Improve codemark skill with fine-grained execution logic queries#25
Improve codemark skill with fine-grained execution logic queries#25DanielCardonaRojas merged 1 commit intomainfrom
Conversation
📝 WalkthroughWalkthroughDocumentation updates to the CodeMark skill introducing fine-grained bookmarking strategies. Enhanced best practices guidance and added language-specific query patterns for targeting specific execution boundaries (call expressions, conditional statements, switch statements) within functions rather than entire function definitions. Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~12 minutes Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 3
🧹 Nitpick comments (1)
extras/skills/codemark/queries/common.md (1)
188-188: Minor wording polish: hyphenate compound modifier.At Line 188, consider “error-handling branches” for cleaner compound-modifier style.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@extras/skills/codemark/queries/common.md` at line 188, Replace the phrase "error handling branches" with the hyphenated compound modifier "error-handling branches" in the text at the location that currently reads "Target the exact decision points, such as authorization guards or error handling branches." to correct compound-modifier style.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@extras/skills/codemark/examples.md`:
- Around line 38-41: The example command uses JS/TS tree-sitter node types
(member_expression, property_identifier) which won't match a Rust file; update
the --query value in the codemark add-from-query example to use Rust node types
(e.g., field_expression and field_identifier) and keep the capture name `@target`
and the method identifier token (currently "update_user_balance") so the command
still targets the same call site; ensure the query replaces
member_expression→field_expression and property_identifier→field_identifier (and
adjust any predicate names if needed) so it matches Rust method calls in
src/db/repo.rs.
In `@extras/skills/codemark/queries/go.md`:
- Around line 112-115: The pattern uses the wrong Go Tree-sitter node kind
`switch_statement`; update the query to target the correct Go node kinds such as
`expression_switch_statement` (or `type_switch_statement` when matching type
switches) instead of `switch_statement`, keeping the capture names `@val` and
`@target` unchanged so the rest of the query logic still binds the switch value
and target capture.
In `@extras/skills/codemark/queries/java.md`:
- Around line 128-132: The Tree-sitter Java query is using the wrong node type
and field: replace uses of switch_statement and its condition field with the
correct node type switch_construct and its selector field; update the pattern
around (switch_statement condition: (parenthesized_expression (identifier) `@val`
(`#eq`? `@val` "action"))) `@target` to reference switch_construct and selector so the
query matches tree-sitter-java v0.23 (preserve the `@val` and `@target` capture
names).
---
Nitpick comments:
In `@extras/skills/codemark/queries/common.md`:
- Line 188: Replace the phrase "error handling branches" with the hyphenated
compound modifier "error-handling branches" in the text at the location that
currently reads "Target the exact decision points, such as authorization guards
or error handling branches." to correct compound-modifier style.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 889e8623-678f-44b5-b988-87c097c5b2a4
📒 Files selected for processing (12)
extras/skills/codemark/SKILL.mdextras/skills/codemark/examples.mdextras/skills/codemark/queries/common.mdextras/skills/codemark/queries/csharp.mdextras/skills/codemark/queries/dart.mdextras/skills/codemark/queries/go.mdextras/skills/codemark/queries/java.mdextras/skills/codemark/queries/python.mdextras/skills/codemark/queries/rust.mdextras/skills/codemark/queries/swift.mdextras/skills/codemark/queries/typescript.mdextras/skills/codemark/templates.md
| codemark add-from-query \ | ||
| --file src/db/repo.rs \ | ||
| --query '(call_expression function: (member_expression property: (property_identifier) @method (#eq? @method "update_user_balance"))) @target' \ | ||
| --note "Critical: Database update point for user balances. Relationships: triggered after payment verification." \ |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Verify Rust node availability in official grammar metadata.
