We have some YRGBLED2 boards that have 5 character long serial numbers, for example YRGBLED2-C3E16. These boards can not be found using yFindColorLedCluster() or yFirstColorLedCluster() because an exception is thrown. We have lots of boards that look like YRGBLED2-121916 and they all work fine.

I just updated to the latest library release, latest VirtualHub and this is still happening, previously my library was from 2017.

Hi,

I'm having some problems with the Yocto3D. I slightly modified the main.cpp for the example code to print out x, y, and z values for the accelerometer and am getting incorrect output. This happens more or less randomly, and whether it works or not seems to be determined at the time the USB cable is connected. That is, if it is working, I can stop and restart the program and everything will continue to work, but if it is producing incorrect output, stopping and restarting the program won't help.

I've seen this on both a 64bit laptop and a 32bit ARM machine, both running Ubuntu 14.04.

This is the output I'm getting for the x, y, and z acceleration values:

teichman@silicon:~/Desktop/yocto$ Examples/Doc-GettingStarted-Yocto-3D/Binary_Linux/64bits/demo any
tilt1
tilt2
compass acc     gyro
 -23.984        -23.964 -23.587 
 -23.984        -23.964 -23.587 
 -23.984        -23.964 -23.587 
 -23.984        -23.964 -23.587 
 -23.984        -23.964 -23.587 
 -23.984        -23.964 -23.587 
 -23.984        -23.964 -23.587 
 -23.984        -23.964 -23.587 

It looks like it is not possible to do:

#include "yocto_display.h"
#include "yocto_dualpower.h"

In file included from .../main.cpp:3:
.../Sources/yocto_dualpower.h:60:3: error: typedef redefinition with different types ('enum Y_POWERSTATE_enum' vs 'enum Y_POWERSTATE_enum')
} Y_POWERSTATE_enum;
  ^
In file included from .../main.cpp:2:
.../Sources/yocto_display.h:60:3: note: previous definition is here
} Y_POWERSTATE_enum;
  ^

To really use the IMU in an accurate and precise way you need to have some way to guarantee that all the sensor readings are synchronized across the device, for example 3 axis accel and 3 axis gyro.

Currently I don't see a way to do this. Can the firmware in the controller make sure that it keeps synchronized results and the software API provide a way to read these results.

Why are you disallowing me to choose my installation prefix?

if(${CMAKE_SYSTEM_NAME} STREQUAL "Linux")
SET(CMAKE_INSTALL_PREFIX "/usr")
endif()

It would be really nice if there were standard install targets in the Makefile (or better yet, if you provided a CMakeLists) for yoctolib and yapi. Is there a reason this isn't the case?

In our use case, I'm building the yocto binaries in a Docker container and would really like to be able to run "make install" to copy:

• the headers to '/usr/local/include'
• executables to '/usr/local/bin'
• .so and .a files to '/usr/local/lib'
• other artifacts maybe to '/usr/local/share'

Obviously these targets could vary based on linux distrobution/OS, but the target directories should always be available as variables.

Having the files installed in the OS path as above would greatly simplify my CMakeLists, and be consistent with other packages manually compiled and installed in our Docker containers.

Extended testing revealed a handle leak in the 1/13/2017 version of Linux CPP source. I fixed it below...well, the handle leak is not occurring anymore anyway.

#define YOCTO_LOCK_PIPE "/tmp/.yoctolock"

// return 1 if we can reserve access to the device 0 if the device
// is already reserved
static int yReserveGlobalAccess(yContextSt *ctx, char *errmsg)
{
    int fd;
    int chk_val, mypid, usedpid = 0;
    size_t res;
    mode_t mode=0666;
    mode_t oldmode = umask(0000);
    char msg[YOCTO_ERRMSG_LEN];

