Merge pull request #953 from Extra-Chill/fix/951-workspace-cli-hang

Keep workspace help and targeted reads bounded

Author: chubes4

data-machine-code.php

@@ -21,6 +21,47 @@
 define( 'DATAMACHINE_CODE_PATH', plugin_dir_path( __FILE__ ) );
 define( 'DATAMACHINE_CODE_URL', plugin_dir_url( __FILE__ ) );
 
+/**
+ * Classify DMC CLI requests which must not initialize runtime services.
+ *
+ * WP-CLI loads WordPress before it dispatches nested help, so command methods
+ * cannot protect help from plugin bootstrap side effects. Inspect only DMC's
+ * own argv namespace here and leave every unrelated WP-CLI request unchanged.
+ *
+ * @param array<int,mixed>|null $argv Raw process arguments.
+ */
+function datamachine_code_is_side_effect_free_cli_request( ?array $argv = null ): bool {
+	if ( ! defined('WP_CLI') || ! WP_CLI ) {
+		return false;
+	}
+
+	$tokens = array_values(array_map('strval', $argv ?? ( is_array($GLOBALS['argv'] ?? null) ? $GLOBALS['argv'] : array() )));
+	if ( ! in_array('datamachine-code', $tokens, true) ) {
+		return false;
+	}
+
+	if ( in_array('--help', $tokens, true) || in_array('-h', $tokens, true) ) {
+		return true;
+	}
+
+	$help_index = array_search('help', $tokens, true);
+	$dmc_index  = array_search('datamachine-code', $tokens, true);
+	return false !== $help_index && false !== $dmc_index && $help_index < $dmc_index;
+}
+
+/**
+ * Whether this request is a targeted, read-only workspace command.
+ */
+function datamachine_code_is_targeted_workspace_read_cli_request( ?array $argv = null ): bool {
+	if ( ! defined('WP_CLI') || ! WP_CLI ) {
+		return false;
+	}
+
+	$tokens          = array_values(array_map('strval', $argv ?? ( is_array($GLOBALS['argv'] ?? null) ? $GLOBALS['argv'] : array() )));
+	$workspace_index = array_search('workspace', $tokens, true);
+	return false !== $workspace_index && 'show' === ( $tokens[ $workspace_index + 1 ] ?? '' ) && in_array('datamachine-code', array_slice($tokens, 0, $workspace_index), true);
+}
+
 // PSR-4 Autoloading.
 require_once __DIR__ . '/vendor/autoload.php';
 
@@ -41,7 +82,7 @@ function datamachine_code_install_schema(): void {
  * Keep schema current for already-active installs after deploy/update.
  */
 function datamachine_code_maybe_upgrade_schema(): void {
-	if ( function_exists('wp_installing') && wp_installing() ) {
+	if ( datamachine_code_is_side_effect_free_cli_request() || datamachine_code_is_targeted_workspace_read_cli_request() || ( function_exists('wp_installing') && wp_installing() ) ) {
 		return;
 	}
 
@@ -75,7 +116,7 @@ function datamachine_code_has_datamachine_integration(): bool {
 function datamachine_code_register_bundle_artifacts(): void {
 	static $registered = false;
 
-	if ( $registered ) {
+	if ( $registered || datamachine_code_is_side_effect_free_cli_request() ) {
 		return;
 	}
 
@@ -90,7 +131,7 @@ function datamachine_code_register_bundle_artifacts(): void {
 function datamachine_code_register_datamachine_integrations(): void {
 	static $registered = false;
 
-	if ( $registered || ! datamachine_code_has_datamachine_integration() ) {
+	if ( $registered || datamachine_code_is_side_effect_free_cli_request() || ! datamachine_code_has_datamachine_integration() ) {
 		return;
 	}
 
@@ -117,7 +158,7 @@ function datamachine_code_register_datamachine_integrations(): void {
 function datamachine_code_bootstrap() {
 	static $bootstrapped = false;
 
-	if ( $bootstrapped ) {
+	if ( $bootstrapped || datamachine_code_is_side_effect_free_cli_request() || datamachine_code_is_targeted_workspace_read_cli_request() ) {
 		return;
 	}
 