python - <<'PY'
import urllib.request
url = "https://raw.githubusercontent.com/tree-sitter/tree-sitter-rust/master/src/node-types.json"
s = urllib.request.urlopen(url).read().decode()
for token in ["member_expression", "property_identifier", "field_expression", "call_expression"]:
print(f"{token}: {token in s}")
PYRepository: DanielCardonaRojas/codemark
Length of output: 2652
🏁 Script executed:
# First, explore the markdown file to understand context and other examples
head -100 extras/skills/codemark/examples.mdRepository: DanielCardonaRojas/codemark
Length of output: 4405
🏁 Script executed:
# Search for other language examples to understand pattern
rg "codemark add-from-query" extras/skills/codemark/examples.md -A 5Repository: DanielCardonaRojas/codemark
Length of output: 3170
🌐 Web query:
Rust tree-sitter grammar node types member_expression property_identifier
💡 Result:
In the tree-sitter Rust grammar (https://github.com/tree-sitter/tree-sitter-rust), method calls like vec.len and field access like obj.field are parsed as field_expression nodes. A field_expression has two fields: receiver (the base expression or scoped_identifier) and field (a field_identifier node containing the property name as text). There are no member_expression or property_identifier node types in this grammar. field_identifier nodes are highlighted as @property in queries/highlights.scm. The full node types are in src/node-types.json (96.7KB file in the repo) and can be used to generate Rust enums via crates like tss-rust or type-sitter. In Rust tree-sitter bindings, access the property name via node.child_by_field_name("field").utf8_text(source_code).to_string.
Citations:
- 1: https://github.com/tree-sitter/tree-sitter-rust
- 2: https://github.com/tree-sitter/tree-sitter-rust/
- 3: https://github.com/tree-sitter/tree-sitter-rust/blob/master/queries/highlights.scm
- 4: https://docs.rs/tss-rust/latest/tree_sitter_symbols_rust/
- 5: https://github.com/Jakobeha/type-sitter
- 6: https://docs.rs/tree-sitter/latest/tree_sitter/struct.Node.html
- 7: https://registry.npmjs.org/tree-sitter-rust
Use Rust-compatible query syntax for the .rs file example.
The query uses JavaScript/TypeScript node types (member_expression, property_identifier) but targets a Rust file (src/db/repo.rs). In Rust's tree-sitter grammar, method calls use field_expression and field_identifier instead. This query will not match.
Suggested fix
- --query '(call_expression function: (member_expression property: (property_identifier) `@method` (`#eq`? `@method` "update_user_balance"))) `@target`' \
+ --query '(call_expression function: (field_expression field: (field_identifier) `@method` (`#eq`? `@method` "update_user_balance"))) `@target`' \📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| codemark add-from-query \ | |
| --file src/db/repo.rs \ | |
| --query '(call_expression function: (member_expression property: (property_identifier) @method (#eq? @method "update_user_balance"))) @target' \ | |
| --note "Critical: Database update point for user balances. Relationships: triggered after payment verification." \ | |
| codemark add-from-query \ | |
| --file src/db/repo.rs \ | |
| --query '(call_expression function: (field_expression field: (field_identifier) `@method` (`#eq`? `@method` "update_user_balance"))) `@target`' \ | |
| --note "Critical: Database update point for user balances. Relationships: triggered after payment verification." \ |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@extras/skills/codemark/examples.md` around lines 38 - 41, The example command
uses JS/TS tree-sitter node types (member_expression, property_identifier) which
won't match a Rust file; update the --query value in the codemark add-from-query
example to use Rust node types (e.g., field_expression and field_identifier) and
keep the capture name `@target` and the method identifier token (currently
"update_user_balance") so the command still targets the same call site; ensure
the query replaces member_expression→field_expression and
property_identifier→field_identifier (and adjust any predicate names if needed)
so it matches Rust method calls in src/db/repo.rs.
| (switch_statement | ||
| value: (identifier) @val | ||
| (#eq? @val "action")) @target | ||
| ``` |
There was a problem hiding this comment.
Use the correct Go switch node kind.
switch_statement is not the Go Tree-sitter node kind; this pattern should target expression_switch_statement (or type_switch_statement when needed).
