RefToBase to Ptr conversion for pat::TriggerObjectStandAlone

[nsmith-]

Description:
Implemented two templated methods: refToBaseToPtr for use in class constructors, etc., where the desire is to keep the RefToBase API available, and then refToBaseToPtr_ioread to be used exclusively in ioread rules to do schema migration of data written with the old type.

Validation:
Managed to forward-copy the data products:

$ edmCopyPickMerge inputFiles=file:../../data/ul17ttsemi_miniaod.root outputFile=../../data/ul17ttsemi_miniaod_copy.root
20-Nov-2023 17:05:19 CST  Initiating request to open file file:../../data/ul17ttsemi_miniaod.root
20-Nov-2023 17:05:24 CST  Successfully opened file file:../../data/ul17ttsemi_miniaod.root
%MSG-w FastCloningDisabled:  AfterModBeginStream 20-Nov-2023 17:05:25 CST  BeforeEvents
Fast copying of file file:../../data/ul17ttsemi_miniaod.root to file ../../data/ul17ttsemi_miniaod_copy.root is disabled because:
The format of a data product has changed.

%MSG
Begin processing the 1st record. Run 1, Event 66298012, LumiSection 66299 on stream 0 at 20-Nov-2023 17:05:25.366 CST
Begin processing the 2nd record. Run 1, Event 66298002, LumiSection 66299 on stream 0 at 20-Nov-2023 17:05:25.805 CST
...

but I cannot verify if non-null origObjCand() is working, since it appears the PATTriggerProducer default value for saveL1Refs is false, and this is not overridden in MiniAOD production (despite l1extra being in MiniAOD)

[Dr15Jones]

@nsmith- were you able to test that the IO rule worked?

[nsmith-]

It worked but only because every RefToBase is null. Indeed the misspelled (non-ioread) function would have caused an error if the member was filled. (that's how I ran into it for reco::GsfElectron conversionRejection_, which I guess would have been a better canary)

[Dr15Jones]

@nsmith- as long as you confirm this compiled, I'm fine with merging.

[nsmith-]

It compiles, yes