/!\ database schema has been modified, do a backup before updating
message have been refactored, here are the main changes:
- languages are now handled
- all messages have an uid (internal to SàT)
- message updating is anticipated
- subject is now first class
- new naming scheme is used newMessage => messageNew, getHistory => historyGet, sendMessage => messageSend
- minimal compatibility refactoring in quick_frontend/Primitivus, better refactoring should follow
- threads handling
- delayed messages are saved into history
- info messages may also be saved in history (e.g. to keep track of people joining/leaving a room)
- duplicate messages should be avoided
- historyGet return messages in right order, no need to sort again
- plugins have been updated to follow new features, some of them need to be reworked (e.g. OTR)
- XEP-0203 (Delayed Delivery) is now fully handled in core, the plugin just handle disco and creation of a delay element
- /!\ jp and Libervia are currently broken, as some features of Primitivus
It has been put in one huge commit to avoid breaking messaging between changes.
This is the main part of message refactoring, other commits will follow to take profit of the new features/behaviour.
log taints can be specified using log_levels_taints_dict in sat.conf, where key is a level name, and value is a list of color or code names. For instance, the following conf would taint INFO level in bold/green and DEBUG in default color:
log_levels_taints_dict = {"debug": [], "info": ["green", "bold"]}
Unknown names are copied as raw strings.
pyfeed and xe where still used only for date format handling, and there is already dateutil which is a dependency of Wokkel.
Furthermore pyfeed and xe are still not on pypi, causing troubles for installation with pip
- new --output option can be added wich use_output. use_output can be True (in which case it wild default to C.OUTPUT_TEXT), or any of C.OUTPUT_TYPES (currently text, list and dict)
- output change the output format mainly to make command result parsing more easy, but it can also be use to add fancy effects (like coloration)
- outputs are added with plugins in the same way as commands (import of both is done in the same method)
- in the command class, output need to be declared with use_output=C.OUTPUT_xxx, then the data only need to be processed with self.output(data)
- outputs can have description (not used yet)
- use_xxx argument handling has been refactored (minor refactoring) to be more generic
- first outputs are "default", "json" (pretty printed) and "json_raw" (compact json)
- the first command to use them is "profile list"
