the database:
73
74        tie %session, 'Apache::Session::MySQL', $session_id,
75           { DataSource => 'dbi:mysql:sessions' };
76
77       Storing and retrieving data to and from the session
78
79       Hey, how much easier could it get?
80
81        $session{first_name} = "Chuck";
82        $session{an_array_ref} = [ $one, $two, $three ];
83        $session{an_object} = new Some::Class;
84
85       Reading the session ID
86
87       The session ID is the only magic entry in the session object, but any‐
88       thing beginning with a "_" is considered reserved for future use.
89
90        my $id = $session{_session_id};
91
92       Permanently removing the session from storage
93
94        tied(%session)->delete;
95

BEHAVIOR

97       Apache::Session tries to behave the way the author believes that you
98       would expect.  When you create a new session, Session immediately saves
99       the session to the data store, or calls die() if it cannot.  It also
100       obtains an exclusive lock on the session object.  If you retrieve an
101       existing session, Session immediately restores the object from storage,
102       or calls die() in case of an error.  Session also obtains an non-exclu‐
103       sive lock on the session.
104
105       As you put data into the session hash, Session squirrels it away for
106       later use.  When you untie() the session hash, or it passes out of
107       scope, Session checks to see if anything has changed. If so, Session
108       gains an exclusive lock and writes the session to the data store.  It
109       then releases any locks it has acquired.
110
111       Note that Apache::Session does only a shallow check to see if anything
112       has changed.  If nothing changes in the top level tied hash, the data
113       will not be updated in the backing store.  You are encouraged to time‐
114       stamp the session hash so that it is sure to be updated.
115
116       When you call the delete() method on the session object, the object is
117       immediately removed from the object store, if possible.
118
119       When Session encounters an error, it calls die().  You will probably
120       want to wrap your session logic in an eval block to trap these errors.
121

LOCKING AND TRANSACTIONS

123       By default, most Apache::Session implementations only do locking to
124       prevent data corruption.  The locking scheme does not provide transac‐
125       tional consistency, such as you might get from a relational database.
126       If you desire transactional consistency, you must provide the Transac‐
127       tion argument with a true value when you tie the session hash.  For
128       example:
129
130        tie %s, 'Apache::Session::File', $id {
131           Directory     => '/tmp/sessions',
132           LockDirectory => '/var/lock/sessions',
133           Transaction   => 1
134        };
135
136       Note that the Transaction argument has no practical effect on the MySQL
137       and Postgres implementations.  The MySQL implementation only supports
138       exclusive locking, and the Postgres implementation uses the transaction
139       features of that database.
140

IMPLEMENTATION

142       The way you implement Apache::Session depends on what you are trying to
143       accomplish.  Here are some hints on which classes to use in what situa‐
144       tions
145

