New PATCH release 07.00.08

Changes since 07.00.07:

Enable computation of amplitude and phase for the total controller output DAQ channel. This is a required input for the diagnostic server.
Should fix #5286

New PATCH release 07.00.07

Changes since 07.00.06:

- fixed some tests
- changes to build environment

New PATCH release 07.00.06

Changes since 07.00.05:

- do not tirgger full table generation when updating the FFC tables (LFF causes high cpu load otherwise)
- PhaseSetter: fix moving output rotation into the wrong direction; improved handling

New PATCH release 07.00.05

Changes since 07.00.04:

- OVC performance optimisation: cache result of forwardVectorSum.useFromSlave()

New PATCH release 07.00.04

Changes since 07.00.03:

added SynchroniseTiming module

New PATCH release 07.00.03

Changes since 07.00.02:

Some bug fixes for XFEL:
- fixed computation of vectorsum error for FSM
- added a small module which switches off the RF at server start
- enable amplitude/phase computation for partial vector sum
- if scaling is set to 0 initially, change it to 1.

New PATCH release 07.00.02

Changes since 07.00.01:

Patch release just to rebuild the Debian package (after a fix in postinst)

New PATCH release 07.00.01

Changes since 07.00.00:

do not install dummy configuration when installing to the DOOCS directory, as this makes problems when installing from the Debian package

New MAJOR release 07.00.00

Changes since 06.10.00:

*** INCOMPATIBLE CHANGES WITH PREVIOUS llrfctrl v6 RELEASES (list may be incomplete) ***
- Changed scheme how the power is computed for the AdcChannel.
- The variable Automation/Rampup/OVC/erorrTarget is now in digits (was before in G.U.)
- /Configuration/globalSamplingPosition no longer exists, use /Configuration/SamplingPosition/offset
- /Controller/errorMessage now needs to be reset by operator (see /Controller/resetErrorMessage)
- only compute amplitude and phase for the controller channels vector sum and error
- Rename the project, the application class and the configuration files into a more consistent naming scheme. The project is called "llrfctrl" and everything is now named accordingly. This completely breaks the compatibility with configuration files of older versions.
- renamed variables in control system:
  "/Board0/CommonCalibration/scale" and ".../angle" -> "/Configuration/CommonCalibration/scale" and ".../angle"
  These act now on all ADC channels (of all boards) where Calibration/CommonCalibration/enable has been set
- removed internal channel name mapping - this should now be done in the control system adapter mapping

New features:

- Implemented support for multi-cavity systems throughout the server (incl. master-slave systems e.g.f for EuXFEL)
- Added support for multiple flattops
- Individual ADC channels are now scaled with /Configuration/lengthScale and the unit has been changed to /Configuration/gradientUnitPerLength. This allows to display the signals in e.g. MV/m, but the old behaviour is still possible.
- Added special UserVariables module needed for EuXFEL/FLASH
- Added support for vector sum bit shift required for multi-cavity systems
- Added support for beam loading compensation
- Added PhaseSetter module (for EuXFEL)
- Learning feed forward: rotate current state when controller setpoint phase is changed (like in llrfctrlsrf recently implemented)
- Publish the global sampling position not only as an index but also in microseconds for each flattop
- Publish server version also as a string
- added amplitude and phase slope, and transition time offset
- introduced trigger limiters (for DCMs)
- Add AmplitudePhaseError module which computes amplitude and phase error and a fault flag (e.g. user for the XFEL FSM)
- made adc channel statistics optional
- make separate FF correction enable variables for user FFC and LFF FFC
- make use of bidirectional variables to put the result of the VM offset calibration into the corresponding control system variables
- added absolute limit for OVC

Performance improvements:

- Allow switching off controller channel statistics
- Allow switching individual adc boards to RAW mode instead of all at the same time
- Do not always read and demultiplex raw adc data

Other changes:

- Error messages are no longer printed to console unless compiled in debug mode
- Improve handling when server starts with out-of-range initial values. This was presumably also causing race-conditions in the tests.
- some minor bug fixes and code cleanup
- prevent table generation if not needed
- link against DOOCS backend instead of symlink gymnastics
- generate tables immediately after parameter change, not only at next trigger
- OVC: prevent averaging window indices to go out of range
- added tests for OVC
- added test for VM offset calibration
- fixed interlock latcher reset happening too quick to be testable (and maybe too quick for the firmware to detect it)
- extended test for interlock detection
- rework test for Automation module
- refactoring to make the code simpler
- do not use slashes in module names and accessor names (now forbidden in ApplicationCore)
- move most directly connected device variables out of the application

New MINOR release 06.10.00

Changes since 06.09.00:

- added support for switching between raw and amplitude/phase mode for controller_slave firmware
- fixes for latest changes in DeviceAccess and VirtualLab

New MINOR release 06.09.00

Changes since 06.08.00:

- Implemented the LearningFeedForward algorithm. The implementation is almost identical to the one from the llrfctrlsrf server. Only known difference:
  - maxControlError now limits the error, not the error squared. This is more logical but incompatible to the llrfctrlsrf server!

New MINOR release 06.08.00

Changes since 06.07.00:

- make data decimation for MicroDAQ configurable
- use units for amplitudes and power from configuration file

*** Control system interface changes:

- renamed FeedBack/Gain/amplitude into FeedBack/Gain/magnitude (in both /Controller and /Automation)
- removed FeedBack/Gain/phase (I and Q gains are now always equal)

New MINOR release 06.07.00

Changes since 06.06.01:

- update for changes in ChimeraTK: use new ReadAnyGroup
- added DataLossCounter

New PATCH release 06.06.01

Changes since 06.06.00:

- for non-DOOCS installations, install the symlink to the right DOOCS backend as /prefix/lib/libChimeraTK-DeviceAccess-DoocsBackend.so.llrfserver
- set WORD_DAQ_SAMPLES register for both the ADC and the controller DAQ

New MINOR release 06.06.00

Changes since 06.05.01:

- write WORD_DAQ_SAMPLES by server (constant from base configuration)
- updated for latest firmware changes
- Changed DOOCS mapping used for the dummy installation to use a single location. This matches our new panels.
- fixed MIMO (and Smith) parameters not properly behaving when leaving un-applied values in the input mask over a server restart

New PATCH release 06.05.01

Changes since 06.05.00:

- added Controller/FeedBack/smithEnable also to the control system and micro DAQ

New MINOR release 06.05.00

Changes since 06.04.00:

- added support for Smith predictor
- install xml generator also for non-doocs setups

New MINOR release 06.04.00

Changes since 06.03.00:

- added output vector correction (implementation as in the llrfctrlsrf server)
- removed some debug output
- do not disable/reset the interlock latcher when going to error state

New MINOR release 06.03.00

Changes since 06.02.08:

- Fixes for multi-board setups

New PATCH release 06.02.08

Changes since 06.02.07:

- fixed dependency name for EPICS adapter
- corrected some engineering units