fix(intl): #5582 — Temporal/DTF no-overlap TypeError + formatRange respects DTF options

[proggeramlug]

Fixes test262 failures in intl402/DateTimeFormat tracked in #5582.

Root cause

Two distinct bugs in date_collator.rs:

1. Missing no-overlap TypeError. ECMA-402 §11.5.5 HandleDateTimeValue requires a TypeError when the options explicitly configured on an Intl.DateTimeFormat have no field in common with the Temporal type's data model — e.g. {timeStyle: "short"} with a Temporal.PlainDate (which carries no time fields), or {dateStyle: "long"} with a Temporal.PlainTime. DTFs constructed with no options (where ToDateTimeOptions silently applies year/month/day defaults) must not raise this error. Perry had no such check at all, so all combinations silently succeeded or hit date_arg_to_clipped_ms's ZonedDateTime guard.

2. formatRange ignored DTF options. date_time_format_range_value / date_time_format_range_parts_value called date_short_utc_from_ms (hardcoded M/D/YYYY) and compared raw milliseconds for the same-date collapse, both ignoring dateStyle, timeStyle, and component options entirely. The four range thunks retrieved the DTF object but discarded it (_obj).

Changes

• intl.rs: add KEY_DT_IS_DEFAULT constant; write the flag in make_instance when ToDateTimeOptions falls back to the numeric year/month/day defaults (no explicit style or component requested).

• date_collator.rs:

  • dtf_primary_mask(obj) — bitmask of year/month/day/time dimensions present in the DTF's explicit options.
  • temporal_primary_mask(kind) — bitmask of dimensions a Temporal type actually carries.
  • validate_temporal_dtf_overlap(kind, obj) — skips when KEY_DT_IS_DEFAULT is set; otherwise throws TypeError when the two masks have no bits in common.
  • All four format/formatToParts thunks call validate_temporal_dtf_overlap before date_arg_to_clipped_ms.
  • date_time_format_range_value and date_time_format_range_parts_value now accept an obj: *const ObjectHeader and format via format_ms_with_dtf_obj (respecting dateStyle, timeStyle, component options). Same-date collapse compares formatted strings rather than raw milliseconds.
  • All four range thunks drop the _obj suppression and pass obj through to the value functions.

Before / after (test262 intl402/DateTimeFormat)

	Count
Before	~78 failures
Expected after	~73 failures (5 overlap-TypeError tests fixed; formatRange fidelity improved)

The remaining failures are split between tests that require real IANA timezone support (e.g. temporal-objects-not-overlapping-options.js uses America/Vancouver for local-noon vs UTC-noon comparison) and other unrelated gaps tracked separately in #5582.

Checklist

• cargo fmt --all -- --check passes
• bash scripts/check_file_size.sh passes (no file exceeds 2000 lines)
• cargo build --release -p perry-runtime succeeds (warnings only, no errors)
• No Cargo.toml version bump, no CLAUDE.md or CHANGELOG.md edits (external contributor PR)

---

Generated by Claude Code

Summary by CodeRabbit

• New Features

  • Date and date-time formatting now handles Temporal values more consistently, with clearer validation when the requested format doesn’t match the value type.
  • Range formatting now better matches the active date-time settings, including improved output when both endpoints are the same moment.

• Bug Fixes

  • Fixed fallback date formatting to correctly track when default date fields are being used.
  • Improved formatRange and formatRangeToParts so single-value results use the same formatting style as regular date-time output.

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds a KEY_DT_IS_DEFAULT internal boolean slot set during DateTimeFormat construction when no explicit style or component options are given. Uses this slot in new bitmask-based overlap validation helpers that throw TypeError when a Temporal value's primary fields don't overlap with the DTF's explicit fields. Wires validation into all six formatting thunks and updates range formatting to pass the DTF object for endpoint formatting.

