|
|
The purpose of DBIx::Safe is to give controlled, limited access to an
application, rather than simply passing it a raw database handle through DBI.
DBIx::Safe acts as a wrapper to the database, by only allowing through the
commands you tell it to. It filters all things related to the database handle -
methods and attributes.
The typical usage is for your application to create a database handle via a
normal DBI call to new(), then pass that to DBIx::Safe->new(), which will return
you a DBIx::Safe object. After specifying exactly what is and what is not
allowed, you can pass the object to the untrusted application. The object will
act very similar to a DBI database handle, and in most cases can be used
interchangeably.
By default, nothing is allowed to run at all. There are many things you can
control. You can specify which SQL commands are allowed, by indicating the first
word in the SQL statement (e.g. 'SELECT'). You can specify which database
methods are allowed to run (e.g. 'ping'). You can specify a regular expression
that allows matching SQL statements to run (e.g. 'qr{SET TIMEZONE}'). You can
specify a regular expression that is NOT allowed to run (e.g. qr(UPDATE xxx}).
Finally, you can indicate which database attributes are allowed to be read and
changed (e.g. 'PrintError'). For all of the above, there are matching methods to
remove them as well.
WWW: http://search.cpan.org/dist/DBIx-Safe/
Feature safe: yes
Notes:
svn path=/head/; revision=286921
|
