Adapts or removes tests that relied on creating durations that are now out
of range. Adds new tests for maximum in-range and minimum out-of-range
durations.
A few results change because the algorithm previously used for rounding
didn't always add duration units to dates in RFC 5545 order, and we also
introduce a special case for rounding with largestUnit years or months and
smallestUnit weeks.
In order to fixtc39/proposal-temporal#2563, we added invocations of
BalanceDurationRelative after some invocations of RoundDuration. These
cause observable calendar calls, which must be accounted for in some
existing tests.
See https://github.com/tc39/proposal-temporal/issues/2563
The old behaviour was encoded in one test in staging, but the behaviour of
largestUnit in duration rounding has changed since that test was written.
Therefore I'm assuming that toString() should've been updated when that
happened.
In PlainYearMonth arithmetic, we need to find the end of the month when
adding a negative duration or subtracting a positive one. This end of the
month can be out of range.
Test case based on one provided by Anba. See issue:
https://github.com/tc39/proposal-temporal/issues/2700
This removes several loopholes where it was possible to return particular
values from user calls that would cause infinite loops, or calculate
zero-length days.
Note the monkeypatch of getPossibleInstantsFor in test/built-ins/Temporal/
TimeZone/prototype/getInstantFor/argument-builtin-calendar-no-array-
iteration.js.
Other than that, all the tests are basically identical.
This shortcut path now exists in all round(), since(), and until()
operations.
In Instant, PlainDate, PlainDateTime, and PlainTime, the change isn't
observable, so no tests could be added. This adds test coverage for
- Duration.p.round()
- PlainYearMonth.p.since()
- PlainYearMonth.p.until()
- ZonedDateTime.p.round()
- ZonedDateTime.p.since()
- ZonedDateTime.p.until()
As well as a few cases where we are testing that certain calendar methods
get called during a round operation, but previously were doing so with
options that now become a no-op and no longer call those calendar methods.
In those cases, round to 2 ns, rather than 1 ns.
Edits Temporal tests to account for changes in
https://github.com/tc39/proposal-temporal/pull/2574.
This PR stops coercing non-string primitive inputs to strings
in Temporal methods, to avoid cases where numbers
are coerced to syntactically valid but often unexpected
string results.
* Add tests for the new PrepareTemporalFields behavior for all direct callers
See https://github.com/tc39/proposal-temporal/pull/2570
* Add tests for the new PrepareTemporalFields behavior for indirect callers through ToTemporalDate
* Add tests for the new PrepareTemporalFields behavior for indirect callers through ToTemporalDateTime
* Add tests for the new PrepareTemporalFields behavior for indirect callers through ToTemporalZonedDateTime
* Add tests for the new PrepareTemporalFields behavior for indirect callers through ToTemporalYearMonth
* Add tests for the new PrepareTemporalFields behavior for indirect callers through ToTemporalMonthDay
* Add tests for the new PrepareTemporalFields behavior for indirect callers through ToRelativeTemporalObject
* Add tests for the new PrepareTemporalFields behavior for indirect callers through AddDurationToOrSubtractDurationFromPlainYearMonth
As per IETF review, an IXDTF string (ISO 8601 with annotations) is no
longer valid if it contains more than one u-ca annotation and at least one
of the annotations is marked critical.
Removes tests where such a string was assumed to be valid, and adds new
ones with a few variations on invalid strings.
This change affects how PlainYearMonth's add and subtract behave. They
now ignore daysInMonth reimplementations in the calendar, and they call
dateadd() a different number of times.
See https://github.com/tc39/proposal-temporal/issues/1315
Previously, "nested" calendar property bags were unwrapped up to one
level. That is, this object:
{
calendar: {
// ...Temporal.Calendar methods
}
}
would not be considered to implement the Calendar protocol, but would have
its calendar property used instead, if it were passed to an API that
required a Calendar protocol object.
These nested property bags are no longer supported. Discussion:
https://github.com/tc39/proposal-temporal/issues/2104#issuecomment-1409549753
Corresponding normative PR:
https://github.com/tc39/proposal-temporal/pull/2485
Compare semantics for custom calendars that _don't_ extend
Temporal.Calendar (and therefore don't have the internal slot) use the
value of the .id property, instead of calling toString().
Normative PR: https://github.com/tc39/proposal-temporal/pull/2482
In several tests involving custom calendars, we need to change the
implementation of dateFromFields/monthDayFromFields/yearMonthFromFields so
that the returned object gets the receiver as its calendar after chaining
up to the builtin implementation.
Normative PR: https://github.com/tc39/proposal-temporal/pull/2482
Has several effects on existing tests:
- Remove PlainTime from various tests that extract a Temporal.Calendar
instance from another Temporal object.
- Remove Temporal.PlainTime.prototype.calendar property.
- Remove calendar property from object returned from
Temporal.PlainTime.prototype.getISOFields().
- Ignore calendar annotation when converting ISO string to PlainTime.
Normative PR: https://github.com/tc39/proposal-temporal/pull/2482
In since() and until() methods, we copy the options object with
CopyDataProperties. Previously, its properties could be read in more than
one place (the method itself, and the calendar method), triggering user
code each time. Now, we pass around null-prototype objects with only data
properties.