🔧 Proposed fix
-(switch_statement
+(expression_switch_statement
value: (identifier) `@val`
(`#eq`? `@val` "action")) `@target`#!/bin/bash
# Verify switch node names in official tree-sitter-go metadata.
python - <<'PY'
import urllib.request
url = "https://raw.githubusercontent.com/tree-sitter/tree-sitter-go/master/src/node-types.json"
s = urllib.request.urlopen(url).read().decode()
for token in ["switch_statement", "expression_switch_statement", "type_switch_statement"]:
print(f"{token}: {token in s}")
PY🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@extras/skills/codemark/queries/go.md` around lines 112 - 115, The pattern
uses the wrong Go Tree-sitter node kind `switch_statement`; update the query to
target the correct Go node kinds such as `expression_switch_statement` (or
`type_switch_statement` when matching type switches) instead of
`switch_statement`, keeping the capture names `@val` and `@target` unchanged so
the rest of the query logic still binds the switch value and target capture.
| (switch_statement | ||
| condition: (parenthesized_expression | ||
| (identifier) @val | ||
| (#eq? @val "action"))) @target | ||
| ``` |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Verify Java switch node/field names in official grammar metadata.
python - <<'PY'
import urllib.request
url = "https://raw.githubusercontent.com/tree-sitter/tree-sitter-java/master/src/node-types.json"
s = urllib.request.urlopen(url).read().decode()
for token in ["switch_statement", "switch_expression", "\"value\""]:
print(f"{token}: {token in s}")
PYRepository: DanielCardonaRojas/codemark
Length of output: 2652
🏁 Script executed:
# First, check the actual file content
cat -n extras/skills/codemark/queries/java.md | sed -n '120,140p'Repository: DanielCardonaRojas/codemark
Length of output: 571
🏁 Script executed:
# Search for any tree-sitter Java grammar files in the repo
find . -name "*.json" -path "*/java*" -o -name "*grammar*" -type f | grep -i javaRepository: DanielCardonaRojas/codemark
Length of output: 53
🏁 Script executed:
# Look for other Java query examples or test files that might show correct usage
fd -t f "java" | head -20Repository: DanielCardonaRojas/codemark
Length of output: 146
🏁 Script executed:
# Check the Java test file for switch examples
cat tests/fixtures/java/AuthService.java | head -100Repository: DanielCardonaRojas/codemark
Length of output: 2884
🏁 Script executed:
# Search for any node-types or grammar references in the repo
find . -name "node-types.json" -o -name "*grammar*" | grep -i javaRepository: DanielCardonaRojas/codemark
Length of output: 53
🏁 Script executed:
# Search for any documentation or comments about tree-sitter Java in the codebase
rg -i "switch_statement|switch_expression" --type md --type txtRepository: DanielCardonaRojas/codemark
Length of output: 607
🏁 Script executed:
# Check if there's a package.json or similar that shows tree-sitter-java version
cat package.json 2>/dev/null || cat package-lock.json 2>/dev/null | head -50Repository: DanielCardonaRojas/codemark
Length of output: 53
🏁 Script executed:
# Look for tree-sitter config files or dependencies
rg -i "tree.sitter|grammar" --type json --type toml --type yaml | head -20Repository: DanielCardonaRojas/codemark
Length of output: 963
🏁 Script executed:
# Check if there's a .tree-sitter or similar directory
find . -type d -name "*tree*" -o -name "*grammar*"Repository: DanielCardonaRojas/codemark
Length of output: 53
🏁 Script executed:
# Search for any comments or documentation in the markdown files about node names
rg "node.*type|grammar|switch" extras/skills/codemark/queries/java.md -B 2 -A 2Repository: DanielCardonaRojas/codemark
Length of output: 162
🏁 Script executed:
# Look at how conditions are used in other query patterns (comparison check)
rg "condition:" extras/skills/codemark/queries/java.mdRepository: DanielCardonaRojas/codemark
Length of output: 148
🌐 Web query:
tree-sitter-java 0.23 switch_statement switch_expression grammar node types
💡 Result:
In tree-sitter-java version 0.23.x (latest v0.23.5 as of Dec 2024), the grammar supports both traditional and enhanced switch statements/expressions from Java 14+ (PR #78 merged Apr 2021) and Java 21 pattern matching (PR #158 merged Aug 2023). Key node types related to switch: - switch_statement: Traditional switch statement (switch (expr) { case ... : statements; }) - switch_expression: Enhanced switch used as expression (switch (expr) { case ... -> expr; }), can yield value - switch_construct: Internal unified construct used in both contexts (mentioned in PR #78 discussion) The grammar.js defines switch_block as either traditional switch_line groups or arrow_switch_line (case -> expr). PR #78 commit fb5038b explicitly renamed/used "switch_expression instead of switch_statement" for the expression form. Static node types are generated in src/node-types.json (79KB file, lists all node types with children/fields). To see exact switch node structure: 1. Clone https://github.com/tree-sitter/tree-sitter-java 2. Run tree-sitter generate 3. Check src/node-types.json for "switch_statement", "switch_expression" entries npm package includes src/node-types.json and grammar.json (182KB). No direct raw content found, but PR #78 confirms distinct nodes for statement vs expression forms, following Java spec separation while unifying parsing via switch_construct.
