Support more types with `ASSERT_*` macros

The `ASSERT_*` macros help to create more readable output on assertion failures. However, they work by converting both operands to `long`. This results in lost precision and confusing printed messages, such as:

```
Assertion failed @ :
expression:
tpsRate.get() >= 0
expands to:
0 >= 0
```

The following 3 improvements could help our use of these macros:
- Avoid losing precision when printing assertion failure details
- Support better floating point arithmetic in assertions (e.g. [googletest assertion macros](https://github.com/google/googletest/blob/main/docs/reference/assertions.md#floating-point-comparison-floating-point))
- Support `ASSERT_EQ` and `ASSERT_NE` for string comparisons.

Start by locating the ASSERT_* macro definitions and their existing assertion tests; inspect how operands are converted before failure details are printed. Compare the requested precision, floating-point, and string-comparison behaviors with the current implementation, then define tests covering each supported case as the completion criteria.

Written by the indexing model from the issue text.