summary |
shortlog | log |
commit |
commitdiff |
tree
first ⋅ prev ⋅ next
Helmut Grohne [Fri, 19 Jul 2013 12:56:05 +0000 (14:56 +0200)]
ThrottledDevice: speed up quick reactivation
When a deactivation is delayed by offdelay and reactivated before the
timer expires, the device is not powered down at all in between. So we
should not incur the ondelay, but immediately transition to ST_ACTIVE
after cancelling the pending deactivation.
Helmut Grohne [Mon, 15 Jul 2013 08:39:06 +0000 (10:39 +0200)]
fix markdown syntax
Helmut Grohne [Mon, 15 Jul 2013 08:28:32 +0000 (10:28 +0200)]
README.md: document dependencies
Helmut Grohne [Mon, 15 Jul 2013 08:12:52 +0000 (10:12 +0200)]
move dbus introspection to dbusutils.list_objects
Helmut Grohne [Mon, 15 Jul 2013 08:01:13 +0000 (10:01 +0200)]
sispmctl_device: permit selecting a device
Helmut Grohne [Mon, 15 Jul 2013 05:13:11 +0000 (07:13 +0200)]
README.md: put the implementations in an itemize
Suggested by Joachim Breitner.
Helmut Grohne [Sun, 14 Jul 2013 19:04:05 +0000 (21:04 +0200)]
rewrite README.md
Thanks to Joachim Breitner for asking good questions. Most of his
questions are answered by this revised version.
Helmut Grohne [Sun, 14 Jul 2013 19:02:46 +0000 (21:02 +0200)]
fix epydoc markup
Helmut Grohne [Sun, 14 Jul 2013 18:54:38 +0000 (20:54 +0200)]
add a sispmctl_device
Helmut Grohne [Sun, 14 Jul 2013 15:09:04 +0000 (17:09 +0200)]
dbus_client: handle multiple transition signals
There bay be multiple changestate signals with ST_ACTIVE|ST_TRANSITION
before a signal for ST_ACTIVE.
Helmut Grohne [Sun, 14 Jul 2013 06:59:33 +0000 (08:59 +0200)]
must not pass float to GObject.timeout_add
Helmut Grohne [Sun, 14 Jul 2013 06:42:34 +0000 (08:42 +0200)]
ThrottledDevice: missing state notifiation
When a throttled device is transitioning to off and we activate it,
activation may be a noop on the actual device thus not emitting any
state change. From a users point of view we are transitioning to on now,
so the state has changed.
Helmut Grohne [Sat, 13 Jul 2013 09:54:24 +0000 (11:54 +0200)]
generalize ThrottledDevice
It gained the capability to delay the activation signal by a specified
amount of time. This should be useful when using it with sispmctl,
because hardware usually takes significant time to power on.
Helmut Grohne [Sat, 6 Jul 2013 11:42:08 +0000 (13:42 +0200)]
add an example dbus_policy.conf
This can be used when running onoff on the system bus. It should be
placed as /etc/dbus-1/system.d/onoff.conf then.
Helmut Grohne [Fri, 28 Jun 2013 10:38:37 +0000 (12:38 +0200)]
dbus_client: add --list to list devices
Helmut Grohne [Fri, 28 Jun 2013 09:46:29 +0000 (11:46 +0200)]
provide a default for --busname
Helmut Grohne [Fri, 28 Jun 2013 09:35:58 +0000 (11:35 +0200)]
move OnoffControl to onoff.dbusutils
What's left in dbus_service.py now is mainly the configuration part.
Maybe configuring onoff should happen by writing a python script for
now?
Helmut Grohne [Sat, 22 Jun 2013 17:01:39 +0000 (19:01 +0200)]
document device configuration
Helmut Grohne [Sat, 22 Jun 2013 11:33:43 +0000 (13:33 +0200)]
change activatefd method and add ThrottledDevice
When activating via activatefd, there is no reason to add an additional
delay. Simply close the socket later. If a device should delay
deactivation, this is a policy decision, that should be made by the
service. This policy decision can be done using the new ThrottledDevice.