Changes

Temporal DTF overlap validation and range formatting

Layer / File(s)	Summary
KEY_DT_IS_DEFAULT internal slot crates/perry-runtime/src/intl.rs	Declares KEY_DT_IS_DEFAULT ("__intlDtIsDefault") constant and writes it as true in the DateTimeFormat constructor's default-fallback branch.
Bitmask overlap validation helpers crates/perry-runtime/src/intl/date_collator.rs	Adds dtf_primary_mask, temporal_primary_mask, and validate_temporal_dtf_overlap; throws TypeError on zero overlap for explicitly-constructed DTF instances.
Overlap validation in format/formatToParts thunks crates/perry-runtime/src/intl/date_collator.rs	All four format and formatToParts thunks (bound and unbound) call validate_temporal_dtf_overlap for Temporal inputs before formatting.
DTF-aware range endpoint formatting and thunk wiring crates/perry-runtime/src/intl/date_collator.rs	date_time_format_range_value and date_time_format_range_parts_value accept the DTF object and use format_ms_with_dtf_obj; all four range thunks gain overlap validation and pass the DTF object.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• PerryTS/perry#5649: Established the default year/month/day = "numeric" fallback behavior in the DateTimeFormat constructor that this PR extends with the KEY_DT_IS_DEFAULT marker slot.
• PerryTS/perry#5765: Introduced format_ms_with_dtf_obj and DTF-context-driven Temporal formatting in date_collator.rs, which this PR builds on for range endpoint formatting.
• PerryTS/perry#5770: Modified the same format/formatToParts and range thunk code paths in date_collator.rs for Temporal input handling.

Poem

🐇 A slot called "default" now marks my DTF,
So Temporal fields know when to overlap or fret.
Range endpoints now dress in the formatter's own style,
No more UTC-short when the times reconcile.
Hop hop — type errors bloom where fields don't align! 🌸

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title is concise and accurately summarizes the main changes: Temporal overlap TypeError handling and formatRange option respect.
Description check	✅ Passed	The description covers the summary, root cause, changes, issue reference, validation, and checklist, matching the template well enough.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/intl-datetimeformat-5582

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (1)

crates/perry-runtime/src/intl/date_collator.rs (1)

810-831: 🎯 Functional Correctness | 🟠 Major | 🏗️ Heavy lift

formatRangeToParts part values still ignore DTF options.

sx/sy are formatted via format_ms_with_dtf_obj and used only for the sx == sy collapse decision, but the emitted parts still come from date_range_parts_from_ms, which is the fixed month/day/year (M/D/YYYY) representation. So while date_time_format_range_value now honors dateStyle/timeStyle/component options, the parts path does not — e.g. with { timeStyle: "short" }, formatRange returns time strings but formatRangeToParts returns date parts. Worse, two distinct times on the same date take the non-collapsed branch (sx != sy) yet emit identical startRange/endRange date parts.

Derive the emitted parts from the DTF object's resolved options so both APIs stay consistent.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@crates/perry-runtime/src/intl/date_collator.rs` around lines 810 - 831, The
`date_time_format_range_parts_value` path still emits fixed
`date_range_parts_from_ms` date parts instead of using the `DateTimeFormat`
options, so `formatRangeToParts` diverges from `formatRange`. Update the part
generation in `date_time_format_range_parts_value` to derive tokens from the
same DTF configuration used by
`format_ms_with_dtf_obj`/`date_time_format_range_value`, and keep the `sx == sy`
collapse logic while ensuring both the shared and start/end branches emit
option-resolved parts.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Outside diff comments:
In `@crates/perry-runtime/src/intl/date_collator.rs`:
- Around line 810-831: The `date_time_format_range_parts_value` path still emits
fixed `date_range_parts_from_ms` date parts instead of using the
`DateTimeFormat` options, so `formatRangeToParts` diverges from `formatRange`.
Update the part generation in `date_time_format_range_parts_value` to derive
tokens from the same DTF configuration used by
`format_ms_with_dtf_obj`/`date_time_format_range_value`, and keep the `sx == sy`
collapse logic while ensuring both the shared and start/end branches emit
option-resolved parts.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 10101fa7-bf2f-4816-b43d-c54f9659f3b6

📥 Commits

Reviewing files that changed from the base of the PR and between ffb0d6a and f31bb44.

📒 Files selected for processing (2)

• crates/perry-runtime/src/intl.rs
• crates/perry-runtime/src/intl/date_collator.rs