- Jun 18, 2014
-
-
Zeeshan Ali Khan authored
-
Zeeshan Ali Khan authored
Clients that are system components should not have to be authorized by agent.
-
Zeeshan Ali Khan authored
We automatically stop all clients if agent sets the max accuracy to 0 but system components should be an exception here.
-
Zeeshan Ali Khan authored
-
- Jun 17, 2014
-
-
Zeeshan Ali Khan authored
-
Zeeshan Ali Khan authored
-
Zeeshan Ali Khan authored
We only needed it for parsing replies from opencellid.org.
-
Zeeshan Ali Khan authored
We now use Mozilla Location Service for 3G-geolocation as well.
-
Zeeshan Ali Khan authored
-
Zeeshan Ali Khan authored
-
Zeeshan Ali Khan authored
Put the code dealing with Mozilla Location Service to a separate module with public API. This is so that we can later re-use most of this code from 3G module for switching from opencellid to Mozilla.
-
Zeeshan Ali Khan authored
It always made sense for this class to derive from WebSource but it wasn't possible before as it already derived from ModemSource class.
-
Zeeshan Ali Khan authored
-
- Jun 10, 2014
-
-
Zeeshan Ali Khan authored
Modem classes now directly use GClueModem and therefore not inherit from GClueModemSource anymore.
-
Zeeshan Ali Khan authored
Instead of deriving from GClueModemSource class, GClueModemGPS now just uses GClueModem singleton to talk to ModemManager. Refer to 9231e644 ("Add GClueModem class") for rationale.
-
Zeeshan Ali Khan authored
Instead of deriving from GClueModemSource class, GClue3G now just uses GClueModem singleton to talk to ModemManager. Refer to 9231e644 ("Add GClueModem class") for rationale.
-
Zeeshan Ali Khan authored
Add a new class which is going to be responsible for dealing with ModemManager for us. This is so that: * We have ModemManager code isolated from rest of the code so its easy to add support for other backends (e.g ofono) in future. * We have only one entity dealing with modem. In the following patch(es), we'll drop GClueModemSource baseclass and, GClueModemGPS and GClue3G will just use GClueModem for talking to modem. * GClue3G can then derive from GClueWebSource instead and be more easily modified to submit cell tower data to Mozilla Location Service. * GClueModemGPS can be easily generalized so that it can be more easily modifed to talk to standalone GPS devices too in future (we'll probably want to change it's name to GClueGPS then).
-
- Jun 06, 2014
-
-
Zeeshan Ali Khan authored
It will be needed in the following patch.
-
- Jun 03, 2014
-
-
Zeeshan Ali Khan authored
This mostl likely means that user disabled geolocation (globally or for the client specifically) so there is no point in keeping it running.
-
Zeeshan Ali Khan authored
An agent usually outlives clients so if the client object remains hooked to agent signal, its handler will still be called on signals from agent and cause a crash since we assume client object to be still around.
-
Zeeshan Ali Khan authored
-
Zeeshan Ali Khan authored
This patch fixes our ability not to timeout if service is in use by system components.
-
Zeeshan Ali Khan authored
Unlike the D-Bus 'InUse' property, this doesn't differentiate between system components and apps.
-
Zeeshan Ali Khan authored
Lets keep this property for apps only and therefore don't report location in use for system components. With this change, gnome-shell won't permanently show location icon in tray if automatic timezone is enabled. One thing this patch breaks is not to timeout if service is in use by system components. It'll be fixed in the following patches.
-
- Jun 02, 2014
-
-
Zeeshan Ali Khan authored
If set to 'true', it means that app is more a part of system than an app-app. You know what I mean. :) Also mark gnome-datetime-panel as system component.
-
- May 25, 2014
-
-
Zeeshan Ali Khan authored
This issue is not resolved so no need to keep a workaround for it: https://github.com/mozilla/ichnaea/issues/165
-
Zeeshan Ali Khan authored
Now that we don't use NetworkManager, we don't need to get ourselves access to it's D-Bus API.
-
Zeeshan Ali Khan authored
-
- May 24, 2014
-
-
Zeeshan Ali Khan authored
Since now wifi interface proxy is created async in Wifi (the only inheriting class), calling _refresh from _start() could easily mean that location query is created while we don't have an interface (or list of BSS proxies, which are also created async) and end-up making a geoip-only query to the server. Also we can easily end-up not making the wifi query at all if geoip-only query is still in process since we only allow one query at a time. Let's leave this to subclasses (we only have 1 right now anyway).
-
Zeeshan Ali Khan authored
wpa_supplicant seems to be a lot more ubiquitous than NetworkManager so lets use that instead of NetworkManager for getting WiFi data.
-
Zeeshan Ali Khan authored
-
Zeeshan Ali Khan authored
We only need access to read properties and receive signals. A patch has been submitted to wpa_supplicant to make this redundant in here but at least for now, we need to get access ourselves.
-
- Apr 29, 2014
-
-
Zeeshan Ali Khan authored
-
- Apr 17, 2014
-
-
Zeeshan Ali Khan authored
This was used for testing geoip client against our own and fedora's server but we dont use any of them anymore and those tests have been removed already so this data is very much redundant.
-
Zeeshan Ali Khan authored
-
Zeeshan Ali Khan authored
-
Zeeshan Ali Khan authored
-
Zeeshan Ali Khan authored
-
Zeeshan Ali Khan authored
-
- Apr 16, 2014
-
-
Zeeshan Ali Khan authored
We'll need this later when we'll override the config parameters with values passed from commandline.
-