Helmut Grohne [Wed, 19 Jun 2013 18:17:47 +0000 (20:17 +0200)]
move the socketpair utility to dbusutils
Helmut Grohne [Wed, 19 Jun 2013 18:15:57 +0000 (20:15 +0200)]
add a README
Outline what onoff is about and what problem it aims to solve.
Helmut Grohne [Wed, 19 Jun 2013 17:44:14 +0000 (19:44 +0200)]
write down some ideas
Helmut Grohne [Wed, 19 Jun 2013 10:00:28 +0000 (12:00 +0200)]
license the code as BSD-3
Helmut Grohne [Tue, 18 Jun 2013 21:59:19 +0000 (23:59 +0200)]
add minimal setup.py
Helmut Grohne [Tue, 18 Jun 2013 21:42:47 +0000 (23:42 +0200)]
dbus_client: improve documentation
Helmut Grohne [Tue, 18 Jun 2013 21:37:33 +0000 (23:37 +0200)]
dbus_client: make duration configurable
Helmut Grohne [Tue, 18 Jun 2013 20:22:27 +0000 (22:22 +0200)]
make clients configurable via argparse
Helmut Grohne [Mon, 17 Jun 2013 21:10:29 +0000 (23:10 +0200)]
ship some documentation
Helmut Grohne [Mon, 17 Jun 2013 18:50:46 +0000 (20:50 +0200)]
add mpd_watcher
It activates a device when mpd plays music.
Helmut Grohne [Mon, 17 Jun 2013 18:15:59 +0000 (20:15 +0200)]
change activatefd signature
The method no longer takes a UnixFd, but returns one. This is copied
from the systemd inhibitor interface. Advantages:
* Not having fds to arbitrary things in the service is probably a good
thing.
* Maybe the interfaces can converge?
Helmut Grohne [Mon, 17 Jun 2013 17:56:37 +0000 (19:56 +0200)]
version dbus interface
The interface is still subject to change. Choosing version 0 as an
unstable version now. Change to 1 on initial release.
Helmut Grohne [Mon, 17 Jun 2013 11:32:49 +0000 (13:32 +0200)]
deduplicate spawn_async invocations
Helmut Grohne [Mon, 17 Jun 2013 10:28:21 +0000 (12:28 +0200)]
add an OnoffCommand device
Helmut Grohne [Mon, 17 Jun 2013 07:47:36 +0000 (09:47 +0200)]
OnoffProcess: don't fail inopportune transitions
Previously a call to activate while transitioning to ST_INACTIVE would
be ignored.
Helmut Grohne [Sun, 16 Jun 2013 22:39:07 +0000 (00:39 +0200)]
attempt to fix for python3
* Wrong way to invoke logging.
* spawn_async returns pid as 0. Bug in library?
Helmut Grohne [Sun, 16 Jun 2013 22:04:17 +0000 (00:04 +0200)]
implement a redshift device
Helmut Grohne [Sun, 16 Jun 2013 22:03:44 +0000 (00:03 +0200)]
draft device api
Helmut Grohne [Sun, 16 Jun 2013 22:02:06 +0000 (00:02 +0200)]
dbus_service: swap print for logging
Helmut Grohne [Sun, 16 Jun 2013 21:57:41 +0000 (23:57 +0200)]
service: more generic activation
Helmut Grohne [Sun, 16 Jun 2013 21:51:09 +0000 (23:51 +0200)]
improve dbus api
* rename state signal to changestate
* provide a state method returning the current state
Helmut Grohne [Sun, 16 Jun 2013 19:55:56 +0000 (21:55 +0200)]
change state meanings
The ST_ACTIVE bit is now to define the desired state. The previous
definition of using the state we are coming from does not make much
sense. Consider the case when being deactivated while transitioning to
active. Then we are coming from inactive and transitioning to inactive.
Helmut Grohne [Sun, 16 Jun 2013 13:33:01 +0000 (15:33 +0200)]
port to python3
Helmut Grohne [Sun, 16 Jun 2013 10:32:08 +0000 (12:32 +0200)]
initial commit
Proof-of-concept for the dbus interaction of the onoff service.