STRATEGIES

147       Apache::Session is mainly designed to track user session between http
148       requests.  However, it can also be used for any situation where data
149       persistence is desirable.  For example, it could be used to share
150       global data between your httpd processes.  The following examples are
151       short mod_perl programs which demonstrate some session handling basics.
152
153       Sharing data between Apache processes
154
155       When you share data between Apache processes, you need to decide on a
156       session ID number ahead of time and make sure that an object with that
157       ID number is in your object store before starting you Apache.  How you
158       accomplish that is your own business.  I use the session ID "1".  Here
159       is a short program in which we use Apache::Session to store out data‐
160       base access information.
161
162        use Apache;
163        use Apache::Session::File;
164        use DBI;
165
166        use strict;
167
168        my %global_data;
169
170        eval {
171            tie %global_data, 'Apache::Session::File', 1,
172               {Directory => '/tmp/sessiondata'};
173        };
174        if ($@) {
175           die "Global data is not accessible: $@";
176        }
177
178        my $dbh = DBI->connect($global_data{datasource},
179           $global_data{username}, $global_data{password}) ⎪⎪ die $DBI::errstr;
180
181        undef %global_data;
182
183        #program continues...
184
185       As shown in this example, you should undef or untie your session hash
186       as soon as you are done with it.  This will free up any locks associ‐
187       ated with your process.
188
189       Tracking users with cookies
190
191       The choice of whether to use cookies or path info to track user IDs is
192       a rather religious topic among Apache users.  This example uses cook‐
193       ies.  The implementation of a path info system is left as an exercise
194       for the reader.
195
196       Note that Apache::Session::Generate::ModUsertrack uses Apache's
197       mod_usertrack cookies to generate and maintain session IDs.
198
199        use Apache::Session::MySQL;
200        use Apache;
201
202        use strict;
203
204        #read in the cookie if this is an old session
205
206        my $r = Apache->request;
207        my $cookie = $r->header_in('Cookie');
208        $cookie =~ s/SESSION_ID=(\w*)/$1/;
209
210        #create a session object based on the cookie we got from the browser,
211        #or a new session if we got no cookie
212
213        my %session;
214        tie %session, 'Apache::Session::MySQL', $cookie, {
215             DataSource => 'dbi:mysql:sessions', #these arguments are
216             UserName   => 'mySQL_user',         #required when using
217             Password   => 'password',           #MySQL.pm
218             LockDataSource => 'dbi:mysql:sessions',
219             LockUserName   => 'mySQL_user',
220             LockPassword   => 'password'
221        };
222
223        #Might be a new session, so lets give them their cookie back
224
225        my $session_cookie = "SESSION_ID=$session{_session_id};";
226        $r->header_out("Set-Cookie" => $session_cookie);
227
228        #program continues...
229

AUTHORS

239       Jeffrey Baker <jwbaker@acm.org> is the author of Apache::Session.
240
241       Tatsuhiko Miyagawa <miyagawa@bulknews.net> is the author of Gener‐
242       ate::ModUniqueID and Generate::ModUsertrack
243
244       Erik Rantapaa <rantapaa@fanbuzz.com> found errors in both Lock::File
245       and Store::File
246
247       Bart Schaefer <schaefer@zanshin.com> notified me of a bug in
248       Lock::File.
249
250       Chris Winters <cwinters@intes.net> contributed the Sybase code.
251
252       Michael Schout <mschout@gkg.net> fixed a commit policy bug in 1.51.
253
254       Andreas J. Koenig <andreas.koenig@anima.de> contributed valuable CPAN
255       advice and also Apache::Session::Tree and Apache::Session::Counted.
256
257       Gerald Richter <richter@ecos.de> had the idea for a tied hash interface
258       and provided the initial code for it.  He also uses Apache::Session in
259       his Embperl module and is the author of Apache::Session::Embperl
260
261       Jochen Wiedmann <joe@ipsoft.de> contributed patches for bugs and
262       improved performance.
263
264       Steve Shreeve <shreeve@uci.edu> squashed a bug in 0.99.0 whereby a
265       cleared hash or deleted key failed to set the modified bit.
266
267       Peter Kaas <Peter.Kaas@lunatech.com> sent quite a bit of feedback with
268       ideas for interface improvements.
269
270       Randy Harmon <rjharmon@uptimecomputers.com> contributed the original
271       storage-independent object interface with input from:
272
273         Bavo De Ridder <bavo@ace.ulyssis.student.kuleuven.ac.be>
274         Jules Bean <jmlb2@hermes.cam.ac.uk>
275         Lincoln Stein <lstein@cshl.org>
276
277       Jamie LeTaul <jletual@kmtechnologies.com> fixed file locking on Win‐
278       dows.
279
280       Scott McWhirter <scott@surreytech.co.uk> contributed verbose error mes‐
281       sages for file locking.
282
283       Corris Randall <corris@line6.net> gave us the option to use any table
284       name in the MySQL store.
285
286       Oliver Maul <oliver.maul@ixos.de> updated the Sybase modules
287
288       Innumerable users sent a patch for the reversed file age test in the
289       file locking module.
290
291       Langen Mike <mike.langen@tamedia.ch> contributed Informix modules.
292
293
294
295perl v5.8.8                       2004-02-24                        Session(3)