Add client and server support to navigate, inspect, reference, and rename block
template and address symbols declared in PSC .def files:
- Client: register language providers for .def (definition, hover, references,
prepare/provide rename) and send the current document text with each request.
- Server: parse .def files into DefDocument/DefDeclaration/DefReference, expose
def-specific LSP endpoints (definition/hover/references/prepareRename/rename),
and integrate .def lookups into existing Tcl hover/definition/references/rename
flows so Tcl calls jump to .def declarations.
- Tcl server keeps a snapshot API for .def documents (current editor content can
replace file on request); only declared names in loaded .def files can be
renamed. Name validation uses DEF_NAME_RE.
Update README and CHANGELOG to document navigation features.
- Pass extension storage path to the server (client/ changes) so the
server can persist a workspace index.
- Introduce IndexCache (server/tools/index_cache.py) and load/save it on
initialization and after background indexing. Index entries are stored
only when the file's stat hasn't changed while being read.
- Add incremental reparse logic (server/tools/incremental_parse.py) and
use a per-file _last_parse cache in the language server to reparse only
the top-level Tcl commands touched by an edit, falling back to a full
parse when necessary.
- Use a new _FileIndex dataclass and _build_file_index helper to unify
what is stored/loaded for a file; update update_poco_completion_for_file
to use the persistent cache for disk-read files (from_disk/source_stat).
- Keep background indexing non-blocking and persist the index at the
end of the run. Add basic unit tests for incremental parse and index cache.
Before: edits and background work always required full parsing of files
and no persistent cross-restart index. After: some edits reuse previous
ASTs and files read from disk can use a persisted index to skip
re-indexing across restarts.
Introduce several changes to reduce full-document reparses, lock contention and
redundant work when handling TclOO analysis and navigation:
- Add a cheap may_contain_classes pre-check and several cursor/receiver
heuristics so completions, signature help and tcloo definitions skip the
expensive marker reparse when the document cannot contain useful OO info.
- Allow passing an existing parsed tree into tcloo completion/signature/definition
helpers; update callers to use the server's cached tree when available.
- Use a thread-local parser for request-time parse_source to avoid blocking the
shared parser during background indexing, and add navigation_state() which
returns cached definition identities (invalidated on index generation changes).
- Add a cheap name pre-filter (_may_resolve_to) for symbol matching and only
run class highlighting when classes may exist.
These changes reduce contention and repeated parsing, improve responsiveness for
requests during background indexing, and cache navigation definition identities.
Tests were added/updated to assert caching and non-blocking behavior.
Add PSC (.psc) indexing and share TclOO class metadata across files so class
definitions discovered via PSC layers can be used for completions, signature
help, inlay hints, and "go to definition". Key behavior changes:
- Client file watcher now includes *.psc and .vscode launch paths/tests updated
to use the postprocessor test folder; .gitignore updated to ignore that folder.
- Server watches .psc changes and refreshes a PSC script index; new
tools/tcloo_navigation.py exposes tcloo_definition used by the language server
to resolve cross-file class/constructor/method definitions.
- Language server uses class_snapshot(document.path) when producing TclOO
completions, signature help, and inlay hints so resolved class metadata is
available across files.
Also includes related docs/changelog updates, minor code formatting cleanups,
and added tests for PSC/TclOO behavior.
Add conservative, document-local TclOO type inference and tooling so the server
can offer method completions, signature help, and inlay parameter hints for
statically resolvable TclOO receivers (including `new`/`create`, `my`, and
simple return-chains). Also surface class names as completion items and emit
semantic tokens for class declarations/references.
Notable changes:
- New tcloo_* tools: completion, symbols, and argument parsing; integrated into
on_completion, signature_help, inlay hint generation and semantic token
highlighting. Completions are returned early when an OO receiver context is
detected.
- Use a completion-friendly parser fallback when the main AST fails (TclSyntaxError)
so editing-in-progress code still yields useful completions.
- Add CompletionItemKind.Class to command kinds, exclude class items from the
poco completion name cache, and include new unit tests for the TclOO helpers.
Adds context-aware completion ranking and document highlights for Tcl.
Adds completion_context to distinguish variables and commands.
Wires per-file completion snapshots and ranking into the flow.
Adds document_highlight provider support and tests for highlights.
- Context-aware ranking of completion items using per-file snapshots
- Document highlight provider wired into initialization and tests
- Tests for completion context, ranking, and document highlights
The changes align the project with the 2025.0.0 lsprotocol
release, removing the old backport and updating type hints
in the protocol hooks to use Sequence where appropriate. The
dist-info and packaging metadata for older lsprotocol
versions are replaced with the new 2025.0.0 artifacts.
- Remove exceptiongroup backport used on Python <3.11
- Use Sequence instead of List in LS protocol hooks
- Replace old dist-info with 2025.0.0 metadata
This change adds cached and incremental analysis for the LSP
server to improve responsiveness. The client now debounces
diagnostic updates to avoid excessive recomputation. The
server introduces per-document line caches and various
caches for completions, inlay hints, and metadata to
support faster, incremental updates.
- Debounce diagnostics on text changes to reduce noise.
- Add caches for completions, inlay hints, and metadata.
- Introduce incremental analysis with per-document line caches.
The changes introduce a Tcl symbol index powering LSP navigation
features across the workspace. A navigation API exposes
snapshots and update hooks, enabling goto-definition,
references, and rename using the index. Background indexing
now watches Tcl files and rebuilds the index to stay in sync.
- Add Tcl symbol index and navigation snapshot API
- Wire go-to-definition, references, and rename using the index
- Watch Tcl files and refresh the index in the background
Adds a thread-safe incremental index and snapshot API for LSP.
Introduces cache invalidation and file operation hooks for delete
and rename. This keeps indices in sync with disk changes.
Supports reindexing TCL files from disk when needed.
- Adds workspace file change handlers to sync indices on delete/rename.
- Introduces locking and snapshot helpers to safely access shared state.
- Refactors to invalidate caches on edits and reindex TCL files.
Christoph
Christoph Brandau