Beckhoff: switch to using AdsDriver

[rerpha]

As a developer I would like to try and use https://github.com/Cosylab/adsDriver as a replacement for tcioc as this means we can avoid having to use a TPY file/fixed PLC memory addresses and instead use some sort of mapping file to just serve the PLC vars we need.

The TPY file has given us grief a few times, i spent ages (more than a whole week) trying to figure out why it wasn't working when working on #6097 so decided to give AdsDriver a try. It has proved to work fairly well against the PLC in r106 as well as a local PLC after a fix which has been submitted upstream. Overall AdsDriver is much simpler as it is only really doing 1 thing (serving ADS vars) rather than 3 (parsing a TPY, mapping these to epics types, then serving them ads read/write)

Acceptance criteria

• both above repos are modified (and pushed upstream if possible?) to compile on windows
• AdsDriver is tested against a real PLC and a simulated one

[rerpha]

I have been a bit cheeky and done this - it seems to work well as a drop in replacement for tcioc and gets round the annoying TPY changes which are taking up lots of time (on SANDALS for example)

Freddie managed to get it building on windows, and working against the system-installed twincat/ads library (rather than the github one) it works very well to provide us with two-way ADS comms.

There are a few quirks which are based on races between ADS, which updates very quickly (every 1ms!) and our epics motor record which updates every 200ms, so I have asked the maintainers if it'd possible to add an ioc shell command or modify their AdsOpen() method to make the poll rate configurable. some of the numbers below dont move due to a deadband as well i think.

Here you can see it in action, with axis 11 being "theta" of a reflectometry virtual instrument:

[rerpha]

PRs for this- i propose that we switch from tcioc to adsdriver as it seems to work well and does not rely on the TPY file. Below should be a drop-in replacement for using tcioc, instead of using a tpy file we now just have a template file for each axis.

ioc: ISISComputingGroup/EPICS-ioc#760
beckhofftestrunner: https://github.com/ISISComputingGroup/BeckhoffTestRunner/pull/16/files
release notes (this closes multiple issues)
#7509
#7513
#7534