From Cerberus Helpdesk Wiki

Cerb4 (4.2.3) was released as a stable update on August 3rd 2009 and contains 23+ improvements from community feedback.

This page will highlight the most interesting changes. You can find the full changelog in the forums.

Contents 1 Highlights of the 4.2.3 Release 1.1 Mail Filters 1.2 Watchers 1.3 Usability 1.4 Support Center (SC) 1.5 Antispam

Highlights of the 4.2.3 Release

Mail Filters

• Plugins can now contribute filter criteria extensions. This compliments the action extensions, as criteria can use any combination of characteristics from a message (body, headers, attachments) to determine if something matches, before handing it to actions to actually do it. A menu-driven interface will always be severely limited compared to what you can accomplish using a scripting language. Now you can create filters to match: combinations of SpamAssassin mail headers with specific values; messages containing 'magic phrases'; messages with specific attachments (that themselves contain specific content); messages that routed along a specific path; messages in specific languages that are detected heuristically; messages that correlate information from a 3rd party system (e.g. billing balance, LDAP); and countless other things. (See: http://vimeo.com/channels/cerb4#5810661)

  

  

  

• [CDH-1333] Plugins can now contribute action extensions on Mail Filters to implement any kind of business logic. As mentioned before, the impetus came from Geoff on the Cerb4 forums asking for a way to look up the phone number from voicemail messages sent to the helpdesk by email and rewrite the 'From:' to the proper Address Book contact. You could also do a million other things, which we'll begin to share. For example: you could provide another level of spam filtering before any mail is accepted (e.g. SpamAssassin); you could verify a sender had an SLA; you could strip specific attachments; you could write the message body to strip excessive quoted content; and so on. (See: http://vimeo.com/channels/cerb4#5810661)

  

  

  

  

• You can now get/set any number of custom properties for Mail Filter criteria and action extensions coming from plugins. These properties will let you *reuse* criteria + actions across multiple filters, with the ability to configure them for each situation. It also means that developers can create and share plugins that non-developers can download and configure without needing to touch any source code. For example: If you want to make a custom action that sets a message header on matching mail, you can create two properties on *one* action that define the header name and the value to set. Without properties you'd have to profligately create an action per desired combination. (See: http://vimeo.com/channels/cerb4#5810661)

  

  

• Pre-parser filters can now read and modify anything about incoming messages: attachments, headers, body. Previously all filters had to reject mail.
• Pre-parser filters have a new "Stop, and Continue delivery" action that will stop processing filters without rejecting a message.
• Pre-parser filters can now stack. This means you can have several filters with actions that each modify part of a message.
• Fixed an issue where Mail Filters wouldn't save if you removed all the actions.

Watchers

• [CHD-365] Superusers can now globally display and edit other workers' Watcher filters from Configuration->Watchers.

  

• Watcher filters can now be set to disabled to preserve their setup while ignoring them. Previously, filters had to be either deleted or their actions had to be removed for them to not trigger.
• [CHD-1249] Disabling a worker now properly disables their existing Watcher filters. The upgrade script will retroactively disable any filters with disabled owners.
• Renamed "E-mail Notifications" to "Watchers" to remove ambiguity between Watchers (copies of mail) and Worker Notifications (ticket/task assignments).
• Added the ability to bulk update Watcher filters from both Preferences and Configuration.
• Watcher filters are now displayed in a sortable/searchable/pageable/customizable worklist. You can 'peek' to display their detailed configuration.
• You can now have a Watcher Filter trigger a Worker Notification in addition to (or instead of) forwarding a copy of the matching email to the worker's private email address. This is especially useful when combined with the Fluid.app+Growl plugin, as you can set up desktop notifications based on any filterable criteria. For example: You could receive desktop alerts for replies to tickets assigned to you; or alerts for tickets opened by customers with an SLA; or alerts for tickets opened during certain days and hours (when you're on shift). The possibilities are endless. The ability to create Worker Notifications will also be handy when we push notifications through the Web-API and read it with various devices, like the iPhone.

  

  

  

Usability

• [CHD-1069] Workers are now sorted in alphabetical order based on first name (given name) rather than last name (surname).
• [CHD-1324] Current KB page numbers are reused when going from topic to topic.
• [CHD-1322] Clicking '(peek)' on the source tickets from a merge throws an error.
• Mail Routing now properly makes sure groups are valid. There were some cases where groups set as destinations could be deleted and mail would still try to deliver to them. A database patch resurrects these tickets.

Support Center (SC)

• [CHD-1329] If a customer opens a new ticket and enters zero for a 'number' custom field then it will not be "saved" in the ticket.
• It's possible that Pre-parser Mail Filters may intentionally reject mail sent by 'Open Ticket' from the Support Center. A bug has been fixed in these situations so an error isn't displayed when attempting to display the newly created ticket ID.
• [CHD-333] Allow manual sorting of Contact Situations display order.
• [CHD-460] You can now require answers to certain follow-up questions from contact situations on Open Ticket by prefixing them with an asterisk (*). For example: "*Which product are you having trouble with?"

Antispam

• Fixed several little issues with the anti-spam system. Forcing words to lowercase with various encodings could corrupt the text (only for spam analysis). The string functions for spam analysis have been converted to their mb_* (mbstring) equivalents. Words weren't always restricted to being composed of alphanumeric characters, which made it easier for some spam to look more unique than it should be (e.g. assorted Cyrillic characters were making it into words). It was possible for some words to be counted twice if their encoding had problems being converted to lowercase; now the word results will be forced lowercase even if they snuck into the database with mixed case. A database patch cleans up the 'bayes_words' table: unused words are purged; nonalpha words are purged; duped words are merged.