Use QNetworkAccessManager for storing the files we receive instead of
directly KIO, then provide the KIO-based QNetworkAccessManager from the
daemon. This way we'll still get the KIO network-transparency, but the
library only depenends on QtNetwork, which will be easier to get in some
platforms.
After this change, we only depend on KConfig, KI18n and KCoreAddons, which
are tier1 and really easy to work with on any platform.
REVIEW: 123325
This fixes a bug where receiving broadcast messages would stop working
in case 2 messages have been received at the same time.
The problem is that readReady will only emit once per batch of datagrams
received, in case we receive 2 datagrams but only consume 1, readReady
will never be emitted again.
Moves the implementation into the actual daemon. This opens the
possibility for different ways to expose these notifications depending on
where the libkdeconnect will be deployed.
REVIEW: 123076
Uses KPluginLoader+KPluginMetadata instead.
Describes plugins in json instead of desktop files. These desktop files are
then embedded into the .so file. All the plugins will be in a kdeconnect/
directory, and those will be the ones to look for.
Note it doesn't drop the KService dependency as KIOCore is a KDE Connect
dependency as well.
REVIEW: 123042
If the phone thinks that it's already paired with the computer when
receiving a pairing request, it will answer stuff back that will cause
the desktop to abort the pairing process.
Centralizing the plugins' config will ensure that all the plugins store it
the same way (ie: not in random files scattered around, like until now).
The base KCM class, together with the already existing base plugin class,
will give easy access to all the plugins to this centralized config. Also,
now the settings are not shared across devices (that is: every device can
have different config for a same plugin).
Note: This commit requires KCMUtils 5.9
REVIEW: 122927
Removes the usage of RenameDialog by moving the functionality away. In
fact, what it does is make the share dialog try to come up with a different
name rather than asking the user. It's a quite common procedure as makes
the interaction simpler (for example, Chromium and Kamoso work like that).
REVIEW: 122813
Benefits:
- We had config files in different paths, now they will be centralized.
- Daemon, LanLinkProvider and NetworkPackage depend only on Qt now.
- KConfig is accessed with strings as keys, where a typo can go unnoticed.
- Daemon is now thinner, as it doesn't need to initialize config values.
- QCA::Initializer is in a single place now (was in Daemon and Plugin).
Only look for it in the writable path, as it won't ever be installed in a
common place.
Don't look for the path in the config file either.
Don't make it depend on the application name.
REVIEW: 122744
Use iconName instead of sending a fixed-size pixmap around, fix wording and use title
capitalization, more meaningful names in notifyrc, QStringLiteral and new style connect.
REVIEW: 122121
This meant to add a lot of dependencies to each plugin since we had
KDELibs4support as PUBLIC link meaning that anything linking against
kdeconnectcore was linking at the same time to mostly all frameworks.
Now each plugin has more or less its dependencies in the CMake some
still depend on KDELibs4Support.
For the mousepad plugin I needed to add a fixX11.h file that basically
undefines/defines again some stuff xlib has that conflcits with normal
C++ and Qt.
Before it was not conflicting because some lib within KDELibs4Support
was including this file, but now we have to do it ourselves.
I had to keep the old kDebug code in kdebugnamespace because other parts
of the code are using it. Will be removed as soon as all code using it
is ported as well.
Straight forward port from QJson to the native Json support.
As a note I have added 2 helper functions that make the code bit more
readable (object2qvariant and qvariatn2qobject).
Probably it was not detecting connections to a wired interface either,
this should do the trick but it's a bit weird having to connect to
QNetworkConfigurationManager::configurationAdded() to know about that.
Only lets the plugins that support something offered on the other end or
that offer something supported on the other end. If we don't have
information about the other device, we enable everything like we used to.
Also refactors a bit the plugin loader, to be able to provide information
step by step.
Currently it works, but we need support for capabilities at least on
Android to start taking good advantage of it.
REVIEW: 118820
CCMAIL: kdeconnect@kde.org
Trying to bind a UDP socket to 255.255.255.255 fails with EADDRNOTAVAIL
at least on FreeBSD.
The sender side on Android should send its message in broadcast mode (as
it already does), but the receiver side should bind to INADDR_ANY, not
INADDR_BROADCAST.
BUG: 324769
REVIEW: 118870
