update the cmssw release and MET recipe

[OlivierBondu]

Getting started on https://github.com/cp3-llbb/Framework/issues/257

Let's try to keep the PR factorized on the different topics in order to properly check that changes are not breaking unexpected things...

First: updating the CMSSW release!

[OlivierBondu]

please test

[OlivierBondu]

Thanks @blinkseb to have restarted things up :)

It looks like the MET branch was updated wrt the commit jenkins was told to pick up, which is the reason why it compiled locally but not on jenkins... anyhow, looks like everything is fine again

Nota Bene: the MET recipe was not updated, but the old recipe was out of date, so updating MET recommendations is still on the todo list

[OlivierBondu]

About the MET commit: what happened is that changing the release worked locally out of the box and not for jenkins. The reason is that

• the readme instructions are about git cms-merge-topic cms-met:METRecipe_8020 -u
• the bootstrap_jenkins.sh is about a specific commit of that branch, namely git cms-merge-topic cms-met:83c1f5d9bfcbc669a0621cdd3c01893f47b268be -u (a change you introduced btw :wink:)

what I did is just update the specific commit being picked up by jenkins

[OlivierBondu]

but actually maybe we should revert the bootstrap script to git cms-merge-topic cms-met:METRecipe_8020 -u

[blinkseb]

Please have a look here: https://twiki.cern.ch/twiki/bin/view/CMS/MissingETUncertaintyPrescription

Recipe looks updated

[pieterdavid]

I missed that the hash change was in the second commit, thanks for explaining! A hash or a branch name in the bootstrap script is a tradeoff between maintainability and reproducibility... I'd still go for the latter, but ideally they would make tags to make this easier ;-)

[OlivierBondu]

I know, I just didn't want to update everything in one go... to factorize and keep track of changes on the ref trees, but indeed, maybe it's worth doing here as the current recipe is broken anyway

let me have a look then

[pieterdavid]

Looks good, thanks!