Some fixes for onnxruntime in TRD, but disabling it since still broken and untested

[davidrohr]

TRD tries to use onnxruntime it seems, but this code is buggy since it checks for ONNXRuntime_FOUND which is never said, so this code was never used. Also the CMake code was wrong lacking the PRIVATE keyword, which was not seen, since it was never used.
This fixes it (requires #12963) and enables the TRD ML.cxx compilation if onnruntime is present.

@martenole : Could you check if this code is actually sane, since it was never used before? The alternative would be to just drop it, but I would not compile something conditionally based on a broken condition.

[github-actions]

REQUEST FOR PRODUCTION RELEASES:
To request your PR to be included in production software, please add the corresponding labels called "async-" to your PR. Add the labels directly (if you have the permissions) or add a comment of the form (note that labels are separated by a ",")

+async-label <label1>, <label2>, !<label3> ...

This will add <label1> and <label2> and removes <label3>.

The following labels are available
async-2023-pbpb-apass
async-2023-pp-apass1
async-2022-pp-apass6
async-2022-pp-apass4
async-mc
async-data

[alibuild]

Error while checking build/O2/fullCI for f3b6288 at 2024-04-01 11:38:

## sw/BUILD/O2-latest/log
c++: error: unrecognized command-line option '--rtlib=compiler-rt'
c++: error: unrecognized command-line option '--rtlib=compiler-rt'
/sw/SOURCES/O2/12964-slc8_x86-64/0/Detectors/TRD/pid/src/ML.cxx:129:14: error: 'trdTRD' was not declared in this scope; did you mean 'trkTRD'?
/sw/SOURCES/O2/12964-slc8_x86-64/0/Detectors/TRD/pid/src/ML.cxx:140:27: error: 'input' was not declared in this scope; did you mean 'int'?
/sw/SOURCES/O2/12964-slc8_x86-64/0/Detectors/TRD/pid/src/ML.cxx:143:115: error: 'trkIn' was not declared in this scope
ninja: build stopped: subcommand failed.

Full log here.

[martenole]

@f3sch could you comment on @davidrohr 's question?

[f3sch]

Hi @davidrohr and @martenole, the changes to the cmake were made in #10672. Since this code is really not used and but can be revived anytime (VC tracked), I would be fine do drop it completely.

[davidrohr]

Thx, I have disabled it now with if (0), but I am leaving the code in so people know how to reenable it in the correct way (just because it was wrong before...)

[martenole]

Ah and now the CI is unfortunately reset. There was also at least one typo in ML.cxx due to which the compilation failed with ONNX enabled. But OK, fine to leave it out for now and once someone would want to work on that part again it cannot directly be enabled, but also some other fixes will need to be done.