See https://github.com/tc39/proposal-temporal/pull/2447
Previously in a few cases (calendar units in Duration) the value for the
roundingIncrement option had no upper limit, other than having to be
finite. These tests cover a normative change limiting it to 1e9.
Normative PR: https://github.com/tc39/proposal-temporal/pull/2480
The with() methods, as well as PlainYearMonth's since() and until(), were
adjusted to read the receiver's fields before the fields of any objects
provided as arguments. This change is observable, so affects several tests
that test the observed order of operations.
Normative PR: https://github.com/tc39/proposal-temporal/pull/2478
I used a script to "unroll" some of the calls to
TemporalHelpers.checkToTemporalCalendarFastPath(), and rewrite it slightly
not to use a specific Calendar instance. This is in preparation for a
future refactor, but also allows testing the PlainTime path which was not
previously covered.
(I didn't unroll all the calls and remove the helper yet, because the
remaining calls have custom assertions that I haven't gotten the script to
work with yet. For the same reason, I didn't yet convert these to use the
test generation facility. These will follow in a future PR.)
For each entry point where a string calendar name is accepted, we should
have a test that ensures the calendar name is case-insensitive. These
tests existed but several were incomplete as they didn't take nested
properties into account, and several entry points were missing this test.
Fix a minor copy-paste issue with double semicolons.
This contains tests for the normative PR
https://github.com/tc39/proposal-temporal/pull/2437, which is to be
presented for consensus to TC39 in the upcoming plenary meeting. That PR
changes the observable order of property accesses to be alphabetical where
possible.
As per the discussion in
https://github.com/tc39/proposal-temporal/issues/2379#issuecomment-1248557100
and the PR https://github.com/tc39/proposal-temporal/pull/2398, which is
to be presented for consensus to TC39 in the upcoming plenary meeting, UTC
offsets and the Z designator should be disallowed after any date-only
strings (YYYY-MM-DD, YYYY-MM, and MM-DD). They should only be allowed to
follow a time component. Z remains disallowed in any string being parsed
into a Plain type.
Annotations become allowed after any ISO string, even YYYY-MM and MM-DD
where they were previously disallowed.
This contains tests for the normative PR
https://github.com/tc39/proposal-temporal/pull/2433, which is to be
presented for consensus to TC39 in the upcoming plenary meeting. That PR
changes ToTemporalCalendar to throw when it encounters a Temporal.TimeZone
instance, and ToTemporalTimeZone to throw when it encounters a
Temporal.Calendar instance.
This adds order-of-operations tests that cover all of the Temporal entry
points that accept options bags, that were not already covered. We'll be
using these tests in the future to verify
https://github.com/tc39/proposal-temporal/issues/2254
However, the tests in this commit reflect the current state of the
proposal, not the potential normative change.
Some of the existing order-of-operations tests didn't observe the calendar
operations. Add a TemporalHelpers.calendarObserver() calendar to the
invocations of these methods in the order-of-operations tests.
Some of the existing order-of-operations tests didn't pass an options bag:
primarily from(), with(), add() and subtract(). (since() and until() were
covered in a previous commit.)
Add a TemporalHelpers.propertyBagObserver() options bag to the invocations
of these methods in the order-of-operations tests.
Use the new collections of strings in TemporalHelpers.ISO to add more
tests for ISO strings in API entry points that convert an ISO string to
Temporal.PlainYearMonth or Temporal.PlainMonthDay.
The idea is to deduplicate more string tests into methods on this object,
that return collections of valid and invalid strings. This adds
collections of valid and invalid PlainYearMonth and PlainMonthDay strings.
This tests some of the normative changes in
https://github.com/tc39/proposal-temporal/pull/2245, which achieved
consensus in the July 2022 TC39 meeting, specifically as they apply to
places where the MergeLargestUnitOperation was called.
Due to the use of the pre-existing spec operation CopyDataProperties, the
order of observable property operations has changed from a batch of
[[GetOwnProperty]] followed by a batch of [[Get]], to a series of
interleaved [[GetOwnProperty]]/[[Get]] pairs. This previously wasn't
tested because TemporalHelpers.propertyBagObserver didn't track
[[GetOwnProperty]] operations, but now it does.
Normally, a plain object passed into an API that takes a Temporal.Calendar
has its 'calendar' property checked (observably) with a Has operation
followed by a Get operation if the property is present. In the normative
change https://github.com/tc39/proposal-temporal/pull/2392 which reached
consensus at the September 2022 TC39 meeting, this was changed so that
this check is skipped for objects which have the Temporal.Calendar
internal slots.
This adds tests to all entry points that pass a user-supplied object to
ToTemporalCalendar, with a "poisoned" calendar object which has the
correct internal slots but a 'calendar' accessor property whose getter
throws. A correct implementation should not cause this getter to throw.
See https://github.com/tc39/proposal-temporal/pull/2316 which eliminated
an observable call to Array.prototype[Symbol.iterator]() in the case where
a calendar's 'fields' property was undefined.
The best way I've thought of to test this is to monkeypatch the
Array.prototype[Symbol.iterator]() method to make it throw. In some cases,
where we are actually expected to iterate the return value from a
Temporal.TimeZone's getPossibleInstantsFor() method, we have to provide a
custom method for that as well, that returns a non-Array iterable so we
don't call the patched Array.prototype[Symbol.iterator]().
This normative change reached consensus at the July 2022 TC39 plenary
meeting.