    HALLOG("old mode (%#o)\n",oldmode);
    HALLOG("create fifo with (%#o)\n",mode);
    if(mkfifo(YOCTO_LOCK_PIPE,mode)<0) {
        HALLOG("unable to create lock fifo (%d:%s)\n",errno,strerror(errno));
    }
    umask(oldmode);
    fd = open(YOCTO_LOCK_PIPE,O_RDWR|O_NONBLOCK);
    if(fd<0){
        HALLOG("unable to open lock fifo (%d)\n",errno);
        if(errno==EACCES) {
            return YERRMSG(YAPI_DOUBLE_ACCES, "we do not have acces to lock fifo");
        }else{
            // we cannot open lock file so we cannot realy
            // check double instance so we asume that we are
            // alone
            return YAPI_SUCCESS;
        }
    }
    chk_val = 0;
    mypid = (int) getpid();
    res = read(fd, &chk_val, sizeof(chk_val));
    if (res == sizeof(chk_val)) {
        //there is allready someone
        usedpid = chk_val;
    } else{
        // nobody there -> store my PID
        chk_val = mypid;
    }
    res = write(fd, &chk_val, sizeof(chk_val));
    if(res != sizeof(chk_val)) {
        YSPRINTF(msg, YOCTO_ERRMSG_LEN, "Write to lock fifo failed (%d)", res);
		close(fd);
        return YERRMSG(YAPI_DOUBLE_ACCES, msg);
    }
    if (usedpid != 0) {
        if (usedpid == 1) {
			close(fd);
            // locked by api that not store the pid
            return YERRMSG(YAPI_DOUBLE_ACCES, "Another process is already using yAPI");
        } else {
            YSPRINTF(msg, YOCTO_ERRMSG_LEN, "Another process (pid %d) is already using yAPI", (u32) usedpid);
			close(fd);
            return YERRMSG(YAPI_DOUBLE_ACCES, msg);
        }
    }
	close(fd);
    return YAPI_SUCCESS;
}


size_t dropwarning;

static void yReleaseGlobalAccess(yContextSt *ctx)
{
    int chk_val;
    int fd = open(YOCTO_LOCK_PIPE,O_RDWR|O_NONBLOCK);
    if(fd>=0){
        dropwarning = read(fd,&chk_val,sizeof(chk_val));
		close(fd);
      }
}

Many classes in yocto_api.h have virtual methods, but no virtual destructor. If they are used
polymorphically then this could result in undefined behaviour. See https://github.com/isocpp/CppCoreGuidelines/blob/master/CppCoreGuidelines.md#Rc-dtor-virtual for discussion of the issue.

As an example, the YMeasure class does not have a virtual destructor.

I just looked in the C++ API website but I still can't find a way to let the compass to read faster, is there any way that I can adjust the speed of the compass sensor to read data faster. thanks a lot!

Hello,

Would it be possible to add a feature to the PWM Rx firmware to implement a counter which increments every time the pulse count rolls overs? This rollover counter could then be read out via an API call. Of course, the counter is reset to zero on soft/hard reset of the device. Thanks in advance.

Best Regards,

Rob Krakora

I am asking if we could select the fusion mode as described in the BNO055 document section 3.3.3. I don't want absolute orientation because I am operating in an area with a lots of metal and high power electrical circuits. With the magnetometer feeding into the estimation this will basically be useless to me. Maybe you can expose a way to select among the 5 fusion methods?

I am creating this ticket so I know when this feature has been implemented. I feel this is pretty fundamental to the BNO055 and other could benefit from this.

The method declaration of every method has no throws(foobar) specified.

Please add them.

I have a dozen of your Yocto-MaxiThermistor boards that are connected to computers that record the measurements at 1Hz continuously. After leaving these running for months and now analyzing the logs I find bad temperature measurements.

First I thought maybe this is a glitch in the software but when I look at the VirtualHub page it also shows the Max Temperature updating when then bad measurements occur, so it isn't my software.

Next I thought maybe it is the particular computer setup, so I connected one various boards to various other computers and after monitoring them for just 10 minutes I see the bad measurements.

So my next guess is that the firmware has a timing issue. The bad measurement seems to show up on any and all channels randomly. The attached image shows two channels that have already captured bogus measurements.

localhost_4445_webapp.pdf

unit_test-Multitech Debug [C/C++ Remote Application]
unit_test [5558] [cores: 0] (0 filtered running threads)
Thread #1 5558 [core: 0] (Suspended : Signal : SIGINT:Interrupt)
__lll_lock_wait() at lowlevellock.c:46 0xb6fc6a9c
__GI___pthread_mutex_lock() at pthread_mutex_lock.c:114 0xb6fc0ec8
Poco::MutexImpl::lockImpl() at Mutex_POSIX.h:61 0x122fc
Poco::Mutex::lock() at Mutex.h:209 0x124bc
Poco::ScopedLockPoco::Mutex::ScopedLock() at ScopedLock.h:40 0x130c4
PollingThread::Stop() at PollingThread.cpp:27 0x37448
DevicePort::Stop() at DevicePort.h:18 0x350e4
PulsePort::Stop() at PulsePort.cpp:175 0x43338
PulseDevice::Stop() at PulseDevice.cpp:100 0x39ee4
PulseTests_FunctionalTest_Test::TestBody() at PulseTests.cpp:461 0x1cd60
<...more frames...>

