Conversation
…getTypeDefinition member lookups ILMethodDefs.lmap was building per-name buckets with repeated Array.append prepends — O(n²) total allocations for a type with N methods. Switch to a ResizeArray accumulator that produces each bucket array in a single ToArray call. TargetTypeDefinition.GetField / GetPropertyImpl / GetEvent previously bypassed the per-type lazy caches (fieldDefs / propDefs / eventDefs) and called the tx* wrapper factories on every invocation. Using Force() on the already-computed lazy arrays returns the same cached wrapper objects, avoids redundant allocations when the arrays are already warm, and guarantees consistent object identity across GetFields/GetField call pairs. 126/126 tests pass. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
dsyme
approved these changes
Apr 3, 2026
2 tasks
github-actions Bot
added a commit
that referenced
this pull request
Apr 7, 2026
- Performance: O(1) convTypeRef assembly-name lookup (#493) - Performance: O(N) ILMethodDefs index construction; lazy caches in TargetTypeDefinition (#497) - Refactor: save-based caching for GetField/GetEvent/GetNestedType; Dictionary in ILNestedExportedTypesAndForwarders (#498) - CI: NuGet and FAKE build caching (#495) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
dsyme
pushed a commit
that referenced
this pull request
Apr 7, 2026
🤖 *This PR was created by [Repo Assist](https://github.com/fsprojects/FSharp.TypeProviders.SDK/actions/runs/24058266578).* ## Summary Bumps the version to **8.5.0** by adding an entry to `RELEASE_NOTES.md` covering all changes merged since 8.4.0: - **Performance**: O(1) assembly-name dictionary lookup in `convTypeRef` (#493) - **Performance**: Avoid O(n²) allocations in `ILMethodDefs` name-index construction; reuse lazy caches in `TargetTypeDefinition` for `GetField`/`GetPropertyImpl`/`GetEvent` (#497) - **Refactor**: `save`-based caching for `GetField`/`GetEvent`/`GetNestedType` on `ProvidedTypeDefinition`; `Dictionary` in `ILNestedExportedTypesAndForwarders` (#498) - **CI**: NuGet and FAKE build caching (#495) No code changes — RELEASE_NOTES.md only. ## Test Status ✅ 126/126 tests pass (`dotnet test tests/FSharp.TypeProviders.SDK.Tests.fsproj -c Release`). > Generated by 🌈 Repo Assist, see [workflow run](https://github.com/fsprojects/FSharp.TypeProviders.SDK/actions/runs/24058266578). [Learn more](https://github.com/githubnext/agentics/blob/main/docs/repo-assist.md). > > To install this [agentic workflow](https://github.com/githubnext/agentics/blob/7ee2b60744abf71b985bead4599640f165edcd93/workflows/repo-assist.md), run > ``` > gh aw add githubnext/agentics@7ee2b60 > ``` <!-- gh-aw-agentic-workflow: Repo Assist, engine: copilot, model: auto, id: 24058266578, workflow_id: repo-assist, run: https://github.com/fsprojects/FSharp.TypeProviders.SDK/actions/runs/24058266578 --> <!-- gh-aw-workflow-id: repo-assist --> --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
🤖 This is an automated PR from Repo Assist.
Summary
Two complementary performance improvements to
TargetTypeDefinitionand the internalILMethodDefsname-index.1. O(1)-per-entry method-name index construction (
ILMethodDefs.lmap)Before: the lazy name→methods dictionary was built by prepending each new entry to an existing array:
For a type with N methods sharing a name, this allocates arrays of sizes 1, 2, 3 … N — O(N²) total elements. The BCL contains types (e.g.
String,Convert) with dozens of overloads per name, making this measurable.After: a
ResizeArrayaccumulator collects entries per name in O(1) each, then a singleToArray()call per bucket produces the final arrays — O(N) total elements.2. Lazy-cache reuse in
GetField,GetPropertyImpl,GetEventBefore: each call to
GetField/GetPropertyImpl/GetEventscanned the rawILFieldDef/ILPropertyDef/ILEventDefarrays and called thetxIL*factory to create a brand-new wrapper object:This created a new allocation on every call even when
fieldDefs/propDefs/eventDefslazy arrays were already fully computed (e.g. after a precedingGetFields(bindingFlags)call).After: the methods call
Force()on the already-computed lazy arrays and filter by name. When the caches are warm (the common case during compilation) zero new wrappers are allocated, and callers get consistent object identity acrossGetFields/GetFieldpairs.Test Status
✅ 126/126 tests pass (
dotnet test tests/FSharp.TypeProviders.SDK.Tests.fsproj).