The 2017 FRC Driver Station supports getting the robot IP via a TCP
connection that returns JSON. Use this to support overriding the
server IP address used for client connections.
Default to using this approach for client connections in the NetworkTable
interfaces.
Add support for setting the server address without stopping and
restarting the client.
SetTeam now also round-robins by default.
During testing, I was seeing a lot of unnecessary code (and allocations
in Java/C#) when appending the path separator to the base path. That
technically is a constant, so this computes this constant during class
construction.
This will allow dependencies such as wpilibc to update to use wpiutil
without breaking "normal" ntcore static library use in the meantime.
This commit also restructures the gradle files by creating a new
(placeholder) wpiutil project, and moving the ntcore project into
a separate gradle file. Added toolchains/native.gradle (refactored from
ntcore).
Also fixes ntcore skipJava on Windows by providing an alternate .def file
for this case.
Moving these headers from src to include enables other libraries to use the
functionality provided.
* tcpsockets
* atomic_static
* raw_istream
* timestamp
* SafeThread
* Base64
* LEB128
* ConcurrentQueue
The classes have been moved into the wpi namespace as they're generic.
There was no way to atomically check for a key in the table,
and then setting if it if non existant. Back before persistent
this was not a problem, however now it is, as its possible for
values to be added before team's robot programs start. This makes
the old method of calling Put*** methods in RobotInit invalid.
This adds SetDefault methods, which do this atomically.
Since we now get ConnectionInfo when setting a connection listener,
there is some good information in there that teams could use.
Implemented using default interface methods, so teams should see no
change if they don't implement the Ex methods.
I noticed that the connection listener methods don't exist at all in
C++, so they did not get added there.
There's a lot of buzz going around the internet about people trying to
get ntcore working on other devices. One of the things that makes it
harder is having to have a separate jar for each platform. What this
change does is if the loading of the extracted library fails, it will
attempt to load ntcore from the path. This means that a program like
GRIP can just provide the libntcore.so and not have to worry about
compiling different versions for different platforms.
JavaGlobal was unconditionally attaching to (okay) and detaching from (bad)
the current thread during destruction. We don't want to do this if the
destructor gets called from an attached thread. Instead, use GetEnv to
first try to get the environment, and only attach and detach if it returns
an error saying the thread is detached.
Also, make sure notifier callbacks appropriately free Java locals to avoid
running out of local variable space.
During JVM shutdown, some JNI calls may not return, so it's not possible to
reliably perform a join() during static variable destruction (which occurs
as the JVM unloads the JNI module).
Also, due to static variable destruction, it's not safe to use any members
of a static class instance from a separate thread of execution.
SafeThread is a templated thread class and a related owner class that's
designed for safe operation and shutdown of threads in the presence of
callbacks that may not return. It also passes ownership of variables from
the static instance to the thread, so the thread can safely operate until
it exits (the last operation of the thread being to destroy its instance).
Notifiers, RpcServer, and Logger now use SafeThread to ensure race-free
destruction in both C++ and Java.
All Java callback threads are now marked as Java daemon threads so they
don't keep the JVM running after main() terminates.
All Java callback threads are now named so their purpose is more easily
identified in a debugger.
Add SetRpcServerOnStart and SetRpcServerOnExit (similar to Listener).
Each call to AttachCurrentThread results in a new Java thread object being
created. This is inefficient and also causes debugging issues with Eclipse
due to constant creation and removal of threads. Now AttachCurrentThread is
only called once for (all) listeners and once for logging (if used).
These are good to have for backwards compatibility, but discouraged for new
development (default-taking functions should be used instead). The reason
is that the exceptions must be explicitly handled and may initially work but
then fail at an inopportune moment.
Mark the similar Java functions as deprecated as well for the same reason.
Update all the docs for consistency.
Mark overridden functions as such in both C++ and Java.
Make IsPersistent and GetFlags const in C++.
no-unused-private-field for Mac builds. Gradle also now works
with the classifier-based dependency system, rather than having
separate repos for every level.
Change-Id: I2eb87391181e91b5675e3e982e4d915be83e14ea
current platform by default, and only builds tests when building for the
current platform. Mac builds and VS2015 builds are fixed.
The other big change in this update is the introduction of Debug and
Release builds. Debug builds are built with -O0 and -g. Release builds are
built with -O2 and -g. For GCC-based builds, the resulting shared object
is copied, stripped of debug information, and a debug link is set up to
the copied shared object. This allows the release build to clock in at
around 600 KB. On Windows, the debug info is already stored in a separate
PDB file, so this copy and strip is not necessary.
ntcore is being separated out from the rest of allwpilib. All other
builds will consume a published maven dependency from this project.
There are 4 possible publishing targets now: release, stable, beta, and
development. These are specified on the command line via -Prepo=<repo
name>.
Change-Id: Ie8cb21f910953e09b80a5192317033eb0866cb70
This enables listeners to be notified of not only value updates, but also flag
changes and deletions by using a bitmask to specify what notifications are
desired. The old API (which only provided a new/not new) flag is still
supported. This also subsumes the feature to listen to local changes (that's
one of the bitmask options).
The default behavior is to only notify remote changes, but for some
applications (e.g. GUI's) it's advantageous to know about local
changes as well.
This is (slightly) optimized in that local changes only result in
additional resources being consumed if (any) local listeners have been
created.