Thread #2 6272 [core: 0] (Suspended : Container)
__pthread_cond_wait() at pthread_cond_wait.c:187 0xb6fc383c
Poco::EventImpl::waitImpl() at Event_POSIX.cpp:99 0x6e328
wait() at Event.h:99 0x72f64
Poco::PooledThread::run() at ThreadPool.cpp:193 0x72f64
Poco::(anonymous namespace)::RunnableHolder::run at Thread.cpp:57 0x66354
Poco::ThreadImpl::runnableEntry() at Thread_POSIX.cpp:349 0x66488
start_thread() at pthread_create.c:314 0xb6fbe424
() at clone.S:92 0xb6af6878
() at clone.S:92 0xb6af6878

Thread #3 6273 [core: 0] (Suspended : Container)
__pthread_cond_wait() at pthread_cond_wait.c:187 0xb6fc383c
Poco::EventImpl::waitImpl() at Event_POSIX.cpp:99 0x6e328
wait() at Event.h:99 0x72f64
Poco::PooledThread::run() at ThreadPool.cpp:193 0x72f64
Poco::(anonymous namespace)::RunnableHolder::run at Thread.cpp:57 0x66354
Poco::ThreadImpl::runnableEntry() at Thread_POSIX.cpp:349 0x66488
start_thread() at pthread_create.c:314 0xb6fbe424
() at clone.S:92 0xb6af6878
() at clone.S:92 0xb6af6878

Thread #4 9997 [core: 0] (Suspended : Container)
poll() at syscall-template.S:81 0xb6ae85d0
handle_events() at io.c:1,888 0xb6d15e38
libusb_handle_events_timeout_completed() at io.c:2,024 0xb6d167e8
event_thread() at ypkt_lin.c:254 0xf10b8
start_thread() at pthread_create.c:314 0xb6fbe424
() at clone.S:92 0xb6af6878
() at clone.S:92 0xb6af6878

Thread #5 9998 [core: 0] (Suspended : Container)
__pthread_cond_destroy() at pthread_cond_destroy.c:75 0xb6fc343c
yCloseEvent() at ythread.c:223 0x905e8
yPktQueueFree() at ystream.c:935 0xd60a4
yyyPacketShutdown() at ypkt_lin.c:665 0xf2204
yyPacketSetup() at ystream.c:1,367 0xd719c
yPacketSetup() at ystream.c:1,425 0xd742c
yStreamSetup() at ystream.c:1,538 0xd792c
StartDevice() at ystream.c:2,091 0xd92c8
enuUpdateDStatus() at ystream.c:2,175 0xd9640
yUSBUpdateDeviceList() at ystream.c:2,320 0xda004
yapiUpdateDeviceList_internal() at yapi.c:2,597 0xcf3e8
yapiUpdateDeviceList() at yapi.c:4,385 0xd3d48
YapiWrapper::updateDeviceList() at yocto_api.cpp:3,739 0x9fc20
YDevice::requestAPI() at yocto_api.cpp:2,536 0x9c470
YFunction::isOnline() at yocto_api.cpp:2,239 0x9ab80
PulseDevice::readValue() at PulseDevice.cpp:498 0x3dc78
PulseDevice::UpdateItemValues() at PulseDevice.cpp:192 0x3b39c
PulseDevice::Poll() at PulseDevice.cpp:161 0x3a910
PulsePort::Poll() at PulsePort.cpp:232 0x43d64
PollingThread::Thread() at PollingThread.cpp:68 0x3794c
PollingThread::StaticPollingThread() at PollingThread.cpp:51 0x377e8
Poco::(anonymous namespace)::CallableHolder::run at Thread.cpp:80 0x66368
Poco::ThreadImpl::runnableEntry() at Thread_POSIX.cpp:349 0x66488
start_thread() at pthread_create.c:314 0xb6fbe424
() at clone.S:92 0xb6af6878
() at clone.S:92 0xb6af6878

