BlogItem is doing deserialisation itself, which is not clean (it should be done at bridge level), and which is causing trouble in jp which doesn't expect serialisation.
As a workaround, the "deserialise" argument has been added for now (default to True), until deserialisation is properly managed in bridge.
So far, and for historical reasons, blog items data where serialised using a unicode: unicode dict, which was causing trouble for many types of values (timestamps, booleans, lists).
This patch changes it by serialising the whole items before going to bridge, and deserialising it when going back. This way, complex data can be used easily in items.
This impact psEvent and serItemsData* methods which are renamed transItemsData* because there are not always serialising anymore (a new argument "serialise" allows to specify it).
When editing a blog post in jp, metadata are now more easy to manipulate, specially lists like tags.
XHTMLBox is a textbox specialised in XHTML, i.e. it renders the XHTML when in read_only, and it allows to edit it.
The XHTML is cleaned by default, the cleaning is done by Text Syntaxes plugin (actually there is a cleaning method which can be set by any plugin, but Text Syntaxes is the one which does it).
metadata are parsed to deserialise some well known values like rsm index and count, and some properties have been added to have easier access.
A "complete" property (or item in metadata) is set to True if we are on the last page, False if we are not, or None if we don't have enough data to know if we are on the last page or not.
- new date_parse_ext method parse a date where relative informations can be added. e.g.: "1 week ago", "now - 2 days", "01/01/01 01:01 + 3 hours".
- new TZ_UTC and TZ_LOCAL constants can be used to use UTC time or local time zone.
- date_parse* can now use a default_tz, which will be used if timezone if not specified in date format (default UTC).
- tz_info can now be used in date_format to format the time in given timezone (default: UTC).
Environment.get_template has been overriden to have always full template name (i.e. template with (site/theme) prefix or absolute template).
This is needed to correctly handle cache, so "base/base.html" is converted to "(some_site/some_theme)base/base.html" and won't conflict with an other theme "base/base.html".
- site can now be specified in template header before theme, for instance: (some_site/some_theme)path/to/template.ext
- absolute template paths are now implemented, but Renderer must be instanciated with trusted to True for security reason (it's the case for jp)
- a new "front_url_filter" callable can be given to Renderer, which will convert template path to URL seen by end-user (default to real path).
- the "front_url_filter" can be used in templates with… "front_url" filter
- template_data is a new named tuple available in templates, which give site, theme and template relative URL
- search order is site/theme, site/default_theme, and default/default_theme where default link to sat_pubsub templates
- when loading CSS files, files with _noscript suffixes are now loaded, and used when javascript is not available
- "styles_extra.css" is also loaded before "styles.css", useful when a theme want to reuse default style, and just override some rules
- new site can be specified in sat.conf [DEFAULT] section, using sites_path_public_dict or sites_path_private_dict (where sites_path_private_dict won't be used in public frontends, like Libervia)
- "private" argument of Renderer tells the renderer to load private sites or not
- templates are now loaded from "templates" subdirectory, to differenciate them from other data like i18n
- jp template output has been updated to handle those changes, and to manage absolute templates
class_map was so far used as a global value, but in some cases, several XMLUI managers may be used (it's the case in Cagou's remote controllers plugins, template_xmlui is used to manage remotes).
This patch set class_map as a create argument, this way it can be set with partial, and several XMLUI managers can be instanced at the same time.
Async trigger manager add an asyncPoint method which can be used with Deferred (or sync method).
This allow a triggers do to some long operations like doing network requests.
MessageReceived and sendMessageData now use asyncPoint.
- removed save_dir_path and .hg_data/pickle method: it is not needed to save repository data anymore as its version is now saved on installation by setuptools_scm
- added "distance" (number of commits since last tag)
- use twisted's which method to find the hg executable
- result is now cached on first execution
- fixed dirstate method
- at last resort, try to find the repository version using the package version, as it is saved there by setuptools_scm
