new methods psNodeWatchAdd and psNodeWatchRemove allows to set a watch for the time of the session on one node, to have a signal called when something change on this node.
This signal (psEventRaw) send raw data (raw XML), in opposition to psEvent which is there to send high level data (e.g. parsed blog data).
Those method are primarely there to let frontends manage local cache for pubsub nodes.
- new internal registerNamespace can be used by a plugin to associate a short name to a namespace
- new NamespacesGet bridge method retrieve those associations
those methods are high level methods specialised for tickets.
ticketsSet will use default tickets node if node is not set, set "created" and "updated" fields, create comments node if the ticket is new, and associate it with "comments_uri" field.
ticketsSchemaGet is like getUISchema with node defaulting to tickets default node.
the date filter is now named "date_fmt" and can be use to specify several useful formats (see docstring), using the current locale.
The "relative" format return the whole human readable relative date.
The new in_the_past test allow to check if a date is passed.
following flags have been added:
- noselect: list selection is not allowed (options are the important data)
- extensible: user can add elements to the list
- reducible: user can remove elements from the list
- inline: presentation hint: the list should be displayed horizontally
- mapping is an option allowing do map imported field to a specific field in the newly created item
- importers can now use complexe data types in options using JSON
- comments are put in "comments" key in import data, it's a list of microblog data
- in resulting item, the comments uri is put in comments_uri field
- (plugin import) root_service and root_node are kept in session
- (plugin ticket) reporter name is now in "reporter" key instead of "reporter_name"
- (plugin bugizlla): if not reporter name is found, first part of email address is used
gidx methods now use the filtered name to return global indexes.
For instance, if "widget" is filtered name, first index will be "widget", then "widget_1", "widget_2" and so one.
If an other name is used, it restart (e.g. for "label": "label", "label_1", "label_2", etc.).
