Choose timezone
Your profile timezone:
ROOT Review planning
Present: ROOT team, Marco, Peter, Benedikt, Bill(R), Valeri(R)
ACTION:
o Axel/Danilo: Mail Rolf+Valeri to confirm the absence of need of tags
o Pere: Check with Philippe and coordinate the beta2 tag
o Danilo: Study the patch of Bill for the IO Rules, which has been submitted
via Jira (ROOT-6036) and test it
o Marco:
- Loading dicts in different order have different effects: entries in the
TClass seem to be shadowed.
- One Release is needed in Feb and one at the end of March, due to new
LHCb plans.
- The plan: not pick up Beta2 BUT whatever is there by the end of March
- We loose therefore 1 Month of validation time, 1 month less to get feedback
- End of March is a HARD DEADLINE, if LHCb manages to tag the full stack
before, better.
- Concretely the test of the packages down to Da Vinci should succed,
Boole has lower priority in the list.
o Pere
- Beta2 is going to be what we have now, modulo the small things Philippe
would like to have in.
o Peter:
- Next week report results of massive tests going on
- Comparisons between ROOT5 and ROOT6
o Valeri:
- Full atlas stack againt 5.34.13 with no Reflex at all compiles!
- Next week: atlas ready to switch to ROOT6
- No special tag is needed for Atlas, as ROOT is an external (see actions)
o Bill:
- Replaced the last Reflex usage un CMSSW!
- All 7 relval tests fail
- 3 different reasons: 2 are linked to ROOT.
- 1: IO rules declared to be different even if the code for the transformation
had different spaces. Patch will be provided (see actions)
- 2: Checkum error: Bill is investigating it
- The IBs were built with LLVM in debug mode: starting from today, they will be
in opt mode. We will gain a factor of ~10.
- A relval is failing with a problem of boosted decision trees: not likely to be
conditions related, rather Reco/BTagging.
o Andreas:
- Might have spotted a relocation issue in the condition area: investigating
- Issue in reading back ROOT6 ROOT5 data: investigating the details
ACTION:
o Axel/Danilo: Mail Rolf+Valeri to confirm the absence of need of tags
o Pere: Check with Philippe and coordinate the beta2 tag
o Danilo: Study the patch of Bill for the IO Rules, which has been submitted
via Jira (ROOT-6036) and test it
o Marco:
- Loading dicts in different order have different effects: entries in the
TClass seem to be shadowed.
- One Release is needed in Feb and one at the end of March, due to new
LHCb plans.
- The plan: not pick up Beta2 BUT whatever is there by the end of March
- We loose therefore 1 Month of validation time, 1 month less to get feedback
- End of March is a HARD DEADLINE, if LHCb manages to tag the full stack
before, better.
- Concretely the test of the packages down to Da Vinci should succed,
Boole has lower priority in the list.
o Pere
- Beta2 is going to be what we have now, modulo the small things Philippe
would like to have in.
o Peter:
- Next week report results of massive tests going on
- Comparisons between ROOT5 and ROOT6
o Valeri:
- Full atlas stack againt 5.34.13 with no Reflex at all compiles!
- Next week: atlas ready to switch to ROOT6
- No special tag is needed for Atlas, as ROOT is an external (see actions)
o Bill:
- Replaced the last Reflex usage un CMSSW!
- All 7 relval tests fail
- 3 different reasons: 2 are linked to ROOT.
- 1: IO rules declared to be different even if the code for the transformation
had different spaces. Patch will be provided (see actions)
- 2: Checkum error: Bill is investigating it
- The IBs were built with LLVM in debug mode: starting from today, they will be
in opt mode. We will gain a factor of ~10.
- A relval is failing with a problem of boosted decision trees: not likely to be
conditions related, rather Reco/BTagging.
o Andreas:
- Might have spotted a relocation issue in the condition area: investigating
- Issue in reading back ROOT6 ROOT5 data: investigating the details
There are minutes attached to this event.
Show them.