ScienceTools: (Jim)
FSSC: (John) is awaiting test files in order to verify that they can handle parallel ingest. (Warren) will supply one.
Documentation: (Chuck) sends the following:
I've been taking a look at the implications of what's to come in light of all the software infrastructure changes and the impact they will have on the workbook. It looks like we should probably archive and replace the Instrument Analysis section, but the real question is should we have a password protected archive in order to "hide" outdated material from the search engines, or should we simply flag the pages as being outdated. So far, the feedback is that we flag the pages, and I'll proceed accordingly.
I'm also working on a tutorial and a help file for the Science Tools modelEditor. The first iteration will be based primarily on the FSSC's version.
Pass various See the ambitious C & A schedule of projects. Some prompted further discussion:
(Julie) Don't we need to have the new ScienceTools release before we switch L1 to p6v8 diffuse responses? (John, Jim) Now that caldb is separated out it suffices to get the caldb release out there before the data, preferably as soon as possible. [update: clearance came through and Jim has created caldb_release-03-01-00]
(Tom G.) We'll need to decide on an order; it wouldn't make sense to do multiple large tasks concurrently.
Tom will be on vacation for a week or two here and there; we need a back-up.
(Tracy) Concerning the other big-ticket item which did not make the C & A schedule (lack of GCR events for the last 8 months): the root cause is that the version of flight software our obf is based on, B1-1-3, has a bug. The version running in the instrument, B2-x-x, has the fix. If we create and use a newer obf, the problem will be fixed. However, fsw was supposed to be working on simplifying so that we didn't have to drag so much of flight software into obf. He would prefer not to take the time (which he doesn't actually have) to prepare an obf for a newer fsw version now and then do it again when the simplification is complete. (Heather) is concerned that the manpower for the promised simplification (and, also critical, adjustments to allow native 64-bit builds) is not there, so this will not happen any time soon. (Anders) Another possibility is to get the information (which does exist in the event, just not in the most convenient place) to GCRRecon. (Tracy) Yes, the information is available, but only (now) after GCRRecon runs. Could contemplate changing order or other patches to get information earlier; this is probably the fastest way to get to a solution (if speed is important here). But any such solution would still involve work.
Pass8: (Tracy) There will be a full meeting this Thursday. There is a lot of activity but nothing critical to report at this time.
Python upgrade: (Heather) Python 2.6.5, including scipy, is not installed for rhel4 and 5, both 32-bit and 64-bit. We still need builds for tiger, snow leopard and Windows. It would be good to try building ST with it. [Update: Jim has done so. Seems fine.]
CLHEP upgrade: (Heather) We now have builds of 2.0.4.6 for rhel 4 and 5, 32-bit and 64-bit.
ROOT: (Heather) Eric C. has asked for a patch, needed by TMine but not of interest to anything else. This will be ROOT v5.26.00a-gl2.
Gaudi upgrade: (Heather) got past one more obstacle having to do with loading libraries belonging to a different package. She is now at work on LdfConverter.
Snow Leopard: (Heather) No news from last week. Xerces and ROOT are still left to do.
Windows builds: (Heather) are still failing, though perhaps for a slightly different reason. Key files, etc., are in place. There seems to be an issue with running ssh on certain machines when it's located in its normal location. It works when copied somewhere else. The machines will be rebooted tomorrow; perhaps that will clear up the problem.
SCons topics
(Heather for Tom S.) Tom has turned on Doxygen generation for SCons builds and it's working.
(Joanne) has been bringing GR SConscripts up to current practice. Updates have been committed for 80 packages; almost all are tagged. There were some other updates, mostly to job option files, to eliminate references of the form $(MYPACKAGEROOT)/someFile. Everything builds all right and test_Gleam almost gets through the initialize phase before crashing, leaving no clues.
She has started to look into SCons-only release tag and HEAD build creation. This is not going to be quick and easy!
previous | minutes index | next |