Core Minutes 1/22/2008

• Navid's Leg:  Navid is recovering, apparently faster than others who suffered the same type of injury that day.  That being said, he really cannot walk yet and has had fluid build up in his leg.  Navid will be meeting with a specialist to discuss in more detail his path to recovery.
   
• Big Run News: GR v13r9p1 about ready to go.  Next up is to get GRB grid going.  Tom reports bookkeeping is half-way done, and is waiting on XML files.  Emmanuel reports that the XML should be ready later today.  With the XML in hand, we may be able to try Interleave "very soon now".
  Leon is concerned about the jobOptions for Interleave.  He was concerned about maintenance and was hoping to slide in some changes.  **Maybe Leon can say more!
  Toby mentioned that the latest tag of astro is not in GR yet.  That patch would be needed for the skyModel.  Richard reminded us that we'll be running the sky twice.  Once for OpsSim2 and again for the big interleave.  The skyModels don't necessarily need to be the same for those two runs.
  Richard also reported that the server for pipeline 2 is down currently - Two of our glast lnx boxes are down.  Hoping to be back up today.  [and they were!]
  This outage is holding up Francesco to run comparison gamma run, to make sure SLAC and Lyon are getting the same answers
   
• Science Tools:  Jim provided the Science Tools update which is linked into the agenda page.  Richard asked about windows and mac coverage for the ST builds.  Eric Winter typed in the following:
  [08:12:22] Eric Winter ScienceTools v9r2p2 builds under Linux and Mac OSX using the HEASARC build system
  [08:12:26] Eric Winter Dave Davis is testing now
  [08:12:55] Eric Winter Windows port under Cygwin is still underwa\y, but still trying to get HEASARC build system to work with Microsoft compilers
  [08:13:34] Eric Winter Will likely update port to use latest tagged release this week
   
• Documentation Update:  Chuck missed a couple days last week, but has made good progress on the configuration section for ISOC web site, with Jim Panetta and Eric Charles.  The next LAT newsletter is coming out soon.
   
• ACD Review:  is set for Thursday morning at 8 AM Pacific.
   
• CELs:  David reported that there is a small bug in the glue between RootIo and rootUtil that he and Heather are investigating.  [This was fixed later in the day and is included in GR v13r9p2]
  David has a few more question concerning the CELs and the Skimmer which he has sent out via email.  Tony, Tom, and Tracy (the three Ts) all desire to start using the new CELs.  The work is ongoing, and should be ready in time for the Big Run work.
   
• SCons:  Navid reports that SCons and RM work has obviously been held up due to his injury.  We are not quite there yet to do full ST SCons checkout.  There are some packages that still require the new style of tagging, such as celestialSources.  Navid will be applying those tags by hand for now.  In the near future, Navid plans to have a tagging script which should make the job of tagging easier for all involved.  We also revisited the need for the unique tags, which is required due to container packages like celestialSources.
  
  Toby asked if there is a place where the new-style tags are in use?  Navid replied that he could modify the CVS notify scripts to send email when the new-style tagging occurs.  Currently, anything that does not conform to the old-style tagging convention is ignored.
  
  The new RM scripts are written but need to be tested.
  
  Toby requested that the how-to Confluence documentation include checkout of st-scons and run scons, as well as how to make a SConstruct file for a single package.  Navid stated it should already be possible to build a single package, and will update the documentation.
  
  Jim Chiang asked about how we will support multiple versions of the same package in the same local build area.  Navid explained that SCons will build the "largest" tag, based on a lexical sort of the tag names.
  An example from Navid::
  Let's say you have 2 versions of astro. When you check out ScienceTools-scons you get: ScienceTools-scons/astro. Let's say you want to play with a different version. You can, inside
  ScienceTools-scons, checkout a different version of astro and call that one astro-v2. Now when you compile it SCons will ass astro and astro-v2 and it will assume astro-v2 is an override of astro and use that one instead.
   

previous

minutes index

next

H. Kelly, Last Modified: 01-Jun-2010 15:46:44 -0700