Using a simple, hysteresis based approach… currently in this first
version the optimizer recursively moves modules out of a thread until a
certain timing target is reached; the "modules moved out" just mentioned
are then partitioned into new threads according to their latest time profile.
Very simple, but very effective.
Todo: Implement merging threads with (again) good behaving workloads.
I thought: Well, isn't it a bit redundant if I go ahead and say in my
config "import this-n-that and temp and load and alsa" while later
just giving register() the modules.
So register() (/ClassFinder as the backend) now support just naming a
module as the first parameter. That module is then imported and
searched for a class as usual.
Just for reference the synopsis of Status.register():
register(mod.SomeChecker())
register(mod, setting1=..., setting2=...)
register(mod, {"setting1:":, "setting2":...})
register("mod", setting=1..., setting2=...)
register("mod", {"setting1:":, "setting2":...})
Fun fact: Actually register() doesn't care for it's arguments.
They're passed straight into ClassFinder.instanciate_class_from_module
(something I should definitely rename), which checks if it's first
parameter is one of:
-Python module
=> It calls ClassFinder.get_class with the same parameters
=> get_class will search the module using ClassFinder.search_module
and return a matching class if and only if there is a single matching
class in the module
-string
=> It calls ClassFinder.get_module to import the module and calls itself
on the result
-something else
=> It returns that something
The actual variation in passing the settings (keyword arguments vs. dict)
is handled in SettingsBase.
ClassFinder:
Removed exclude argument, which is basically covered with
obj.__module__ == module.__name__ which ensures that no imported
classes are found, which was the only use case for exclude.
This didn't change the public "API" used by modules.
.core.io contains the IO classes, namely IOHandler, StandaloneIO and JSONIO
.core.util contains SettingsBase and ClassFinder
.core.exceptions contains all custom exceptions
This is 100pct. the same functionality as a complete module before :-)
But still, have to come up with some better way to manage these "templates".
And a place to stash them.
So I can have my old representation back... indeed you can
do pretty much anything now with it:
status.register(battery,
format="{consumption:.2f}W {percentage:.2f}% [{percentage_design:.2f}%] {remaining_hm}"
)
Mail checkers are not separate modules anymore; there is only one
mail checker module (called mail) that can make use of various backends
, which were formerly modules on their own.
This greatly simplifies code and reduces redundance.
This commit only contains the base classes for this.
You can now pass a module (the Python thing, e.g. modsde) into register
and it will just "do the right thing" (find a class derived from Module
in that module and instanciate it with any extra arguments)
Btw. nothing has changed for the config files. You can pass into
__init__ a dict as before, or you can use keyword arguments.
Either way, a module defines the settings that can be specified
and those which are required. __init__ basically checks that all
required options are set and that no invalid options are used.
Specify standalone=True to the I3statushandler constructor
to run i3pystatus without i3status in the background.
i3pystatus won't read input from stdin or any other
file object specified with input_stream.
The keyword argument interval specifies how often output should
be generated. The default is 1 (second).
Sorry guys for changing the way i3pystatus "way of operation"
is set so often. If you're want the "self-contained" mode
(you execute i3pystatus, i3pystatus automatically starts
i3status), don't set the file attribute, but pass the file
descriptor of the pipe as input_stream like this:
process = subprocess.Popen(["i3status", "-c", "~/.i3/status"], stdout=subprocess.PIPE, universal_newlines=True)
status = i3pystatus(input_stream=process.stdout)
On a side note:
The main class name has been changed to i3pystatus, but I3statusHandler
is still available as an alias. Use whichever you prefer :-)
(Linux is about choice after all)
Notmuch checker does not work in this state
NotmuchMailChecker does not work because of a conflict between the notmuch Python API and the file for NotmuchMailChecker which is also named notmuch.
I suggest to rename i3pystatus/notmuch.py to i3pystatus/notmuchchecker.py and modify main.py.dist appropriately.
Reported by mjepronk
Also removed some superfluous whitespace
If one wants to run i3status as a subprocess of
i3pystatus, do it like this:
import subprocess
import io
...
process = subprocess.Popen(["i3status", "-c", "~/.i3/status"], stdout=subprocess.PIPE)
stdin = io.TextIOWrapper(process.stdout)
status = I3statusHandler(stdin)
exploited_language_features += 2;
(I also exploit the mutability of the list-object here,
yield j binds the list to the context,
when the context is leaved execution continues and the
modified j is written back)
Allows to run i3status directly from your __main__, like this:
status.register(...) # and so on
process = subprocess.Popen("i3status", stdout=subprocess.PIPE)
status.fd = process.stdout
# start the handler
status.run()
