Attendees: Fabrice, Serge, Bill, AndyS, Daniel, AndyH, Jacek
Issues raised by AndyS:
1) std::list --> std::vector
- why? Default now is vector, iterating over vector
is much more efficient than over list
- reason why we use lists: want to convey the intent
(put things in the end, remove from the end etc),
yes can be implemented through vector
- revisit on case by case bases, do not blindly replace
- preferred solution: typedef, and name it in a way that
conveys the intent (e.g., might call it a "container"),
underneath use vector
2) "boost::shared_ptr =(new T())" --> boost::make_shared()
- yes, do it wherever it can be done
3) remove obsolete strToDoubleFunc (and more) in util/stringUtil.h
- yes, ok to drop it if really not used
--> Jacek will create jira issues for the above
Side discussion: test the code early with C++11
- CERN has a script that sets envvar which hooks to cern's
version of c++11 and pulls it from publicly visible afs space
--> AndyH will post the information to lsst-data
AndyS
- playing with latest code from Fabrice, chasing some
issues with tests
Daniel
- trac ticket cleanup
- #3166
- reviewing CSS ticket
- went to Amazon AWS meeting in SF yesterday, met local
Amazon people, they want to try our use cases
(near neighbor etc), we are planning to meet with
them to follow up, perhaps we will get a chance to
test Qserv code on AWS with them at no cost?
Bill
- implemented hierarchical logging and implemented
changes suggested by K-T
--> push the changes
- will update trac docs ASAP
- we should schedule mtg w/broader DM, maybe next week?
- need to package log4cxx, ideally would be good if
someone could do that for us, don't want to dump
too much on Fabrice
AndyH
- preparing to integrate changes done for us into the
main xrootd master branch (integrated the new interfaces
with all clustering code, etc)
- planning to document it all now (for lsst needs, if we
find others who would like to plus different system instead
of xrootd to that interface, than will augment the docs
as needed)
- we can throw oss interface now
--> we need to start migrating qserv to the new xrootd
client, maybe a week after the mysql conf
Fabrice
- the in2p3 team loaded all partitioned W14 data, 1 TB,
validating and comparing to non-partitioned data now
- will integrate the new partitioner into the new integration tests
- another engineer (Sebastian) from LAL is now 30% on lsst,
playing with docker (https://www.docker.io/),
already packaged Qserv into docker image ! :)
- very useful if people want to just get a working, pre-canned
qserv with some sample data and try it out
- closing the trac #3100 / Jira DM-231 (refurbish configuration)
- large ticket
- AndyS will review
Serge
- several very large code reviews
- done with reviewing Daniel's join-representation ticket
- CSS next
--> close the small subtasks and review them as part of the
main CSS review.
Jacek
- finished and packaged CSS for review
--> document how to setup qserv with zookeeper
- reviewed #3166
- module-specific namespaces done and reviewed by AndyS
- major surgery
- want to push CSS to master first, than this.
- lots of minor issues found/reported thorugh qserv-l
- switching to std::pair vs using super simple css/IntPair?
- document in swig trac page
--> stay with IntPair, but rename to the actual names
FYIs
- one candidate for SLAC db position will be at slac Apr 9-11
- Fabrice 6 months at slac, starting some time in October
--> Jacek will initiate paperwork work (visa, travel)
- mtg w/Emmanuel earlier this week, interesting work on parser,
too early stage to judge, will have another meeting in ~2 months
- No qserv meeting next week, Jacek, Daniel and AndyS all gone
Jacek
########################################################################
Use REPLY-ALL to reply to list
To unsubscribe from the QSERV-L list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=QSERV-L&A=1
|
