Configuring pin as non-touch after being in touch mode leads to crash

[dpgeorge]

When calling pin.isTouched the pin automagically goes into touch mode and creates a MicroBitButton instance. It registers itself with the system ticker using addSystemComponent. Then if you call pin.getDigitalValue it deletes the MicroBitButton instance but does not seem to unregister itself with the system ticker. As such the system hangs.

Example code to reproduce:

void app_main() {
    while (1) {
        printf("digital mode\r\n");
        for (int i = 0; i < 20 * 10; ++i) {
            if (uBit.io.P0.getDigitalValue()) {
                uBit.display.print('D');
            } else {
                uBit.display.clear();
            }
            uBit.sleep(50);
        }
        printf("touch mode\r\n");
        for (int i = 0; i < 20 * 10; ++i) {
            if (uBit.io.P0.isTouched()) {
                uBit.display.print('T');
            } else {
                uBit.display.clear();
            }
            uBit.sleep(50);
        }
    }
}                                                                                                     

It works the first time round, but on the second loop it hangs.

[finneyj]

thanks @dpgeorge.

@jamesadevine - can you please fix this one up? we should deregister in the appropriate destructor...

[dpgeorge]

Destructors are evil: #72 :)

[finneyj]

:-)

Well, let's face it, C++ is evil... But when in leagues with the devil, his minions should be turned to do our bidding. :-)

[jamesadevine]

@finneyj @dpgeorge will be fixed momentarily...

@dpgeorge as an aside printf pulls in 1k of stuff, so uBit.serial.printf() should be used instead where possible.

(I assume you were aware though? In which case ignore me!)

[dpgeorge]

@dpgeorge as an aside printf pulls in 1k of stuff, so uBit.serial.printf() should be used instead where possible.

Ok, thanks for the tip! It was just for debugging purposes in this case. MicroPython has its own internal printf implementation that we use (but I wouldn't be surprised if there were some stray native printf's in there somewhere...).

[dpgeorge]

Thanks @jamesadevine!

[jamesadevine]

@dpgeorge I'm revising the patch a little bit - will update in a short while! :)