During shared library loading, a different libLLVM can be pulled in, causing
llvm symbols from dependent libraries to resolve to that library instead of
this one. This has been seen in the wild with the Mesa OpenGL implementation
in JavaFX applications (see wpilibsuite/shuffleboard#361).
This is clearly a very breaking change. For some level of backwards
compatibility, a namespace alias from llvm to wpi is performed in the "llvm"
headers. Unfortunately, forward declarations of llvm classes will still break,
but compilers seem to generate clear error messages in those cases
("namespace alias 'llvm' not allowed here, assuming 'wpi'").
This change also moves all the wpiutil headers to a single "wpi" subdirectory
from the previously split "llvm", "support", "tcpsockets", and "udpsockets".
Shim headers will be added for backwards compatibility in a later commit.
In roborio GCC 5.5, __FRC_ROBORIO__ has been added directly to the
compiler. So we can instead use that to detect if we are build for the
roborio, and in a much more reliable way.
Also adds function that can register all the callbacks at once.
Since all of the callbacks issue a string identifier, it makes it
possible and easy to have one function callback, and differentiate the
path to take based on the string. Hooking up all the callbacks at once
makes it easier for the simulator developer to know when something was
added to wpilib rather than looking at the commits.
This race was caused by holding a lock while calling into FRC_ChipObject,
which was waiting for the callback to exit before returning, and our
callback wanted to grab the same lock.
Added callback scheme for a pass through to something higher
level. Since the ID is embedded into the arbitration ID, and some
devices can use different schemes whether it is plugged in through a
device or put into the daisy chain (pigeonImu), I made one "internal
data object" for max reusability.
This avoids a direct byte buffer allocation on every read/write/transaction
on the byte[] variants.
Changes HAL I2C interfaces to use const for dataToSend.
This avoids a direct byte buffer allocation on every read/write/transaction
for the byte[] variants.
Also change spiGetAccumulatorOutput() to directly set the AccumulatorResult
object, avoiding a ByteBuffer allocation.
Changes HAL SPI interfaces to use const for dataToSend.
Fixes#733.
