 |
 |
Hi, Maurik. You wrote in the software meeting notes: https://confluence.slac.stanford.edu/display/hpsg/2015.10.01+--+Software+Meeting "Not all the steering files work with all the versions of the software. This may be getting confusing (e.g. current trunk cannot run the EngineeringRun2015FullRecon.lcsim but work OK with EngineeringRun2015FullRecon_Pass2.lcsim). We are currently not keeping track of what works with what. How can we do better?" I am seeing the same thing now, that EngineeringRun2015FullRecon.lcsim is broken from a null pointer exception and the "Pass2" steering works okay. I haven't checked "Pass1" but I'm assuming it is probably broken as it is similar to the one that is failing. I believe the solution to doing a better job here is relatively simple. We can add test cases that run an lcim job for each critical steering file and perform basic sanity checks on the output like making sure tracks/clusters were reconstructed. And when the Hudson build fails because one of these steering files is broken, we would then know which checkin caused it to fail and/or on what time scale this occurred (e.g. if something else other than a code change caused it like different conditions data being found in the db, etc.). On a related issue, there are likely many other broken steering files. Some of them cannot be easily fixed and so these should be sandboxed (if they will eventually be fixed up) or deleted if they are no longer needed or have been replaced by something more recent. --Jeremy ######################################################################## Use REPLY-ALL to reply to list To unsubscribe from the HPS-SOFTWARE list, click the following link: https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1