@@ -321,7 +362,7 @@ function datamachine_code_register_cli_commands() {
  * Only load when Data Machine core's AI engine is available.
  */
 function datamachine_code_load_chat_tools() {
-	if ( ! class_exists('DataMachine\Engine\AI\Tools\BaseTool') ) {
+	if ( datamachine_code_is_side_effect_free_cli_request() || datamachine_code_is_targeted_workspace_read_cli_request() || ! class_exists('DataMachine\Engine\AI\Tools\BaseTool') ) {
 		return;
 	}
 

inc/Abilities/WorkspaceAbilities.php

@@ -2888,21 +2888,22 @@ public static function getCapabilities( array $input ): array {   // phpcs:ignor
 	 */
 	public static function showRepo( array $input ): array|\WP_Error {
 		$workspace = new Workspace();
-		if ( RemoteWorkspaceBackend::should_handle() ) {
-			$local_result = self::showLocalWorkspaceHandleIfPresent($workspace, (string) ( $input['name'] ?? '' ));
-			if ( null !== $local_result ) {
-				return $local_result;
-			}
+		$handle    = (string) ( $input['name'] ?? '' );
+		$local     = $workspace->show_repo($handle);
+		if ( ! is_wp_error($local) && empty($local['is_context']) ) {
+			return $local;
 		}
 
+		// Registered remote state remains authoritative for local misses, bounded
+		// local failures, and context aliases whose checkout is not mounted.
 		if ( RemoteWorkspaceBackend::should_handle() ) {
-			$result = ( new RemoteWorkspaceBackend() )->show($input['name'] ?? '');
+			$result = ( new RemoteWorkspaceBackend() )->show($handle);
 			if ( ! self::shouldFallbackToLocalWorkspace($result) ) {
 				return $result;
 			}
 		}
 
-		return $workspace->show_repo($input['name'] ?? '');
+		return $local;
 	}
 
 	/**

inc/Cli/Commands/WorkspaceCommand.php

@@ -17,6 +17,7 @@
 
 use WP_CLI;
 use DataMachine\Cli\BaseCommand;
+use DataMachineCode\Abilities\WorkspaceAbilities;
 use DataMachineCode\Cli\CliResponseRenderer;
 use DataMachineCode\Cli\CliRepeatableOptionParser;
 use DataMachineCode\Cli\WorkspaceCompactOutput;
@@ -2510,13 +2511,9 @@ public function show( array $args, array $assoc_args ): void {
 			return;
 		}
 
-		$ability = wp_get_ability('datamachine-code/workspace-show');
-		if ( ! $ability ) {
-			WP_CLI::error('Workspace show ability not available.');
-			return;
-		}
-
-		$result = $ability->execute(array( 'name' => $args[0] ));
+		// This targeted read is also the lightweight startup path used before the
+		// Abilities API runtime has been bootstrapped.
+		$result = WorkspaceAbilities::showRepo(array( 'name' => $args[0] ));
 
 		if ( is_wp_error($result) ) {
 			WP_CLI::error($result->get_error_message());

inc/Workspace/WorkspaceCoreUtilities.php

@@ -692,7 +692,17 @@ private function build_primary_freshness_report( string $repo_path, string $hand
 			);
 		}
 
-		$header   = strtok( (string) ( $status_result['output'] ?? '' ), "\n");
+		return $this->build_primary_freshness_report_from_status_output((string) ( $status_result['output'] ?? '' ), $handle);
+	}
+
+	/**
+	 * Build primary freshness metadata from an already-bounded local status probe.
+	 *
+	 * @return array<string,mixed>
+	 */
+	private function build_primary_freshness_report_from_status_output( string $status_output, string $handle ): array {
+
+		$header   = strtok($status_output, "\n");
 		$header   = false === $header ? '' : trim($header);
 		$branch   = null;
 		$detached = false;