Citations:
- 1: Support enhanced switch statements tree-sitter/tree-sitter-java#78
- 2: https://github.com/tree-sitter/tree-sitter-java
- 3: feat(java 21): complete pattern matching for switch statements tree-sitter/tree-sitter-java#158
- 4: https://github.com/tree-sitter/tree-sitter-java/blob/master/src/node-types.json
- 5: Stricter declarations tree-sitter/tree-sitter-java#170
- 6: https://github.com/tree-sitter/tree-sitter-java/blob/a7db5227/queries/highlights.scm
- 7: https://github.com/tree-sitter/tree-sitter-java/tree/a7db5227ec40fcfe94489559d8c9bc7c8181e25a
- 8: https://registry.npmjs.org/tree-sitter-java
- 9: https://crates.io/crates/tree-sitter-java/0.23.5
🌐 Web query:
tree-sitter-java grammar "switch_expression" OR "switch_statement" node types
💡 Result:
In tree-sitter-java, the enhanced switch construct is represented by the node type "switch_construct" (it’s both an expression and a statement), rather than separate node types named "switch_expression" or "switch_statement". So, for your query: - switch_expression node type: not found as a distinct node type name in tree-sitter-java; the grammar uses switch_construct instead. - switch_statement node type: not found as a distinct node type name in tree-sitter-java; the grammar uses switch_construct instead. The related internal node names mentioned in the grammar update include "switch_block" and "switch_line" / "arrow_switch_line" (for the traditional and arrow forms of cases).
Citations:
- 1: Support enhanced switch statements tree-sitter/tree-sitter-java#78
- 2: https://tree-sitter.github.io/tree-sitter/using-parsers/6-static-node-types
- 3: https://github.com/tree-sitter/tree-sitter-java/blob/master/src/node-types.json
🌐 Web query:
tree-sitter-java src/node-types.json switch_construct field names
💡 Result:
The tree-sitter-java grammar defines a "switch_construct" node type for enhanced switch statements/expressions (introduced in Java 14+). This node appears in src/node-types.json, which is auto-generated from grammar.js via tree-sitter generate. No search results provided the exact JSON entry for "switch_construct" fields, as the file content is not directly quoted in snippets. However, Tree-sitter node types with fields are defined in grammar.js using field(name, rule) syntax, which populates the "fields" object in node-types.json with keys as field names. From PR #78 (merged), switch_construct is defined as both expression and statement, with structure: switch ( selector ) { switch_block } Where switch_block uses switch_line (traditional cases) or arrow_switch_line (case -> expr). Typical fields for similar constructs (e.g., method_declaration has "body", "name") suggest switch_construct likely has fields like: - "selector" or "discriminant": the expression after 'switch' - "body" or "cases" or "switch_block": the { ... } content Later PRs (#158 for Java 21 pattern matching) added "guard" nonterminal, potentially adding a "guard" field to case labels. To view exact current fields: 1. Clone https://github.com/tree-sitter/tree-sitter-java 2. Run tree-sitter generate 3. Check src/node-types.json for { "type": "switch_construct", ..., "fields": { ... } } Tree-sitter docs confirm fields enable named child access: ts_node_field_name_for_child(node, child_index).
