The MRAA clang build was getting overridden with gcc.
This commit allows MRAA to build with clang and adds nodejs
0.10 to the build matrix. Note, this means 8 builds
(2 compilers x 4 versions of nodejs).
* Added compiler directive for clang/gcc.
* Added build for node010.
* Added make install to cover the MRAA install.
Signed-off-by: Noel Eck <noel.eck@intel.com>
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
Partly rollsback f2c4d95c15 which stopped
mraa_i2c_init from switching i2c bus silently. If you don't want this to
happen, you need to use mraa_i2c_init_raw. Closes#497.
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
This fixes previous broken commit because dev is not defined before
_internal_init but since fp is taken during internal init we need to know the
channel number. This does not affect the internal override API.
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
Using NPM, this module can be cross-compiled for Android using NDK.
But, when compiling for Android, we need to add glob.c to source files
(as in Android.mk).
Also, re-arranged the layout in binding.gyp for better readability.
Signed-off-by: Robert Chiras <robert.chiras@intel.com>
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
This commit changes also the _replace function and adds exceptions to the C++
API for errors in AIO read
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
Enum was deprecated in 0.10.x and causes issues with strict C89 compliance so
is now removed
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
Edited syslog messages. Changed some return types to used signed and using -1
to indicate errors
Signed-off-by: Martin G Lane-Smith <mlanesmith@gmail.com>
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
This commit introduces support for Dallas Semiconductor (DS) 1-wire
compliant device support using an available UART device.
The principle of operation is described in the following Application
note by Maxim Electronics:
https://www.maximintegrated.com/en/app-notes/index.mvp/id/214
with help (1-wire search) from:
https://www.maximintegrated.com/en/app-notes/index.mvp/id/187
It has been tested on Galileo 2 and Edison, with 2 DS 1-wire devices,
the DS18B20 and DS2413 connected to the bus. A UPM driver for the
DS2413 is already complete and a PR will be submitted after this one.
It is important that you use a UART with CMOS/TTL level voltages
(3.3v/5v) RX and TX lines. DO NOT use standard RS232 level voltages
or you are going to have a bad day.
In order for this to work, a simple interface circuit, using a single
diode must be constructed:
(forgive my "Asciihematic" :)
-|
U|
A| TX---|<--+
R| |
T| RX-------o--------o 1-wire data bus
-|
The diode on TX is a 1N4148 (cheap and common), with the cathode
connected to TX, and the anode connected to RX and the 1-wire data
line.
The 1-wire data line requires a pull-up resistor, as the DS 1-wire
spec requires. 4.7-5K is typical for DS 1-wire buses.
NOTE: DHT-type (temp/humidity sensor) 1-wire devices ARE NOT DS 1-wire
compliant, and will not work with this code/circuit unfortunately.
Also note, this will use up one of your UARTs, which cannot be used
for any other purpose (ie: to access true UART-type serial devices).
You can however, connect as many DS 1-wire devices as feasible to this
UART, as it will function as a DS 1-wire bus master.
Signed-off-by: Jon Trulson <jtrulson@ics.com>
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
LIB_SUFFIX is no longer used. Replaced with LIB_INSTALL_DIR
which provides the correct lib dir with arch suffix.
Moved the call to include(GNUInstallDirs) prior to checking the
CMAKE_INSTALL_LIBDIR - which is set by include(GNUInstallDirs).
Tested with cmake 2.8 and cmake 3.5.
Signed-off-by: Noel Eck <noel.eck@intel.com>
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
We rely on driver-supplied information in mraa_spi_frequency(),
so let's be consistent.
Closes#132.
Signed-off-by: Alex Tereschenko <alext.mkrs@gmail.com>
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
ioctl() returns -1 on error, or otherwise a value certainly not within the
mraa_result_t enum. Fixes#480.
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
Whilst not the cleanest and the headers could do with an update, this is a
short term solution to stop the warnings
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
Currently, node modules install under the current environment's
node install path. This happens even if cmake is passed a
CMAKE_INSTALL_PREFIX.
This change checks if a CMAKE_INSTALL_PREFIX has been provided,
in which case it prepens the install path to the NODE_MODULE_INSTALL_PATH.
If a CMAKE_INSTALL_PREFIX has NOT been provided, the the mraa node
modules get install according to the NODE_ROOT_DIR.
This fixes a problem with an install case where the caller does not
have write access to the NODE_ROOT_DIR.
Signed-off-by: Noel Eck <noel.eck@intel.com>
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
mraa_LIB_PLAT_SRCS_NOAUTO no longer used for firmata unlike USBPLAT but
mraa_LIB_SRCS_NOAUTO and this is now loaded before the standard src files for
mraa. The package still needs to be generated with IMRAA & FIRMATA enabled for
this to work.
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
