Free memory allocated for previous system interrupt handler

[sergeuz]

Fixes #927

---

Doneness:

• Contributor has signed CLA
• Problem and Solution clearly stated
• Code peer reviewed
• API tests compiled
• Run unit/integration/application tests on device
• Add documentation
• Add to CHANGELOG.md after merging (add links to docs and issues)

[m-mcgowan]

how about adding some tests for this? registering thousands of successive handlers to the same interrupt should be sufficient to show that there is no memory leak and that deleting the old handlers works without causing a fault. We also have System.freeMemory() which could be used to validate that memory isn't being excessively leaked.

[sergeuz]

Added slightly less involved test case that I believe has similar coverage

[m-mcgowan]

• Tested on the Photon.
• Tested on the Core

Made the tests compile on the core by commenting out GPIO, PWM and Servo tests, which are unrelated. I think we will split the no_fixture tests into Peripheral tests and System tests on the core via a conditional define. so we can continue to compile and run the tests on the core.