Add LSP server for integration package validation#3405
Draft
Add LSP server for integration package validation#3405
Conversation
Wrap existing package-spec validation in an LSP server (stdio) so editors can surface lint errors inline on open/save. Uses tliron/glsp for the JSON-RPC transport. Includes debounced per-package validation, error-to-file routing, and diagnostic clearing. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…guide Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
… LSP YAML field path resolution maps validation errors to exact lines. Document store tracks open file contents for hover/completions. Notify function captured at initialize for safe async use in debouncer. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Completions: field names from fields/*.yml in pipelines, field types in field definitions, manifest keys with context awareness. Hover: manifest key docs (nested paths resolved via indentation), field type/unit/metric_type docs, field reference info from field index. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
![macroscopeapp[bot]](https://avatars.githubusercontent.com/in/900172?s=60&v=4){kind=small}
Better error attribution for dashboard refs and folder items. Completion prefix filtering for types and manifest keys. Overlay FS enables validation with unsaved file contents. Expose ValidateAndFilterFromFS. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Tests for completions, hover, field index, workspace, overlay FS, document store, server lifecycle, and live diagnostics with real package validation. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Remove unused validatePackage function, range over string directly instead of []rune(string), align struct tags, regenerate README. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Manifest completions now resolve nested YAML context (e.g. completing keys inside policy_templates.inputs or streams.vars). Value completions suggest enum values and booleans from the package-spec JSON schemas. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- position.go: use len(quotedValue)-2 for JSON range end to handle escaped characters correctly - manifest_schema.go: store normalized schemaPath in valueCandidates so recursive branch calls use the resolved base path - lsp-demo.sh: fix \n in sed for BSD/macOS compatibility; fix trap quoting to defer $BROKEN_DIR expansion Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
| lines := splitLines(text) | ||
| quotedValue := strconv.Quote(value) | ||
|
|
||
| for lineIndex, line := range lines { |
There was a problem hiding this comment.
🟢 Low lsp/position.go:194
findJSONPropertyValueRange uses strings.Index byte offsets for protocol.Position.Character, but LSP requires UTF-16 code unit offsets. When the line contains multi-byte UTF-8 characters before the matched value, the computed positions are incorrect, causing diagnostic highlights to misalign in editors. Consider converting the byte offset to a UTF-16 code unit offset before constructing the position.
🚀 Reply "fix it for me" or copy this AI Prompt for your agent:
In file internal/lsp/position.go around line 194:
`findJSONPropertyValueRange` uses `strings.Index` byte offsets for `protocol.Position.Character`, but LSP requires UTF-16 code unit offsets. When the line contains multi-byte UTF-8 characters before the matched value, the computed positions are incorrect, causing diagnostic highlights to misalign in editors. Consider converting the byte offset to a UTF-16 code unit offset before constructing the position.
Evidence trail:
1. internal/lsp/position.go lines 190-213 at REVIEWED_COMMIT: shows `findJSONPropertyValueRange` using `strings.Index` (returns byte offset) and `len(quotedValue)` (returns byte count) directly as `Character` values in `protocol.Position`.
2. LSP specification reference from https://github.com/microsoft/language-server-protocol/issues/872 and https://github.com/microsoft/language-server-protocol/issues/376: Confirms LSP Position.Character requires UTF-16 code unit offsets by default, not byte offsets.
3. Go documentation confirms `strings.Index` returns byte index and `len(string)` returns byte count.
…ions Editors without YAML auto-indent (e.g. Neovim) place the cursor at column 0 on a new blank line. With pos.Character=0 the line prefix was empty, yamlIndent returned 0, and path resolution fell back to root level — showing all top-level keys instead of the current block's children. Fix: when the cursor is at column 0 on a blank/whitespace-only line, infer the expected indent from the nearest preceding non-blank line. A bare key (no inline value) implies indent+2; a key-value pair implies the same indent (sibling position). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…and subscription Three related completion fixes: 1. Root keys shown even at non-zero column: linePrefixAtPosition clamps to line length, so pos.Character>0 on a truly blank line still yields linePrefix="". Widen the blank-line inference guard from pos.Character==0 to yamlIndent(linePrefix)==0 to cover this. 2. categories (and other array-of-enum fields): valueCandidates only looked at the top-level schema node. For array types the enum lives in items, not the array node itself. Add items traversal so both "- |" list-item and inline "categories: sec" positions return the full enum list. 3. When completing in list-item position and the array holds scalar values (no object keys), fall back to completeArrayEnumItems instead of returning nil. conditions.elastic.subscription values were already working; added a regression test to pin the behaviour. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Collaborator
💔 Build Failed
Failed CI Steps
History
|
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
2 participants
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.
Summary
elastic-package lspcommand — a Language Server Protocol server (stdio) that wraps existing package-spec validation to surface lint errors inline in editorsfields/*.yml), manifest key completions, and field type completionsTest plan
go test ./internal/lsp/...— unit tests for error parsing, position lookup, URI conversion, document store, manifest schema./scripts/lsp-demo.sh ./elastic-package— end-to-end demo: valid package (no errors), broken package (errors routed to correct files), nested file triggers package-root discoverystreams.input) — verify docs shownfield:in a pipeline file — verify field name completions appear🤖 Generated with Claude Code / Opus 4.6