Craig Macomber (Microsoft)

This commit performs a **client version bump**, updating the internal version of numerous packages from `2.92.0` to `2.93.0`. This **maintenance** task affects a broad spectrum of **Distributed Data Structures (DDSs)**, **drivers**, **runtime components**, **testing utilities**, and other **core client packages** by modifying their respective `packageVersion.ts` files. The change ensures consistent versioning across the client's internal dependencies, reflecting the progression to a new release `2.93.0`. This update is crucial for maintaining release alignment and dependency management within the project.

This commit **prepares the documentation** for the **client 2.92.0 release** by generating and adding the corresponding `RELEASE_NOTES/2.92.0.md` file. It specifically details **new features, deprecations, and API changes** introduced in this version, ensuring a comprehensive record of updates. This **documentation update** is a crucial step in the **release management process**, providing users and developers with a clear overview of changes for the new **client** version.

Update build-tools in gitrest and historian (#26906)

This commit **standardizes the formatting and content of four `.changeset` files** to comply with internal guidelines, ensuring consistency in release notes. It involves adding missing blank lines, correcting summary tenses to imperative present, removing backtick formatting from summaries, and eliminating internal issue references from the body of the changesets. Notably, one change's section was reclassified from `breaking` to `legacy` for a deprecated client API, aligning with the intended scope for major releases. This is a **maintenance task** that improves the **accuracy and consistency of generated changelogs** by adhering to established documentation standards.

This commit **refactors** the **`id-compressor`** package by **tagging an assertion** within the `toIdCompressorWithCore` function. A new short code, `0xced`, was added to the `assertionShortCodesMap` in `test-runtime-utils` to support this change. This **maintenance** effort aims to improve **error identification** and debugging clarity for a specific runtime assertion, making it easier to pinpoint issues related to ID compression.

This commit performs a **refactoring** and **maintenance** effort across several **Fluid Framework packages** to enhance diagnostic capabilities. It **adds numeric short codes to assertion messages** within various core components, including `dds/tree`'s `anchorSet`, `cursor`, `sharedTree`, and `treeNodeKernel`, as well as `runtime/container-runtime` and `runtime/id-compressor`. This work also updates the `assertionShortCodesMap.ts` utility to reflect these changes. The primary goal is to improve the clarity and traceability of assertion failures by providing concise identifiers, which is crucial for debugging and error reporting.

This commit primarily **refactors** the **`dds/tree` package** by **deduplicating logic** within the `charactersCopy` and `getString` methods in the `textDomainFormatted` module, leveraging a new private helper `getCharactersSubarray`. Concurrently, it introduces a **new utility function `forEachNodeSubsequence`** in the **`core/tree/cursor` module** to facilitate efficient iteration over node subranges. This work also **improves input validation** for `getString` by making it more strict and resolves a **duplicated assert short code**, enhancing the overall robustness and maintainability of the tree and text handling subsystems.

Rework Benchmark Package (#26638)

This commit **refactors** the **chunked forest codec** to explicitly separate and manage different versions of the **field batch encoding format**. It introduces distinct definitions for `V1` and `V2` formats, with `V2` adding support for incremental chunk shapes, and updates encoding/decoding logic to handle these versions. This ensures that **schema validation** correctly rejects `V2`-specific shapes when encoding to `V1`, improving data integrity within the `packages/dds/tree` library. The change enhances maintainability and **prepares the system for the introduction of future codec versions**.

This commit introduces and widely applies the `JsonCodecPart` interface across the **DDS Tree package's codec infrastructure**, particularly within `core/rebase` and various `feature-libraries` such as `modular-schema`, `optional-field`, and `sequence-field`. This **refactoring** effort clarifies that specific codec components, like `RevisionTagCodec` and those within `makeModularChangeCodecV1`, are parts of a larger codec and do not handle schema validation or version dispatch independently. By standardizing codec definitions to use `JsonCodecPart` and `RevisionTagSchema`, the change improves the **maintainability and clarity** of the serialization/deserialization logic, making the internal codec system more robust for future development. The impact is primarily internal, enhancing developer understanding and consistency without altering external APIs or user-facing behavior.

This commit **refactors** the **message codec** creation and version dispatching within the **`shared-tree-core`** module to standardize its implementation. It ports the `makeMessageCodec` and `makeMessageCodecs` functions to utilize the `ClientVersionDispatchingCodecBuilder`, mirroring the pattern established for `makeEditManagerCodec`. This involves introducing a new `makeMessageCodecBuilder` and updating `SharedTreeCore` and `sharedTree` to consume it, while also **removing deprecated APIs** that previously handled version dispatching manually. This **maintenance** effort simplifies the codebase by eliminating redundant logic and ensures all version-dispatching codecs follow a consistent, robust pattern, with no functional changes to external behavior.

