Make event emission match between cli and service build paths

databuild/graph/analyze.rs

@@ -618,6 +618,44 @@ async fn main() {
         None
     };
     
+    // Check if this is a CLI build (not service-initiated)
+    let is_cli_mode = env::var("DATABUILD_CLI_MODE").is_ok();
+    
+    // Emit orchestration events for CLI mode to match Service behavior
+    if is_cli_mode {
+        if let Some(ref event_log) = build_event_log {
+            // Emit "Build request received" event
+            let event = create_build_event(
+                build_request_id.clone(),
+                crate::build_event::EventType::BuildRequestEvent(BuildRequestEvent {
+                    status: BuildRequestStatus::BuildRequestReceived as i32,
+                    requested_partitions: args.iter()
+                        .map(|p| PartitionRef { str: p.clone() })
+                        .collect(),
+                    message: "Build request received".to_string(),
+                })
+            );
+            if let Err(e) = event_log.append_event(event).await {
+                error!("Failed to log build request received event: {}", e);
+            }
+            
+            // Emit "Starting build planning" event
+            let event = create_build_event(
+                build_request_id.clone(),
+                crate::build_event::EventType::BuildRequestEvent(BuildRequestEvent {
+                    status: BuildRequestStatus::BuildRequestPlanning as i32,
+                    requested_partitions: args.iter()
+                        .map(|p| PartitionRef { str: p.clone() })
+                        .collect(),
+                    message: "Starting build planning".to_string(),
+                })
+            );
+            if let Err(e) = event_log.append_event(event).await {
+                error!("Failed to log build planning event: {}", e);
+            }
+        }
+    }
+    
     match mode.as_str() {
         "plan" => {
             // Get output refs from command line arguments

databuild/graph/execute.rs

@@ -367,7 +367,28 @@ async fn main() -> Result<(), Box<dyn std::error::Error>> {
 
     info!("Executing job graph with {} nodes", graph.nodes.len());
 
-    // Log build request execution start
+    // Check if this is a CLI build (not service-initiated)
+    let is_cli_mode = std::env::var("DATABUILD_CLI_MODE").is_ok();
+
+    // Emit orchestration events for CLI mode to match Service behavior
+    if is_cli_mode {
+        if let Some(ref event_log) = build_event_log {
+            // Emit "Starting build execution" event (matches Service pattern)
+            let event = create_build_event(
+                build_request_id.clone(),
+                EventType::BuildRequestEvent(BuildRequestEvent {
+                    status: BuildRequestStatus::BuildRequestExecuting as i32,
+                    requested_partitions: graph.outputs.clone(),
+                    message: "Starting build execution".to_string(),
+                })
+            );
+            if let Err(e) = event_log.append_event(event).await {
+                error!("Failed to log build execution event: {}", e);
+            }
+        }
+    }
+
+    // Log build request execution start (existing detailed event)
     if let Some(ref event_log) = build_event_log {
         let event = create_build_event(
             build_request_id.clone(),
@@ -648,7 +669,36 @@ async fn main() -> Result<(), Box<dyn std::error::Error>> {
 
     info!("Execution complete: {} succeeded, {} failed", success_count, failure_count);
 
-    // Log final build request status
+    // Emit orchestration completion events for CLI mode to match Service behavior
+    if is_cli_mode {
+        if let Some(ref event_log) = build_event_log {
+            let orchestration_message = if failure_count > 0 || fail_fast_triggered {
+                "Build request failed"
+            } else {
+                "Build request completed successfully"
+            };
+            
+            let orchestration_status = if failure_count > 0 || fail_fast_triggered {
+                BuildRequestStatus::BuildRequestFailed
+            } else {
+                BuildRequestStatus::BuildRequestCompleted
+            };
+            
+            let event = create_build_event(
+                build_request_id.clone(),
+                EventType::BuildRequestEvent(BuildRequestEvent {
+                    status: orchestration_status as i32,
+                    requested_partitions: graph.outputs.clone(),
+                    message: orchestration_message.to_string(),
+                })
+            );
+            if let Err(e) = event_log.append_event(event).await {
+                error!("Failed to log build request completion event: {}", e);
+            }
+        }
+    }
+
+    // Log final build request status (existing detailed event)
     if let Some(ref event_log) = build_event_log {
         let final_status = if failure_count > 0 || fail_fast_triggered {
             BuildRequestStatus::BuildRequestFailed

databuild/rules.bzl

@@ -635,7 +635,10 @@ def _databuild_graph_build_impl(ctx):
         output = script,
         is_executable = True,
         content = "#!/bin/bash\n\n" + RUNFILES_PREFIX + """
-# TODO these rlocation commands resolve to nothing - does it expect to be
+# Set CLI mode to indicate this is a direct CLI build (not service-initiated)
+export DATABUILD_CLI_MODE=true
+
+# Run analysis and execution phases
 $(rlocation _main/{analyze_path}) $@ | $(rlocation _main/{exec_path})
         """.format(
             analyze_path = ctx.attr.analyze.files_to_run.executable.short_path,