Citations:
- 1: https://github.com/tree-sitter/tree-sitter-java/blob/master/src/node-types.json
- 2: https://tree-sitter.github.io/tree-sitter/using-parsers/6-static-node-types
- 3: Support enhanced switch statements tree-sitter/tree-sitter-java#78
- 4: feat(java 21): complete pattern matching for switch statements tree-sitter/tree-sitter-java#158
- 5: Documentation about tree node fields tree-sitter/tree-sitter#416
- 6: https://tree-sitter.github.io/tree-sitter/creating-parsers/3-writing-the-grammar.html
🌐 Web query:
site:github.com tree-sitter-java "switch_construct" field value OR condition
💡 Result:
In the tree-sitter-java grammar, "switch_construct" is the node type for Java switch statements and expressions (introduced for enhanced switch support in Java 14+). It serves as both an expression and statement. No sources explicitly document specific named fields like "value" or "condition" for this node, nor "OR condition". The grammar defines switch_construct with substructures like switch_block containing switch_line or arrow_switch_line, but detailed field names are in src/node-types.json (not directly accessible in results). Pattern matching updates (Java 21) added "guard" nonterminal for when-clauses, but no "value" or "condition" fields mentioned. Likely fields include selector (the switch expression), cases (list of switch labels/rules), and body (switch block), based on typical Tree-sitter Java grammar conventions from PR discussions, but unconfirmed explicitly.
Citations:
- 1: feat(java 21): complete pattern matching for switch statements tree-sitter/tree-sitter-java#158
- 2: Support enhanced switch statements tree-sitter/tree-sitter-java#78
- 3: https://github.com/tree-sitter/tree-sitter-java
- 4: https://github.com/tree-sitter/tree-sitter-java/blob/master/src/node-types.json
🌐 Web query:
site:github.com/tree-sitter/tree-sitter-java "switch_construct" query scm
💡 Result:
"switch_construct" is a non-terminal (grammar rule name) in the Tree-sitter Java parser grammar (defined in grammar.scm). It represents the enhanced switch statement/expression syntax introduced in Java 14+, supporting both traditional switch statements and switch expressions with arrow syntax (case L -> expr). It is used in both expression and statement contexts, as noted in development PRs for Tree-sitter Java grammar updates. "SCM" refers to the Scheme-based grammar file format used by Tree-sitter (grammar.scm). The "query scm" likely refers to Tree-sitter queries (in queries/*.scm files), but no direct query named or using "switch_construct" was found; highlights.scm exists but does not reference it explicitly. No connection to general Source Control Management (SCM) tools found.
Citations:
- 1: feat(java 21): complete pattern matching for switch statements tree-sitter/tree-sitter-java#158
- 2: Support enhanced switch statements tree-sitter/tree-sitter-java#78
- 3: Emma's switch changes tree-sitter/tree-sitter-java#79
- 4: https://linkedin.com/company/sinoprintingscm
- 5: https://github.com/tree-sitter/tree-sitter-java/blob/a7db5227/queries/highlights.scm
🌐 Web query:
site:github.com/tree-sitter/tree-sitter-java queries highlights "switch"
💡 Result:
The query "queries highlights 'switch'" refers to Tree-sitter syntax highlighting queries for the Java grammar, where "switch" is highlighted as a keyword. Tree-sitter uses .scm files like queries/highlights.scm to define syntax highlighting rules for code editors and tools. In the Java grammar, "switch" is explicitly listed under @keyword for highlighting.
