Add new parallel HLTBeginSeq for L1Decoder and CacheCreators, add HLTEndSeq for summary algorithms

Also put L1Decoder into the sequence at the correct point already in runHLT_standalone.py. Addresses ATR-21657, and in general means we can run parallel jobs more stably.

Trigger/TrigValidation/TrigUpgradeTest/CMakeLists.txt

@@ -12,11 +12,13 @@ atlas_add_component( TrigUpgradeTest
 atlas_add_test( emu_step_processing
 		SCRIPT test/test_emu_step_processing.sh
 		LOG_SELECT_PATTERN "TrigSignatureMoniMT.*INFO HLT_.*|TrigSignatureMoniMT.*-- #[0-9]+ (Events|Features).*|TriggerSummaryStep.* chains passed:|TriggerSummaryStep.*+++ HLT_.*|TriggerSummaryStep.*+++ leg.*"
+    PROPERTIES WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/unitTestRun_emu_step_processing
 		)
 
 atlas_add_test( emu_step_menu_processing
 		SCRIPT test/test_emu_step_menu_processing.sh
 		LOG_SELECT_PATTERN "TrigSignatureMoniMT.*INFO HLT_.*|TrigSignatureMoniMT.*-- #[0-9]+ (Events|Features).*|TriggerSummaryStep.* chains passed:|TriggerSummaryStep.*+++ HLT_.*|TriggerSummaryStep.*+++ leg.*"
+    PROPERTIES WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/unitTestRun_emu_step_menu_processing
 		) 
 
 # Install files from the package: