Skip to content

CHANGELOG.md

Latest commit

 

History

History
136 lines (86 loc) · 7.49 KB

CHANGELOG.md

File metadata and controls

136 lines (86 loc) · 7.49 KB

Change log

All notable changes to the project will be documented in this file. This project adheres to Semantic Versioning.

2.1.2 (2026-04-06)

Bug Fixes

  • Add unbounded context kind to segment type (#39) (aa2504f)

2.1.1 (2026-03-06)

Bug Fixes

2.1.0 (2026-03-02)

Features

  • Add support for flag / segment builders (#33) (fd63046)

2.0.2 (2026-01-26)

Bug Fixes

2.0.1 (2025-01-28)

Bug Fixes

2.0.0 (2024-07-10)

⚠ BREAKING CHANGES

  • Add exclude from summary field to flag (#18)
  • Add migration and sampling ratio fields (#17)

Features

  • Add exclude from summary field to flag (#18) (ca935a3)
  • Add method to strip anonymous contexts (#21) (f76faa6)
  • Add migration and sampling ratio fields (#17) (5e4e969)
  • Add serialize trait to Detail (#16) (683231e)

Bug Fixes

1.2.0 (2024-03-14)

Features

  • Support redacting all attributes from anonymous contexts (#14) (76d5e52)

[1.1.1] - 2023-08-07

Fixed:

  • Fixed an issue with evaluation of segments which included/excluded users when that user was part of a multi-context.

[1.1.0] - 2023-05-05

Changed:

  • Updated MSRV from 1.60.0 to 1.64.0

Fixed:

  • Fix invalid encoding for fully qualified context key generation.

[1.0.0] - 2022-12-06

This release of the evaluation engine corresponds to the upcoming v1.0.0 release of the LaunchDarkly server-side Rust SDK (launchdarkly-server-sdk), and is not compatible with earlier SDK versions.

Added:

  • Added the types Context, Kind, ContextBuilder, MultiContextBuilder, Reference, and supporting types. Context defines the new context-based evaluation model. Context replace User, which was removed.
  • Added: urlencoding @ 2.1.0
  • Added: maplit @ 1.0.1
  • Added: itertools @ 0.10.3
  • Added: serde_with @ 2.1.0

Changed:

  • The MSRV is now 1.60.0
  • Updated: chrono from 0.4.15 to 0.4.23; only enable the "std" feature
  • Updated: semver from 0.10.0 to 1.0.14
  • evaluate now takes a Context instead of a User
  • User keys could previously be empty strings. With contexts, the key cannot be empty.
  • The "secondary" meta-attribute which affected percentage rollouts has been removed. If you set an attribute with that name in a context, it will be a custom attribute like any other.
  • For backwards-compatibility, it is possible to enable "secondary" evaluation logic within the evaluation engine by enabling the secondary_key_bucketing flag. This will only affect contexts that were created via deserialization, since it is not possible to set "secondary" via a builder method.

Removed:

  • User, UserBuilder, UserAttributes. See Context, ContextBuilder and MultiContextBuilder instead.

[1.0.0-beta.5] - 2022-04-04

Changed

  • Update to edition 2021.
  • Change store signature to support persistent stores.

Added

  • Support flag serialization of old and current schema.
  • Add versioned trait (implemented by flag and segment).

[1.0.0-beta.4] - 2022-03-07

Changed

  • Bump sha1 and test_case dependencies.

Fixed

  • When serializing a user, custom attributes were incorrectly being flattened in the JSON output. We now correctly encode them under a custom key.

[1.0.0-beta.3] - 2022-02-16

Added

  • Add support for globally and user specific private attributes.

Changed

  • If we receive a negative variation index, or an unsupported operation, the SDK should not fail to parse the payload.

[1.0.0-beta.2] - 2022-01-21

Changed

  • Modified the try_map function to accept a default value when returning a new Detail instance.

[1.0.0-beta.1] - 2022-01-19

Initial release of flag evaluation support code that will be used with the LaunchDarkly Server-Side SDK for Rust.