Draft: Respect OnlineEnv.OutputLevel in HLT1

/cc @miolocco

we can remove these lines if not needed

Maybe we can keep the default 3 but instead pass --output-level=3 only in the qmt.

diff --git MooreScripts/tests/qmtest/hlt1slim.qmt MooreScripts/tests/qmtest/hlt1slim.qmt
index 4fa4934..9199351 100644
--- MooreScripts/tests/qmtest/hlt1slim.qmt
+++ MooreScripts/tests/qmtest/hlt1slim.qmt
@@ -22,6 +22,7 @@ Run an HLT1 job in the Online testbench
   <text>--tfdb-nfiles=2</text>
   <text>--hlt-type=hlt1_pp_forward_then_matching_no_ut</text>
   <text>--measure-throughput=0</text>
+  <text>--output-level=4</text>
 </set></argument>
 <argument name="prerequisites"><set>
   <tuple><text>create_hlt1_tck</text><enumeral>PASS</enumeral></tuple>

we should add a check on how much output we produce in the qmt validator, see #1

added RTA label

mentioned in merge request LHCb!4306 (merged)

• [2023-10-03 16:31] Validation started with lhcb-master-mr#9367

We should also find where the printout for the following messages happens in Allen. Ideally, we'll only print them if the output level is 3 (INFO) or less.

https://lblogs.cern.ch/goto/b3b34f5597c49d7b47cf1aa0f263c062

Regarding the first screenshot, we can give the output_level as an argument to the function here (that prints the sequence) and add an if output_level>3.

The same applies to the second screenshot for the function here regarding Requested option

I think only the second case is relevant for running online. The fix is probably something like

diff --git main/src/Allen.cpp main/src/Allen.cpp
index 02aed13b01..889a62b39c 100644
--- main/src/Allen.cpp
+++ main/src/Allen.cpp
@@ -223,7 +223,9 @@ int allen(
   }
 
   // Show call options
-  print_call_options(options, device_name);
+  if (logger::verbosity() >= logger::info) {
+    print_call_options(options, device_name);
+  }
 
   number_of_buffers = number_of_threads + n_mon + 1;

done in https://gitlab.cern.ch/lhcb/Allen/-/tree/miolocco_reduce_verbosity?ref_type=heads

Might be useful to assign the output_level to the default verbosity here?

for online running allen.py is irrelevant. As far as I see the verbosity argument to allen.py is the way to configure the output level.

MooreOnline/run MooreOnline/MooreScripts/scripts/testbench.py MooreOnline/MooreScripts/tests/options/HLT1Slim/Arch.xml --working-dir=hlt1slim --partition=TESTHLT1SLIM --test-file-db-key=2022_mep_253895 --tfdb-nfiles=1 --hlt-type=hlt1_pp_forward_then_matching_no_ut --measure-throughput=10 --output-level=4

less hlt1slim/HLT1_0.log

added 118 commits

• 9fe6880e...56f192dc - 117 commits from branch master
• 4f336838 - Respect OnlineEnv.OutputLevel in HLT1

Compare with previous version

assigned to @miolocco

added 1 deleted label

changed target branch from master to 2024-patches

should we also silence https://gitlab.cern.ch/lhcb/Allen/-/blob/master/stream/sequence/src/HostBuffersManager.cpp#L57? (No empty buffer available)

mentioned in commit Allen@41ab9a9a