core.id — unique identifiers

Generators for identifiers that must not collide.

They lived in core/base/ until the stdlib's ring law measured the consequence: base is ring 0, the language's own vocabulary, and these modules mount core.time.system_time (ring 2). core.base.snowflake was claiming that a Snowflake ID is a primitive of the LANGUAGE, alongside Maybe and Result. It is not — it is a timestamp, a machine id and a sequence number.

module	shape
core.id.uuid	RFC 9562 UUIDs
core.id.ulid	lexicographically sortable, timestamp-prefixed (Crockford base32)
core.id.nanoid	short, URL-safe, entropy only — 21 chars, 126 bits
core.id.snowflake	64-bit: timestamp + machine + sequence

An identifier is not a secret

All four draw from core.random.secure, so the entropy is sound. That is a property of the GENERATOR, not of the format: ulid and snowflake embed a timestamp by design and reveal their creation time to anyone holding one. If you need a value an observer cannot correlate, nanoid carries no structure — or draw the bytes yourself.

None of these is a hash. For content addressing see core.hash.crypto; for a keyed tag see core.mac.

Per-module status

file	status	detail
uuid.vr	partial	core-tests/base/uuid — Uuid.parse/from_bytes stable; Uuid.to_text previously lenient-skipped on Text.new (one of ~400 Text.* method cascade entries at May-11 baseline). Tracked under task #16 (twice attempted reland regressed; runtime sentinel handler durably landed via commit b5f5462d4) (commits b5f5462d4 runtime sentinel handler + 5b657aa16 stage-1 + 98fced985 stage-2 + 7cdb334b2 + 86784eebf stub-overwrite discipline): caller-modules now see Text.new / List.from / AtomicInt.new / Duration.from_secs / variant constructors as stubs from compile-time-zero; producing modules' real bodies overlay the stubs via the sentinel-ID-range overwrite gate at stdlib_bootstrap.rs:1453; stuck stubs (when producing module itself fails precompile) degrade gracefully via the runtime sentinel handler with a clean [lenient] task#16 stage-N stub never resolved panic. Stage 1 targets 386 Text-method skips; combined with Stage 2's 131 variant-constructor skips, ~631 of 1333 fresh-baseline skips (≈47%) close as one architectural fix.
ulid.vr	partial	core-tests/base/ulid — 4/4 test files green under --interp (was 1/3 pre-fix). 28 unit + 11 property + 11 integration + 8 regression pins. Pure-data surface (from_parts(int, int) factory, Ulid.from_parts_seeded, parse validation, ULID_ALPHABET Crockford pin, UlidError 2-variant ADT, 48-bit timestamp round-trip) fully covered. Two @ignore'd defect classes pinned: (a) task §2.1 — Ulid.to_text() lenient panic-stub because Text.from_utf8_unchecked was unsafe fn (private) at core/text/text.vr:455; visibility fix staged in this branch (public unsafe fn); flips green on next precompiled-stdlib refresh. (b) task #17/#39 — SystemTime.now() static-method dispatch defect: Ulid.new() / Ulid.now() / generate() internally call SystemTime.now().timestamp_millis() which mis-routes to SysTimeOpsInstant.now() (same root as snowflake §D). Workaround: deterministic from_parts(int, int) constructors throughout test suite.
nanoid.vr	partial	core-tests/base/nanoid — 4/4 test files green under --interp (was 1/3 pre-fix). 16 unit + 14 property + 9 integration + 12 regression pins. Pure-data surface — NANOID_ALPHABET (64-char URL-safe alphabet: A-Z + a-z + 0-9 + _ + -), NANOID_DEFAULT_LENGTH = 21, alphabet composition, URL-safety, 126-bit-entropy invariant — fully covered. Three @ignore'd pins for two cascading stdlib defects (both fixed in this branch, activate on next precompiled-stdlib refresh): (a) §A Text.from("") undefined — core/base/nanoid.vr:138 called a non-existent Text.from factory; fix: → Text.new(). (b) §B Text.from_utf8_unchecked private — same fix as ulid §A (text.vr:455 unsafe fn → public unsafe fn). Combined effect: every code path through generate_with_alphabet (including generate() and generate_len) lenient-stubbed at precompile, panics at runtime. Workaround: constant + alphabet-invariant surface exclusively.
snowflake.vr	partial	core-tests/base/snowflake — 4/4 test files green under --interp (was 2/3 pre-fix). 24 unit + 14 property + 10 integration + 9 regression pins. Pure-data surface (Snowflake.new validation, layout constants, SnowflakeError variants + Eq matrix, parse · build identity, SnowflakeParts construction, accessors) fully covered. Live next_id() gated on task #17/#39 — SystemTime.now() static-method dispatch defect: next_id internally calls SystemTime.now().timestamp_millis(), but the dispatcher mis-routes to SysTimeOpsInstant.now() (a sibling type with a 1-field layout), causing field access out of bounds: field index 1 (offset 8+8 = 16) exceeds object data size 8 type_id=... type='SysTimeOpsInstant'. 3 @ignore'd pins (regression_test.vr §D) for live next_id / monotonicity / live-ID-parse-round-trip. Workaround in test suite: synthetic IDs built via (ts_offset << TIMESTAMP_SHIFT) | (worker << WORKER_SHIFT) | seq, parsed via parse(id, epoch_ms).