ROOT 6 planning
Present: The root team, andrea, marcin, peter, david L., markus, rolf S.
ACTION:
o Danilo: Follow up with Bill the 200 MB increase.
Bill
o 6523: now version of a patch for invoking functions with references w/o crashing. CMS thinks that this is the right solution. PyROOT crashes with the
o Unit tests and Relval tests: 12 failures. A small number. 1 of them can be linked to a root: the macros with library loading. Each problem does not seem to be ROOT6 issue but this cannot be rule out until solved though. It looks good modulo the 200 MB jump which has to be understood.
o There is one problem which is likely to be root related: the dictionary issue in relval 1001 is an issue in root.
o The cutparser issue: wrong type of an argument, an enumeration. Likely to be a CMSSW one.
Peter
o Lead-lead collisions: same speed root5 and root6
o Memory: +200MB RSS
o Some events: difference in tracking. difference in 10% of the events
o Issue ROOT-6839: still waiting of a fix. It should be simple to solve.
o Some more reports are to be submitted.
Marcin
o Nightly based on dev4: fantastic achievement. Patches can be tested immediately.
o Fixes in PyCOOL are effective
o One test pending: taujet. Axel claims is fixed, no double check on atlas side yet.
o Root include path: now reduced from 180KB to 90KB. Workaroundable for the moment: keeping it as a side issue.
Andrea
o Nothing much to report. Marco, Axel and Marcin orchestrated some patches.
o ProcessLine replaced by a Declare: improvement for injecting headers in ROOT yet to be integrated in pyCOOL