tests/end_to_end/lib/db_utils.sh

@@ -69,6 +69,46 @@ compare_build_events() {
     fi
 }
 
+# Compare core build events (job and partition only, ignoring request orchestration)
+compare_core_build_events() {
+    local db1="$1"
+    local db2="$2"
+    
+    log_info "Comparing core build events (job and partition) between databases" >&2
+    
+    # Count core events by type
+    local db1_job=$(count_build_events "$db1" "job")
+    local db1_partition=$(count_build_events "$db1" "partition")
+    local db2_job=$(count_build_events "$db2" "job") 
+    local db2_partition=$(count_build_events "$db2" "partition")
+    
+    log_info "Database 1: $db1_job job events, $db1_partition partition events" >&2
+    log_info "Database 2: $db2_job job events, $db2_partition partition events" >&2
+    
+    # Check if core events match
+    if [[ "$db1_job" -eq "$db2_job" ]] && [[ "$db1_partition" -eq "$db2_partition" ]]; then
+        log_info "Core build events match" >&2
+        return 0
+    else
+        log_error "Core build events differ" >&2
+        return 1
+    fi
+}
+
+# Get detailed event breakdown
+get_event_breakdown() {
+    local db_path="$1"
+    
+    if [[ ! -f "$db_path" ]]; then
+        echo "Database not found: $db_path" >&2
+        return 1
+    fi
+    
+    echo "Event breakdown for $db_path:" >&2
+    sqlite3 "$db_path" \
+        "SELECT event_type, COUNT(*) FROM build_events GROUP BY event_type ORDER BY event_type;" >&2
+}
+
 # Count events in database
 count_build_events() {
     local db_path="$1"

tests/end_to_end/simple_test.sh

@@ -117,10 +117,35 @@ else
     exit 1
 fi
 
-# Compare event counts (should be similar)
+# Compare event counts with detailed analysis
 if [[ "$CLI_EVENTS" -gt 0 ]] && [[ "$SERVICE_EVENTS" -gt 0 ]]; then
     echo "[INFO] Both CLI and Service generated events successfully"
     echo "[INFO] CLI: $CLI_EVENTS events, Service: $SERVICE_EVENTS events"
+    
+    # Detailed event comparison
+    echo "[INFO] Analyzing event distribution..."
+    
+    CLI_JOB_EVENTS=$(sqlite3 /tmp/simple_test_cli.db "SELECT COUNT(*) FROM build_events WHERE event_type = 'job';" 2>/dev/null || echo "0")
+    CLI_PARTITION_EVENTS=$(sqlite3 /tmp/simple_test_cli.db "SELECT COUNT(*) FROM build_events WHERE event_type = 'partition';" 2>/dev/null || echo "0")
+    CLI_REQUEST_EVENTS=$(sqlite3 /tmp/simple_test_cli.db "SELECT COUNT(*) FROM build_events WHERE event_type = 'build_request';" 2>/dev/null || echo "0")
+    
+    SERVICE_JOB_EVENTS=$(sqlite3 "$SERVICE_DB_PATH" "SELECT COUNT(*) FROM build_events WHERE event_type = 'job';" 2>/dev/null || echo "0")
+    SERVICE_PARTITION_EVENTS=$(sqlite3 "$SERVICE_DB_PATH" "SELECT COUNT(*) FROM build_events WHERE event_type = 'partition';" 2>/dev/null || echo "0")
+    SERVICE_REQUEST_EVENTS=$(sqlite3 "$SERVICE_DB_PATH" "SELECT COUNT(*) FROM build_events WHERE event_type = 'build_request';" 2>/dev/null || echo "0")
+    
+    echo "[INFO] Event breakdown:"
+    echo "[INFO]   Job events:       CLI=$CLI_JOB_EVENTS, Service=$SERVICE_JOB_EVENTS"
+    echo "[INFO]   Partition events: CLI=$CLI_PARTITION_EVENTS, Service=$SERVICE_PARTITION_EVENTS"
+    echo "[INFO]   Request events:   CLI=$CLI_REQUEST_EVENTS, Service=$SERVICE_REQUEST_EVENTS"
+    
+    # Validate core events are identical
+    if [[ "$CLI_JOB_EVENTS" -eq "$SERVICE_JOB_EVENTS" ]] && [[ "$CLI_PARTITION_EVENTS" -eq "$SERVICE_PARTITION_EVENTS" ]] && [[ "$CLI_REQUEST_EVENTS" -eq "$SERVICE_REQUEST_EVENTS" ]]; then
+        echo "[INFO] ✅ Core build events (job and partition) are identical"
+    else
+        echo "[ERROR] ❌ Core build events differ - this indicates a problem"
+        exit 1
+    fi
+
 else
     echo "[ERROR] One or both builds generated no events"
     exit 1