inc/Workspace/WorkspaceRepositoryLifecycle.php

@@ -712,13 +712,28 @@ public function remove_repo( string $handle ): array|\WP_Error {
 	 * @return array{success: bool, name?: string, path?: string, branch?: string, remote?: string, commit?: string, dirty?: int}|\WP_Error
 	 */
 	public function show_repo( string $handle ): array|\WP_Error {
-		$context_policy = WorkspaceAliasResolver::context_policy_for($handle);
+		$requested_handle = $handle;
+		$context_policy   = null;
+		$parsed           = $this->parse_handle($handle);
+		$repo_path        = $this->workspace_path . '/' . $parsed['dir_name'];
+		$inspection       = WorkspaceTargetInspector::inspect($repo_path, $parsed['dir_name']);
+		if ( is_wp_error($inspection) ) {
+			return $inspection;
+		}
+
+		if ( empty($inspection['exists']) ) {
+			$context_policy = WorkspaceAliasResolver::context_policy_for($handle);
+		}
 		if ( null !== $context_policy ) {
 			$target    = (string) ( $context_policy['target'] ?? $handle );
 			$parsed    = $this->parse_handle($target);
 			$repo_path = $this->workspace_path . '/' . $parsed['dir_name'];
 			$ref       = (string) ( $context_policy['ref'] ?? '' );
-			if ( ! is_dir($repo_path) ) {
+			$inspection = WorkspaceTargetInspector::inspect($repo_path, $handle);
+			if ( is_wp_error($inspection) ) {
+				return $inspection;
+			}
+			if ( empty($inspection['exists']) ) {
 				return array(
 					'success'          => true,
 					'name'             => (string) $context_policy['alias'],
@@ -736,35 +751,37 @@ public function show_repo( string $handle ): array|\WP_Error {
 			$handle = $target;
 		}
 
-		$resolved_handle = $this->resolve_primary_repo_name($handle);
-		if ( ! is_wp_error($resolved_handle) ) {
-			$handle = $resolved_handle;
+		if ( empty($inspection['exists']) && null === $context_policy ) {
+			$resolved_handle = $this->resolve_primary_repo_name($handle);
+			if ( ! is_wp_error($resolved_handle) && $resolved_handle !== $handle ) {
+				$handle     = $resolved_handle;
+				$parsed     = $this->parse_handle($handle);
+				$repo_path  = $this->workspace_path . '/' . $parsed['dir_name'];
+				$inspection = WorkspaceTargetInspector::inspect($repo_path, $parsed['dir_name']);
+				if ( is_wp_error($inspection) ) {
+					return $inspection;
+				}
+			}
 		}
 
-		$parsed    = $this->parse_handle($handle);
-		$repo_path = $this->workspace_path . '/' . $parsed['dir_name'];
-
-		if ( ! is_dir($repo_path) ) {
-			return new \WP_Error('repo_not_found', sprintf('Workspace handle "%s" not found.', $parsed['dir_name']), array( 'status' => 404 ));
+		if ( empty($inspection['exists']) ) {
+			return new \WP_Error('repo_not_found', sprintf('Workspace handle "%s" not found.', $requested_handle), array( 'status' => 404 ));
 		}
 
-		$branch = GitRunner::current_branch($repo_path);
-		$remote = GitRunner::remote_url($repo_path);
-		$commit = GitRunner::latest_commit_summary($repo_path);
-		$status = GitRunner::dirty_count($repo_path);
-
 		$result = array(
 			'success'           => true,
 			'name'              => null !== $context_policy ? (string) $context_policy['alias'] : $parsed['dir_name'],
 			'repo'              => $parsed['repo'],
 			'is_worktree'       => $parsed['is_worktree'],
 			'is_context'        => null !== $context_policy,
 			'path'              => $repo_path,
-			'branch'            => $branch,
-			'remote'            => $remote,
-			'commit'            => $commit,
-			'dirty'             => $status,
-			'primary_freshness' => ! $parsed['is_worktree'] ? $this->build_primary_freshness_report($repo_path, $parsed['dir_name']) : null,
+			'branch'            => $inspection['branch'] ?? null,
+			'remote'            => $inspection['remote'] ?? null,
+			'commit'            => $inspection['commit'] ?? null,
+			'dirty'             => (int) ( $inspection['dirty'] ?? 0 ),
+			'primary_freshness' => ! $parsed['is_worktree'] && is_string($inspection['branch_status'] ?? null)
+				? $this->build_primary_freshness_report_from_status_output((string) $inspection['branch_status'], $parsed['dir_name'])
+				: null,
 		);
 		if ( null !== $context_policy ) {
 			$result['workspace_policy'] = WorkspaceAliasResolver::policy_attestation($handle);

