Use infix evaluation

[esseivaju]

Use infix evaluation of logic expressions. We probably want to give a compile-time or run-time option to pick between infix and postfix notation.

[github-actions]

Test summary

  432 files    679 suites   31s ⏱️
1 354 tests 1 200 ✅ 40 💤 114 ❌
2 467 runs  2 240 ✅  8 💤 219 ❌

For more details on these failures, see this check.

Results for commit a8542c5.

♻️ This comment has been updated with latest results.

[esseivaju]

@sethrj @elliottbiondo Is this failing because we need to update all the orange test data which is in postfix notation, e.g. here

[sethrj]

Oh boy. Let's chat about this on slack? Namely if we're going to hardcode an option, whether and where to test it... And if we're going to break all the CSG representation strings, one thing I would like to do is flip the values of Sense so that "inside sphere 0" gives a CSG representation of "0" rather than "~0". 🤔

[esseivaju]

Yep, we can discuss this on Slack. I wanted to get numbers from the regression suite, so this is a quick, hardcoded implementation, but there are also logic strings in the regression suite.

[sethrj]

@esseivaju I think you can rebuild those manually by calling

$ export GDML=foo.gdml
$ build/test/orange/g4org_Converter '--gtest_filter=*arbitrary' --gtest_also_run_disabled_tests

and that'll spit out the necessary JSON files locally.

That might not be true for the custom TestEM3 🤔

[elliottbiondo]

If infix/postfix could be toggled as a runtime option that would side-step the issue with the orange test data, correct?

[sethrj]

@elliottbiondo yeah, but since it requires different implementations deep in SimpleUnitTracker that would mean propagating a lot of templates... I think we'd want this switch to be all or nothing.