Fix test runner warning caused by previous commit, and rename calculateDuration() to calculateElapsed() to avoid confusing/overloading similar terms.

[juj]

I removed the odd stub addDuration() function in the previous PR: https://github.com/emscripten-core/emscripten/pull/24396/files#diff-fd5a5983b044aa841287bf197f2b5e46781162bf0e6a0ca8d1ccaddf1cc1c969L110-L111

turns out that Python unit test runner starts to warn if this function is not present.

test_sse4 (test_core.strict_js.test_sse4) ... ok (1.14s)
C:\python\Lib\unittest\case.py:580: RuntimeWarning: TestResult has no addDuration method
  warnings.warn("TestResult has no addDuration method",
test_sse4 (test_core.bigint.test_sse4) ... ok (1.10s)
C:\python\Lib\unittest\case.py:580: RuntimeWarning: TestResult has no addDuration method
  warnings.warn("TestResult has no addDuration method",
test_sse4 (test_core.instance.test_sse4) ... ok (1.13s)
C:\python\Lib\unittest\case.py:580: RuntimeWarning: TestResult has no addDuration method
  warnings.warn("TestResult has no addDuration method",
test_sse4 (test_core.core_2gb.test_sse4) ... ok (1.25s)

Add it back in, and use it to register the elapsed time in the test (looks like Python internally tracks it).

Rename the calculateDuration() function to a calculateElapsed() method so it doesn't look too similar to the addDuration() one.

[sbc100]

What is this different to reporting self.test_duration ?

[juj]

It is an processing order thing with Python's test harness. self.test_duration has not been calculated yet at this point, i.e. BufferedParallelTestResult.addDuration() has not yet been called by the time this function is called, so self.test_duration would still report zero.