Adds a test that verifies that Intl.supportedValuesOf('timeZones')
includes those timezones those don't correspond to any continent. This
includes the oceanic Etc/* timezones and UTC.
Fixes: https://github.com/tc39/ecma402/issues/778
"islamic" is underspecified because it doesn't include any information
which location and algorithms should be used to approximate new moon
observations. Instead switch to "islamic-tbla".
For each character class escape (\d, \D, \s, \S, \w, \W), check
positive cases (the escape matches all characters it's supposed to
match) and negative cases (the escape doesn't match any of the
characters it should not match). Each of these checks is also done in
Unicode mode and with the v flag.
This uses regenerate.js from the unicode-property-escapes-tests
repo to generate strings that contain exactly the characters that
are supposed to be matched or not matched for each escape.
Comparison is done with regex test instead of regex replace to
optimize the tests.
This is part of my work at the SYSTEMF lab at EPFL.
Avoid modifying the regenerate library object prototype.
The test for https://github.com/tc39/ecma402/pull/724 (added in
https://github.com/tc39/test262/pull/4328) didn't take the Time Zone
Canonicalization proposal into account; but it should, because that
proposal is stage 3.
As of that proposal, the [[TimeZone]] slot of DateTimeFormat gets the
case-regularized original identifier, no longer the primary identifier. So
the resolvedOptions().timeZone property also no longer returns the primary
identifier.
The objects it provides are also available in another harness file,
wellKnownIntrinsicObjects.js. There's no point in duplicating that in the
harness. Rewrite each test that used hidden-constructors.js to use
getWellKnownIntrinsicObject instead.
