Avoid using a (possibly invalid) EventContext pointer - attempt 2

First version (!719 (merged)) had a bug, but it only became apparent after it was merged and used in Athena. The merge deleted the source branch, so I'm starting again here.

Motivation remains the same: fixing an intermittent issue documented here: https://its.cern.ch/jira/browse/ATEAM-471

An EventContext pointer had the potential to be used after deletion. Since the pointer wasn't really necessary I've refactored the updateStates() method to avoid it.

GaudiHive/src/AvalancheSchedulerSvc.cpp

@@ -581,7 +581,8 @@ StatusCode AvalancheSchedulerSvc::tryPopFinishedEvent( EventContext*& eventConte
  * * No algorithms have been signed off by the data flow
  * * No algorithms have been scheduled
 */
-StatusCode AvalancheSchedulerSvc::updateStates( int si, const int algo_index, EventContext* inputContext )
+StatusCode AvalancheSchedulerSvc::updateStates( int si, const int algo_index, const int sub_slot,
+                                                const int source_slot )
 {
 
   StatusCode global_sc( StatusCode::SUCCESS );
@@ -614,16 +615,14 @@ StatusCode AvalancheSchedulerSvc::updateStates( int si, const int algo_index, Ev
     if ( algo_index >= 0 ) {
       Cause cs = {Cause::source::Task, index2algname( algo_index )};
 
-      // Pass sub-slots to precedence service if necessary
-      if ( !inputContext || iSlot != (int)inputContext->slot() || inputContext == thisSlot.eventContext ) {
+      // Run in whole-event context if there's no sub-slot index, or the sub-slot has a different parent
+      if ( sub_slot == -1 || iSlot != source_slot ) {
         if ( m_precSvc->iterate( thisSlot, cs ).isFailure() ) {
           error() << "Failed to call IPrecedenceSvc::iterate for slot " << iSlot << endmsg;
           global_sc = StatusCode::FAILURE;
         }
       } else {
-        // An input context that doesn't match the event context for that slot number implies a sub-slot
-        unsigned int const subSlotIndex = thisSlot.contextToSlot.at( inputContext );
-        if ( m_precSvc->iterate( thisSlot.allSubSlots[subSlotIndex], cs ).isFailure() ) {
+        if ( m_precSvc->iterate( thisSlot.allSubSlots[sub_slot], cs ).isFailure() ) {
           error() << "Failed to call IPrecedenceSvc::iterate for sub-slot of " << iSlot << endmsg;
           global_sc = StatusCode::FAILURE;
         }
@@ -1033,13 +1032,14 @@ StatusCode AvalancheSchedulerSvc::promoteToExecuted( unsigned int iAlgo, int si,
                      : AState::ERROR;
 
   // Update states in the appropriate slot
+  int subSlotIndex = -1;
   if ( eventContext == thisSlot.eventContext ) {
     // Event level (standard behaviour)
     sc = thisSlot.algsStates.set( iAlgo, state );
   } else {
     // Sub-slot
-    unsigned int const subSlotIndex = thisSlot.contextToSlot.at( eventContext );
-    sc                              = thisSlot.allSubSlots[subSlotIndex].algsStates.set( iAlgo, state );
+    subSlotIndex = static_cast<int>( thisSlot.contextToSlot.at( eventContext ) );
+    sc           = thisSlot.allSubSlots[subSlotIndex].algsStates.set( iAlgo, state );
   }
 
   ON_VERBOSE if ( sc.isSuccess() ) verbose() << "Promoting " << algo->name() << " on slot " << si << " to " << state
@@ -1050,9 +1050,9 @@ StatusCode AvalancheSchedulerSvc::promoteToExecuted( unsigned int iAlgo, int si,
 
   // Schedule an update of the status of the algorithms
   ++m_actionsCounts[si];
-  m_actionsQueue.push( [this, si, iAlgo, eventContext]() {
+  m_actionsQueue.push( [this, si, iAlgo, subSlotIndex]() {
     --this->m_actionsCounts[si]; // no bound check needed as decrements/increments are balanced in the current setup
-    return this->updateStates( -1, iAlgo, eventContext );
+    return this->updateStates( -1, iAlgo, subSlotIndex, si );
   } );
 
   return sc;
@@ -1088,13 +1088,14 @@ StatusCode AvalancheSchedulerSvc::promoteToAsyncExecuted( unsigned int iAlgo, in
                      : AState::ERROR;
 
   // Update states in the appropriate slot
+  int subSlotIndex = -1;
   if ( eventContext == thisSlot.eventContext ) {
     // Event level (standard behaviour)
     sc = thisSlot.algsStates.set( iAlgo, state );
   } else {
     // Sub-slot
-    unsigned int const subSlotIndex = thisSlot.contextToSlot.at( eventContext );
-    sc                              = thisSlot.allSubSlots[subSlotIndex].algsStates.set( iAlgo, state );
+    subSlotIndex = thisSlot.contextToSlot.at( eventContext );
+    sc           = thisSlot.allSubSlots[subSlotIndex].algsStates.set( iAlgo, state );
   }
 
   ON_VERBOSE if ( sc.isSuccess() ) verbose() << "[Asynchronous] Promoting " << algo->name() << " on slot " << si
@@ -1105,9 +1106,9 @@ StatusCode AvalancheSchedulerSvc::promoteToAsyncExecuted( unsigned int iAlgo, in
 
   // Schedule an update of the status of the algorithms
   ++m_actionsCounts[si];
-  m_actionsQueue.push( [this, si, iAlgo, eventContext]() {
+  m_actionsQueue.push( [this, si, iAlgo, subSlotIndex]() {
     --this->m_actionsCounts[si]; // no bound check needed as decrements/increments are balanced in the current setup
-    return this->updateStates( -1, iAlgo, eventContext );
+    return this->updateStates( -1, iAlgo, subSlotIndex, si );
   } );
 
   return sc;