IO::Iron Policies - No Typing Errors to Iron.io Services!
Policies is a way to limit the names of message queues, code packages, caches and items (item keys) to a predefined group of possible strings. This can limit the chances for typos and enforce an enterprise policy. The policies are loaded from a JSON file which is specified either when creating a IO::Iron::Iron*::Client object, or in the config file .iron.json (or equivalent).
Policies in Config file
Add the item policies to the config file. The value of the item is the file name of the policies file.
Example config file:
{
"project_id":"51bdf5fb2267d84ced002c99",
"token":"-Q9OEHZPhdZtd0KHBzzdUJIqV_E",
"host":"cache-aws-us-east-1.iron.io",
"policies":"iron_policies.json"
}
Policies file specified when creating the client
my $policies_filename = '/etc/ironmq/global_policies.json';
my $client = IO::Iron::IronCache::Client->new('policies' => $policies_filename);
Examples of Policies File and Explanation of Configuration
The 'default' policies JSON file:
(
'definition' => {
'character_group' => {
},
'no_limitation' => 1, # There is an unlimited number of alternatives.
},
'queue' => { 'name' => [ '[:alnum:]{1,}' ], },
'cache' => {
'name' => [ '[:alnum:]{1,}' ],
'item_key' => [ '[:alnum:]{1,}' ]
},
'worker' => { 'name' => [ '[:alnum:]{1,}' ], },
);
The above file would set an open policy for IronMQ, IronCache and IronWorker alike. The file is divided into four parts: definition for defining meta options, and queue|cache|worker parts for defining the changing strings (queue|cache|worker names and item keys). The character group alnum covers all ascii alphabetic characters (both lower and upper case) and digits (0-9).
N.B. The option no_limitation controls the open/closed policy. If no_limitation is set (1=set), the policy control is turned off.
An example of policies file
{
"__comment1":"Use normal regexp. [:digit:] = number:0-9, [:alpha:] = alphabetic character, [:alnum:] = character or number.",
"__comment2":"Do not use end/begin limitators '^' and '\$'. They are added automatically.",
"__comment3":"Note that character groups are closed inside '[::]', not '[[:]]' as normal POSIX groups.",
'definition' => {
'character_group' => {
"[:lim_uchar:]":"ABC",
"[:low_digit:]":"0123"
},
},
"cache":{
"name":[
"cache_01_main",
"cache_[:alpha:]{1}[:digit:]{2}"
],
"item_key":[
"item.01_[:digit:]{2}",
"item.02_[:lim_uchar:]{1,2}"
]
}
}
This policies file sets policies for cache names and item keys. Both have two templates. Template "cache_01_main" is without wildcards: the template list can also only contain predefined names or keys. Sometimes this could be exactly the wanted behaviour, especially in regard to cache and message queue names.
Items beginning with '__' are considered comments. Comments can not be
inserted into lists, such as I The definition part contains the list character_group for user-defined
groups. The following groups are predefined:
All lower ASCII (7-bit) characters are allowed in names and in character
groups, except for the reserved characters (RFC 3986):
A character group definition is closed inside characters '[::]',
not '[[:]]' as normal POSIX groups. Only the equivalents of the POSIX groups
mentioned above can be used; e.g. POSIX group [[:graph:]] is not available.
When using the character groups in a name or key, only two markings are allowed:
[:group:]{n} and [:group:]{n,n}, where 'n' is an integer.
This limitation (not being able to use any regular expression) is due to the
double functionality of the policy: a) it acts as a filter when creating
and naming new message queues, code packages, caches and cache items; 2) it
can be used to list all possible names, for example when quering for
cache items.
!$&'()*+,;=:/?#[]@