Thread #6 9999 [core: 0] (Suspended : Container)
__lll_lock_wait() at lowlevellock.c:46 0xb6fc6a9c
__GI___pthread_mutex_lock() at pthread_mutex_lock.c:79 0xb6fc0d20
yEnterCriticalSection() at ythread.c:601 0x90aac
devPauseIO() at ystream.c:496 0xd5750
yUsbReadBlock() at ystream.c:2,832 0xdb358
yapiRequestWaitEndUSB() at yapi.c:3,238 0xd1524
yapiHTTPRequestSyncStartEx_internal() at yapi.c:3,326 0xd18b8
yapiHTTPRequestSyncStartOutOfBand() at yapi.c:4,549 0xd4298
YDevice::HTTPRequest() at yocto_api.cpp:2,501 0x9c140
YDevice::requestAPI() at yocto_api.cpp:2,533 0x9c450
YFunction::isOnline() at yocto_api.cpp:2,239 0x9ab80
PulseDevice::readValue() at PulseDevice.cpp:498 0x3dc78
PulseDevice::UpdateItemValues() at PulseDevice.cpp:192 0x3b39c
PulseDevice::Poll() at PulseDevice.cpp:161 0x3a910
PulsePort::Poll() at PulsePort.cpp:232 0x43d64
PollingThread::Thread() at PollingThread.cpp:68 0x3794c
PollingThread::StaticPollingThread() at PollingThread.cpp:51 0x377e8
Poco::(anonymous namespace)::CallableHolder::run at Thread.cpp:80 0x66368
Poco::ThreadImpl::runnableEntry() at Thread_POSIX.cpp:349 0x66488
start_thread() at pthread_create.c:314 0xb6fbe424
() at clone.S:92 0xb6af6878
() at clone.S:92 0xb6af6878
Remote Shell
/opt/mlinux/3.2.3/sysroots/x86_64-mlinux-linux/usr/bin/arm-mlinux-linux-gnueabi/arm-mlinux-linux-gnueabi-gdb (7.6.2)

On Unix you do not have to include the "dirty active wait loop" before exiting if the argument of the program is wrong. The error will just be displayed in the console and the user will get the prompt faster.

Maybe the waiting loop is needed on Windows. But on Unix it is just annoying.

Please use something like:

static void usage(void)
{
    cout << "usage: demo <serial_number> " << endl;
    cout << "       demo <logical_name>" << endl;
    cout << "       demo any                 (use any discovered device)" << end+++l;
#ifdef WIN32
    u64 now = yGetTickCount();    // dirty active wait loop
    while (yGetTickCount()-now<3000);
#endif
    exit(1);
}

In yoctolib_cpp/Examples/Prog-QT/QTDemo/inventory.cpp, you're including an header called ui_inventory.hbut I can't find it... Am i missing something?

The following code runs without error:

int main(int argc, const char * argv[])
{
  string         errmsg;
  string         target;
  YDisplay      *disp;

  // Setup the API to use local USB devices
  if (yRegisterHub("usb", errmsg) != YAPI_SUCCESS) {
    cerr << "RegisterHub error: " << errmsg << endl;
    return 1;
  }

  target = "pouet"; // or anything else
  disp =  yFindDisplay(target + ".display");   

  if (!disp->isOnline()) {
    cout << "Module is offline (check USB cable)" << endl;
    return 1;
  }
}

The documentation (French or English) describes a method display→get_displayId() but the real method name looks like to be get_functionId() instead.

There is something I havn't understood.

Every GNUmakefile use the regular g++ to compile arm version of the demos.

Since the lib GNUmakefile is ok, I don't understand what have been done.
I've edited a lot of the GNUmakefile to make the buzzer demo to compile for armel, it was a lot of pain.

YPwmInput::set_pulseCounter(10000) does not work in version of released on 1/13/2017.

"GET /api/pwmInput2/pulseCounter?pulseCounter=100000&. \r\n\r\n"

I think that storing in github binary files that are the result of a compilation is wrong:

• it is big (147 MB or 70% of the total size) and is then very slow to fetch from the server
• I always get git conflicts with the versions I rebuild locally

I fully understand that you want to provide pre-built binaries to your customers. But if I use github it is because I want to rebuild from the source code.

I can always use the official .zip archives if I want to use pre-built binaries.

I have a Raspberry Pi rev 1 model B using Raspbian

$ LANG=C ./build.sh 
Build static and dynamic library for C++
========================================
compiling linux/armel/objs/yapi.o
In file included from /usr/include/features.h:398:0,
                 from /usr/include/stdint.h:25,
                 from /usr/lib/gcc/arm-linux-gnueabihf/4.9/include/stdint.h:9,
                 from ../Sources/yapi/ydef.h:191,
                 from ../Sources/yapi/yapi.h:47,
                 from ../Sources/yapi/yapi.c:40:
/usr/include/arm-linux-gnueabihf/gnu/stubs.h:7:29: fatal error: gnu/stubs-soft.h: No such file or directory
 # include <gnu/stubs-soft.h>
                             ^
compilation terminated.
GNUmakefile:273: recipe for target 'linux/armel/objs/yapi.o' failed
make: *** [linux/armel/objs/yapi.o] Error 1