Releases: TiManGames/hpl3-language-tools
Releases · TiManGames/hpl3-language-tools
Release list
v1.1.0
- Moves open-document parsing and workspace indexing/diagnostics into separate workers so completion, hover, definitions, colors, and other editor requests no longer queue behind large-file analysis or validation.
- Adds atomic immutable project snapshots, resolver-catalog transfer, granular declaration/include invalidation, changed-range metadata, cancellable background diagnostics/references/rename, and throttled latest-version-first document processing.
- Adds a versioned binary project cache in VS Code global storage with file and grammar fingerprints, safe corruption recovery, forced-rebuild bypass, and cache/index progress in the HPL3 status tooltip.
- Adds cached transferable semantic-token segments with full delta and range support, cached color ranges, viewport-bounded inlay hints, stable completion symbol IDs, and structured verbose-trace timing budgets.
- Adds a bundled-server benchmark against the real 650 KB Bunker agent script and 22,000-symbol API. The verified local run reaches 4.7 ms edit brokerage, 33.5 ms completion P95, 99 ms document snapshots, 887 ms scheduled diagnostics, 1.28 s cold indexing, and 489 ms cached startup.
- Adds declaration-aware references, read/write document highlights, and conservative code-symbol rename. Literal callback references are discoverable but are never rewritten by rename.
- Adds class/interface type hierarchy and implementation navigation.
- Adds configurable parameter-name, out/inout, and inferred-type inlay hints.
- Adds literal cColor(...) editor color previews and presentations.
- Adds quick fixes for array.clear(), reliably described missing callbacks, and inherited method stubs.
- Expands semantic highlighting to functions, methods, properties, and variables using resolved occurrences.
- Resolves completion documentation lazily and ranks compatible expected types ahead of unrelated symbols.
- Keeps large-script editing responsive by caching visible include symbols and semantic occurrences, and by classifying semantic tokens in one lexical pass.