# Changes Logs

## v3.0.0

- build: migrated from CommonJS to ES Modules
- refactor: changed main entry to namespace re-exports (`Snowflake`, `SnowflakeSI`, `UUIDv4`, `UUIDv5`, `UUIDv7`, `Constants`, `Errors`)
- feat: added subpath exports (`@litert/uuid/snowflake`, `@litert/uuid/snowflake-si`, `@litert/uuid/uuid4`, `@litert/uuid/uuid5`, `@litert/uuid/uuid7`, `@litert/uuid/errors`, `@litert/uuid/constants`, `@litert/uuid/utils`)
- chore: updated minimum Node.js version to 20.0.0
- feat: added utility functions for UUID string/buffer conversion
- doc: updated documentation for ESM import patterns

## v2.2.0

- feat(uuid-v7): added implements of UUIDv7 algorithm.
- fix(uuid-v4): fixed the variant of generated UUIDv4 IDs.
- feat(uuid-v4): added `generateAsBuffer` method to generate UUIDv4 as a 16-byte buffer.
- feat(uuid-v5): added `generateAsBuffer` method to generate UUIDv5 as a 16-byte buffer.
- feat(uuid-v5): added static `generate` and `generateAsBuffer` methods to generate a UUIDv5 string or buffer.

## v2.1.0

- feat(algo): added `bulkGenerate` for both Snowflake and Snowflake-SI algorithms.

## v2.0.3

- fix(algo): reverted the update of `v2.0.2` that removed the reset of sequence number to 0.

- feat(algo): added `onTimeReversed` option to Snowflake (SI) algorithms.

    This option allows the user to specify a strategy for handling time reversals in the Snowflake (SI) algorithm.
    The default strategy is `THROW_ERROR`, to keep the compatibility with previous versions.

- feat(algo): added `onTimeChanged` option to Snowflake (SI) algorithms.

    This option allows the user to customize the behavior of the sequence number, when the time is changed.
    The default strategy is to `RESET` the sequence number to 0, which is compatible with previous versions.

    This option could cooperate with the `sequenceResetThreshold` option, so that only when the current sequence
    number is up to the threshold, the sequence number will be reset to 0.

## ~~v2.0.2~~

- ~~fix(algo): snowflake/snowflake-si class should not reset sequence to 0 avoiding sequence overflow.~~

    ~~To ensure the randomness of the generated IDs, the sequence number must keep increasing even if the time is~~
    ~~changed. Thus an additional counter will be used to track the quantity of IDs generated in the same millisecond.~~

## v2.0.1

- fix(algo): snowflake class constructor should check if `epoch` is invalid.

## v2.0.0

- feat(algo): added implements of GUIDv4 algorithm.
- feat(algo): added implements of GUIDv5 algorithm.
- fix(algo): allow Snowflake using custom epoch.
- fix(algo): implement Snowflake-SI with more customization options.
- deprecate(algo): removed the Snowflake-SI-vA algorithm, which is no longer recommended.
- test(project): added unit tests.
- deprecate(deps): removed all runtime dependencies.

## v1.0.0

- Refactored the project.
- Added implements of standard Snowflake algorithm.

## v0.1.2

- Remove useless dependencies.

## v0.1.1

- Fixed the position of MID of UUID in Snowflake-SI-vA.