1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
|
John,
I have been an extremely satisfied ADODB user for several years now.
To give you something back for all your hard work, I've spent the last 3
days rewriting the adodb-session.php code.
----------
What's New
----------
Here's a list of the new code's benefits:
* Combines the functionality of the three files:
adodb-session.php
adodb-session-clob.php
adodb-cryptsession.php
each with very similar functionality, into a single file adodb-session.php.
This will ease maintenance and support issues.
* Supports multiple encryption and compression schemes.
Currently, we support:
MD5Crypt (crypt.inc.php)
MCrypt
Secure (Horde's emulation of MCrypt, if MCrypt module is not available.)
GZip
BZip2
These can be stacked, so if you want to compress and then encrypt your
session data, it's easy.
Also, the built-in MCrypt functions will be *much* faster, and more secure,
than the MD5Crypt code.
* adodb-session.php contains a single class ADODB_Session that encapsulates
all functionality.
This eliminates the use of global vars and defines (though they are
supported for backwards compatibility).
* All user defined parameters are now static functions in the ADODB_Session
class.
New parameters include:
* encryptionKey(): Define the encryption key used to encrypt the session.
Originally, it was a hard coded string.
* persist(): Define if the database will be opened in persistent mode.
Originally, the user had to call adodb_sess_open().
* dataFieldName(): Define the field name used to store the session data, as
'DATA' appears to be a reserved word in the following cases:
ANSI SQL
IBM DB2
MS SQL Server
Postgres
SAP
* filter(): Used to support multiple, simulataneous encryption/compression
schemes.
* Debug support is improved thru _rsdump() function, which is called after
every database call.
------------
What's Fixed
------------
The new code includes several bug fixes and enhancements:
* sesskey is compared in BINARY mode for MySQL, to avoid problems with
session keys that differ only by case.
Of course, the user should define the sesskey field as BINARY, to
correctly fix this problem, otherwise performance will suffer.
* In ADODB_Session::gc(), if $expire_notify is true, the multiple DELETES in
the original code have been optimized to a single DELETE.
* In ADODB_Session::destroy(), since "SELECT expireref, sesskey FROM $table
WHERE sesskey = $qkey" will only return a single value, we don't loop on the
result, we simply process the row, if any.
* We close $rs after every use.
---------------
What's the Same
---------------
I know backwards compatibility is *very* important to you. Therefore, the
new code is 100% backwards compatible.
If you like my code, but don't "trust" it's backwards compatible, maybe we
offer it as beta code, in a new directory for a release or two?
------------
What's To Do
------------
I've vascillated over whether to use a single function to get/set
parameters:
$user = ADODB_Session::user(); // get
ADODB_Session::user($user); // set
or to use separate functions (which is the PEAR/Java way):
$user = ADODB_Session::getUser();
ADODB_Session::setUser($user);
I've chosen the former as it's makes for a simpler API, and reduces the
amount of code, but I'd be happy to change it to the latter.
Also, do you think the class should be a singleton class, versus a static
class?
Let me know if you find this code useful, and will be including it in the
next release of ADODB.
If so, I will modify the current documentation to detail the new
functionality. To that end, what file(s) contain the documentation? Please
send them to me if they are not publically available.
Also, if there is *anything* in the code that you like to see changed, let
me know.
Thanks,
Ross
|