This commit **refactors** the **`EditManager` codec** within the **`shared-tree-core`** and **`shared-tree`** modules by porting its creation to the `ClientVersionDispatchingCodecBuilder`. This **standardizes** the codec's version dispatching and schema validation, **deduplicating logic** and improving maintainability across the system. The change impacts `editManagerCodecs.ts`, `sharedTreeCore.ts`, and `sharedTree.ts`, which now leverage the new builder pattern for codec instantiation. This is a significant **refactoring** effort, aligning the `EditManager` with a more robust and consistent codec infrastructure.

This commit **refactors** the **codec infrastructure** within `packages/dds/tree` by moving **schema validation** for **schema change codecs** up to the **tree change codec** level. This is an **architectural improvement** aimed at centralizing validation, allowing for more consistent handling of encoded data and enabling future snapshotting of schemas per codec version. Specifically, it introduces `JsonCodecPart` and enhances `withSchemaValidation` in `codec.ts`, while `makeSchemaChangeCodec` in `schema-edits` and `makeSharedTreeChangeCodecFamily` in `shared-tree` are updated to integrate this new validation flow. This change **reduces redundant schema validation** and improves the robustness of data format checks, as verified by new test cases that confirm rejection of malformed schema-change data.

This commit **optimizes the testing process** for the **`packages/dds/tree`** module by conditionally disabling specific benchmark configurations. It modifies `sharedTree.bench.ts` and `simpleTree.bench.ts` to only run certain wide and deep tree benchmarks when a dedicated performance testing mode is active. This **maintenance** change significantly **reduces test execution time**, particularly on CI, leading to faster feedback cycles for developers.

This commit performs a **maintenance and refactoring** task by **removing the unused and redundant `SchemaChangeFormatVersion.v2`** from the **schema-edits codec system** within `dds/tree`. Since `v2` was functionally identical to `v1` and was never persisted, this change is **safe** and introduces no compatibility issues. It also **simplifies the codec family creation logic** in `schemaChangeCodecs.ts`, streamlining the `makeSchemaChangeCodecs` function and renaming a related function to `makeSchemaChangeCodecV1` for clarity. This improves code hygiene and reduces unnecessary complexity in the **schema change serialization infrastructure**.

This commit **refactors** the project's benchmarking infrastructure by **removing the dependency on an old benchmark package**. It **migrates and adds new performance benchmarks** within the `packages/dds/shared-object-base` module, specifically for the `FluidSerializer`'s `encode`, `stringify`, and `parse` methods. This **maintenance** effort ensures that performance testing for critical serialization logic is conducted using a modern and supported framework. The change provides more accurate and reliable performance insights for the **FluidSerializer** component, improving the overall quality of performance measurements.

This commit **refactors** the **schema change codec system** to **decouple** the schema codec version from the schema change codec, enhancing flexibility in schema evolution. The `makeSchemaChangeCodecs` function within the **`schema-edits` feature library** (`packages/dds/tree/src/feature-libraries/schema-edits/schemaChangeCodecs.ts`) was updated to remove explicit schema format version overrides, thereby **simplifying codec family creation**. This **architectural improvement** streamlines how schema changes are encoded and decoded, making the **schema versioning infrastructure** more robust. Additionally, documentation for the `metadata` property in `TreeNodeSchemaDataFormat` (`packages/dds/tree/src/core/schema-stored/formatV2.ts`) was clarified to improve understanding of schema data formats.

This commit **refactors the schema change codec creation and usage** within the `dds/tree` package, marking the first part of a broader cleanup effort. It introduces a `SchemaChangeFormatVersion` enum and transitions core components from direct `makeSchemaCodec` calls to a more structured `schemaCodecBuilder.build` approach. This **refactoring** simplifies internal schema codec management, impacting the `SharedTreeKernel` and `extractPersistedSchema` functions, as well as various test utilities. The changes primarily affect the `feature-libraries` and `shared-tree` modules, improving the robustness and maintainability of the system for handling schema evolution.

This commit focuses on **documentation improvement** for the **`SharedTree` and `SimpleTree` APIs**, enhancing clarity and conciseness across several key areas. It **refactors** and rewords documentation for tree events and the overall `SharedTree` overview, making explanations more direct. A significant terminology update involves consistently referring to 'node kinds' instead of 'node types' within the `docs/data-structures/tree` section. Additionally, JSDoc remarks were added to interfaces like `TreeBranch`, `ITree`, and `TreeChangeEvents` in the `simple-tree/api` to improve cross-referencing and clarify API usage. This work aims to provide developers with a more precise and easier-to-understand guide for interacting with the tree data structures.

This commit **updates and simplifies** the **documentation** for **schema evolution**, specifically focusing on the **allowed types rollout** process. The file `docs/docs/data-structures/tree/schema-evolution/allowed-types-rollout.mdx` has been revised to provide clearer explanations and more relevant code examples. This **documentation improvement** ensures that users have accurate and practical guidance for implementing schema changes, moving away from test-centric examples to real-world usage scenarios. The change primarily impacts developers seeking to understand and apply **data structure schema evolution**.