Citations:
- 1: https://github.com/tree-sitter/tree-sitter-java/blob/a7db5227/queries/highlights.scm
- 2: feat(java 21): complete pattern matching for switch statements tree-sitter/tree-sitter-java#158
- 3: Support enhanced switch statements tree-sitter/tree-sitter-java#78
- 4: fix(highlights): missing highlights when accessing static methods and referencing methods tree-sitter/tree-sitter-java#95
🏁 Script executed:
# Check if there are any test cases or examples using switch in the codemark repo
rg "switch" tests/ -A 3 -B 3Repository: DanielCardonaRojas/codemark
Length of output: 53
🌐 Web query:
tree-sitter-java highlights.scm switch_construct selector expression
💡 Result:
In tree-sitter-java, the highlights.scm file defines syntax highlighting queries for Java code using Tree-sitter queries. The keyword "switch" is highlighted as @keyword, matching the literal token "switch" in switch statements and expressions. There is no specific query targeting "switch_construct" (the unified node for both switch statements and switch expressions, introduced in PR #78 for enhanced switch support), its "selector", or "expression". The selector (the expression after "switch (" and before ")") would typically be captured by general expression or identifier queries, such as (identifier) @variable, depending on its type. Switch cases ("case") are also @keyword. The highlights.scm prioritizes broad captures like keywords, identifiers, types, and functions, without specialized highlighting for switch substructures. For example: - Keywords: ["switch", "case", "default", ...] @keyword - Identifiers: (identifier) @variable To highlight switch selectors specifically, a custom query could be added like: (switch_construct selector: (expression) @variable.parameter) but none exists in the official queries/highlights.scm (as of commit a7db5227).
Citations:
- 1: https://github.com/tree-sitter/tree-sitter-java/blob/a7db5227/queries/highlights.scm
- 2: Support enhanced switch statements tree-sitter/tree-sitter-java#78
- 3: https://github.com/tree-sitter/tree-sitter-java
🌐 Web query:
"tree-sitter-java" "switch_construct" example pattern TreeSitter query
💡 Result:
A Tree-sitter query example pattern matching "switch_construct" in the tree-sitter-java grammar is: (switch_construct) @switch This captures any switch construct node, which represents enhanced switch expressions/statements in modern Java (introduced in Java 14+), as per the grammar updates in PRs like #78 and #79. For more specific matching, such as highlighting from queries/highlights.scm: "switch" @keyword To match the full structure: (switch_construct selector: (_) @condition body: (switch_block) @switch.body) Use the Tree-sitter playground (tree-sitter.github.io/playground or uber.github.io/piranha/tree-sitter-playground with Java selected) to test queries against Java switch code and visualize the "switch_construct" node.
Citations:
- 1: Support enhanced switch statements tree-sitter/tree-sitter-java#78
- 2: https://github.com/tree-sitter/tree-sitter-java/blob/a7db5227/queries/highlights.scm
- 3: https://github.com/tree-sitter/tree-sitter-java
- 4: https://tree-sitter.github.io/tree-sitter/cli/playground.html
- 5: https://uber.github.io/piranha/tree-sitter-playground/
Correct the Java switch query node type and field.
The pattern uses switch_statement with condition, which don't exist in tree-sitter-java v0.23. The correct node type is switch_construct with the selector field.
🔧 Proposed fix
-(switch_statement
- condition: (parenthesized_expression
- (identifier) `@val`
- (`#eq`? `@val` "action"))) `@target`
+(switch_construct
+ selector: (identifier) `@val`
+ (`#eq`? `@val` "action")) `@target`📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| (switch_statement | |
| condition: (parenthesized_expression | |
| (identifier) @val | |
| (#eq? @val "action"))) @target | |
| ``` | |
| (switch_construct | |
| selector: (identifier) `@val` | |
| (`#eq`? `@val` "action")) `@target` |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@extras/skills/codemark/queries/java.md` around lines 128 - 132, The
Tree-sitter Java query is using the wrong node type and field: replace uses of
switch_statement and its condition field with the correct node type
switch_construct and its selector field; update the pattern around
(switch_statement condition: (parenthesized_expression (identifier) `@val` (`#eq`?
`@val` "action"))) `@target` to reference switch_construct and selector so the query
matches tree-sitter-java v0.23 (preserve the `@val` and `@target` capture names).
1 participant
This PR updates the codemark skill and query guides to support targeting finer nodes like call expressions, conditionals, and branching logic. It also emphasizes the preference for specific bookmarks over general ones.
Summary by CodeRabbit
Release Notes