The concept does not work because the hardware needs half a second before the value shows.

It takes about half a second, so we do 10 triggers.

The hw protection on the hw takes some time so the next trigger does not have the cleared hw protection flag yet.

Each channel state machine has one output, and you can only have one feeder per network.

- use clang-format-14
- sort included headers
- break comments at 120 characters

Enable by adding
  include(cmake/enable_code_style_check.cmake)
to the project's CMakeLists.txt

The logical name mapping used for testing is now much closer to the real config.

The changes introduced in c2b47a83 are only working for the PZ16M.
The PZT4 is not using a bit field to enable the feedforward table, but
integer because it has different modes. It never was tested, and
tests are failing since then.

* renamed /Channelx/enableControlTable to /Channelx/controlTableMode
* changed its type to integer

The PZ16M xlmap needs a type hint modifier in addition to the renaming.

Dependency to ApplicationCore was missing

Ignore eclipse python project file.

- All variables which are conceptually bool now use ctk::Boolean
- Part of the changes were necessary because ApplicationCore now treats
  1-bit variables as ctk::Boolean, and the types in the server have to be
  consistent
- Causes history incompatibilities (at least to for previous piezoctrl4)

Consolidate coverage options

 - Remove Makefile.coverage.in
 - Drop BUILD_COVERAGE_ON_DEBUG for ENABLE_COVERAGE_REPORT (defaults to
    on if included)
 - Add clean-gcda target

simplified state machine implementation for hwProtectionMonitor, which also solves the bug that a persistent monitor error was not causing a second transition to error when reset was pressed.