Add command for running a query suite

[nickrolfe]

The command is visible in the context menu for a .qls file in the Explorer pane, only when canary mode is enabled.

This uses a new query-server command for running multiple queries, so that a single evaluator log will be produced for the entire run. The intended use case is an internal QL author who wants to judge performance by viewing the evaluator log for the entire query suite.

It would have been much simpler to use the existing functionality for running multiple queries, but that functionality does not work very well when one wants to inspect the evaluator logs – since each query gets started in turn, which truncates the log for the previous query and starts a new one.

To avoid too much code duplication, I have updated a lot of the code paths involved in running local queries to work with multiple query paths. This also required some refactoring to explicitly associate an output basename (used to produce the .bqrs, .csv, etc. paths) with each input query, where before those output filenames were hard-coded. As a result, the diff is fairly big (but the changes should be easily understandable).

Since this is a canary feature intended for internal users, and in the interest of shipping it quickly, there are a few rough edges, particularly with the UI for the query history pane:

• There's one entry while the suite is running, and then, when it completes, we add entries for each query in the suite.
• Since all those entries point to the same results directory, deleting an entry in the query history pane will now also delete any others that have the same results directory.
• The summary text for each additional entry in the query history pane will not include the number of results – this number is normally computed when showing the results window, as we decode the BQRS. It would be too slow and visually disturbing to do this for every query in the suite.

In the long term, it would be ideal to have a first-class concept of a multi-query run in the query history pane, which would allow us to create hierarchical entries, where run-specific commands like "Show evaluator log summary" would appear on the root node for the run, while query-specific commands like "Show CSV" would appear on the child nodes for each query. But I think what's implemented here is good enough for a canary/internal feature.

Also, if the CLI is too old to support the new evaluation/runQueries query-server command, the run will fail with an error message explaining this. I decided not to implement a fallback to running the queries one by one.

This is a canary feature, so I won't update the changelog.

[Copilot]

Pull Request Overview

This PR introduces a new “Run Query Suite” canary feature for .qls files and refactors local query execution to support multiple queries in one run.

• Adds codeQL.runQuerySuite command, context menu entries, and CLI feature guard
• Refactors query-run flows to use queryInputsOutputs lists and per-query outputBaseName
• Updates QueryOutputDir to generate .bqrs, .csv, and interpreted paths dynamically

Reviewed Changes

Copilot reviewed 42 out of 42 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
suggestion-queries.ts	Decode BQRS with per-query basename
model-editor-queries.ts	Read model-editor results via per-query basename
generate.ts	Decode generate query with per-query basename
run-query.ts	Wrap single-run in queryInputsOutputs; lookup by path
results-view.ts	Pass separate resultsPath and interpretedResultsPath
query-output-dir.ts	Replace fixed paths with getBqrsPath/getCsvPath/getIntereptedResultsPath
local-query-run.ts	Adapt history and eval-log summary to multi-query runs
local-queries.ts	Implement runQuerySuite command and multi-query runner
template-provider.ts	Use per-query BQRS path
query-resolver.ts	Single-query runner updated to use queryInputsOutputs
location-finder.ts	Fetch links from multi-query BQRS via dynamic path
ast-builder.ts	Consume BQRS path directly
debugger-ui.ts & debug-session.ts	Adapt debugger to per-query result structures
debug-protocol.ts	Add outputBaseName to protocol
interpreted-results.ts & compare-view.ts	Reference new resultsPath properties
common/commands.ts	Register codeQL.runQuerySuite command
cli-version.ts	Expose queryServerRunQueries feature flag
package.json	Add command and menu entries for runQuerySuite

[Copilot]

The method name getIntereptedResultsPath is misspelled; it should be getInterpretedResultsPath.

Suggested change

	getIntereptedResultsPath(
	getInterpretedResultsPath(

[Copilot]

[nitpick] This pattern for extracting and validating a single query result is repeated in multiple locations; consider extracting it into a shared helper to reduce duplication.

Suggested change

	const queryResults = Array.from(completedQuery.results.values());
	if (queryResults.length !== 1) {
	throw new Error(
	`Expected exactly one query result, but got ${queryResults.length}`,
	);
	}
	const bqrs = await cliServer.bqrsDecodeAll(
	completedQuery.outputDir.getBqrsPath(queryResults[0].outputBaseName),
	const bqrs = await extractAndDecodeSingleQueryResult(
	completedQuery.results,
	cliServer,
	completedQuery.outputDir,

[Copilot]

[nitpick] Hard-coding the outputBaseName string "results" in multiple places could be error-prone; consider defining a constant for the default basename.

Suggested change

	queryInputsOutputs: [{ queryPath, outputBaseName: "results" }],
	queryInputsOutputs: [{ queryPath, outputBaseName: DEFAULT_OUTPUT_BASENAME }],