inc/Workspace/WorkspaceTargetInspector.php

@@ -0,0 +1,95 @@
+<?php
+/**
+ * Bounded, read-only inspection of one workspace target.
+ *
+ * @package DataMachineCode\Workspace
+ */
+
+namespace DataMachineCode\Workspace;
+
+use DataMachineCode\Support\CommandSpec;
+use DataMachineCode\Support\ProcessRunner;
+
+defined('ABSPATH') || exit;
+
+final class WorkspaceTargetInspector {
+
+	private const DEFAULT_TIMEOUT_SECONDS = 5;
+
+	/**
+	 * Inspect exactly one path in a killable child process.
+	 *
+	 * @return array<string,mixed>|\WP_Error
+	 */
+	public static function inspect( string $path, string $handle ): array|\WP_Error {
+		$timeout = function_exists('apply_filters')
+			? (int) apply_filters('datamachine_code_workspace_target_lookup_timeout_seconds', self::DEFAULT_TIMEOUT_SECONDS, $handle)
+			: self::DEFAULT_TIMEOUT_SECONDS;
+		$timeout = max(1, $timeout);
+
+		$diagnostic = array(
+			'phase'           => 'filesystem',
+			'resource'        => $path,
+			'probe_owner'     => 'workspace-show',
+			'timeout_seconds' => $timeout,
+		);
+		if ( function_exists('do_action') ) {
+			do_action('datamachine_code_workspace_lookup_waiting', $diagnostic);
+		}
+
+		$probe_path       = __DIR__ . '/workspace-target-probe.php';
+		$filesystem_probe = '';
+		$git_command      = 'git';
+		if ( function_exists('apply_filters') ) {
+			$filesystem_probe = (string) apply_filters('datamachine_code_workspace_target_filesystem_probe_command', '', $handle, $path);
+			$git_command      = (string) apply_filters('datamachine_code_workspace_target_git_command', 'git', $handle, $path);
+		}
+		$command = CommandSpec::from_argv(array( PHP_BINARY, $probe_path, $path, $filesystem_probe, $git_command ));
+		if ( is_wp_error($command) ) {
+			return $command;
+		}
+
+		$result = ProcessRunner::run(
+			$command,
+			array(
+				'timeout_seconds' => $timeout,
+				'separate_streams' => true,
+				'output_cap_bytes' => 32768,
+				'error_code'       => 'workspace_target_lookup_failed',
+			)
+		);
+		if ( is_wp_error($result) ) {
+			$data   = is_array($result->get_error_data()) ? $result->get_error_data() : array();
+			$stderr = (string) ( $data['stderr'] ?? '' );
+			if ( preg_match_all('/^DMC_BOUNDARY:([a-z_]+):([^\r\n]+)/m', $stderr, $matches) && ! empty($matches[1]) ) {
+				$diagnostic['phase']     = (string) end($matches[1]);
+				$diagnostic['operation'] = (string) end($matches[2]);
+			}
+
+			if ( isset($data['timeout']) ) {
+				return new \WP_Error(
+					'workspace_target_lookup_timeout',
+					sprintf('Workspace lookup for "%s" timed out during the %s probe after %d second(s).', $handle, $diagnostic['phase'], $timeout),
+					array_merge($data, $diagnostic, array( 'status' => 504 ))
+				);
+			}
+
+			return new \WP_Error(
+				'workspace_target_lookup_failed',
+				sprintf('Workspace lookup for "%s" failed during the %s probe.', $handle, $diagnostic['phase']),
+				array_merge($data, $diagnostic, array( 'status' => 500 ))
+			);
+		}
+
+		$decoded = json_decode((string) ( $result['stdout'] ?? '' ), true);
+		if ( ! is_array($decoded) ) {
+			return new \WP_Error(
+				'workspace_target_lookup_invalid_response',
+				sprintf('Workspace lookup for "%s" returned an invalid probe response.', $handle),
+				array_merge($diagnostic, array( 'status' => 500 ))
+			);
+		}
+
+		return $decoded;
+	}
+}