ScienceTools: Jim presented the report.
(Toby) notes that it is not just the tip test program which fails for the latest tag on Windows, but essentially any use of tip. (Jim) It turns out the original reason for putting large file support in tip was a red herring; we don't actually need it. He will contact James Peachey and ask him to put a patch in the code so that it will revert to old behavior on Windows. James P. has done this and tagged tip v2r14, which now compiles and passes unit tests in the RM on Windows.
FSSC Report: (John) The new release was made late last week. There are known problems with 64-bit machines. The work-around will be to use the 32-bit build with appropriate compatibility libraries. There are also problems with Ubuntu which are being worked on.
Documentation: (Chuck) sends this summary:
In addition to routine updates to the workbook, I'm continuing to work on a software overview for the ACD, TKR, and CAL. At the moment,this appears to be a rather daunting task, and if anyone knows of a writeup or presentation on any one — or all — of these that is particularly useful, I would be grateful if you would let me know about it.
I would also like to remind everyone about the new Change Log linked in from the Workbook's splash page and located on the right, just below the illustration of the LAT.
GlastRelease on RHEL4: (Heather) Johann tested GR with the G4 optimized build on RHEL4. Crashes are gone! And running time is improved. We should be ok when, before too long, Flight Software stops supporting RHEL3.
GR tags: (Heather) v15r50p16 has all the latest Overlay tags. They are also now in HEAD and will be in the GR tag in progress, v17r1.
(Leon) The latest AnalysisNTuple tag has new variables requested by Jonathan Ormes concerning magnetic latitude, rocking angle, and livetime.
skimmer: (Heather) The skimmer has had problems merging merit files with different numbers of columns. (Leon) stumbled on the reason; a fix is now in progress. (Heather) David Chamont is working on it. It turns out the default "fast merge" will not work when the files have different structure. David is planning to keep that as the default, but add the ability to turn it off. There still is the issue of how the user will know it needs to be turned off.
Overlay: (Leon We have a 20-million event all-gamma dataset with and without overlay and also a 1-day sample. People are looking at them now. (Tracy) One more feature is yet to come. Would like to get delta event time (time since last read out event in ticks of 50 ns clock) and delta window time (time to last trigger window) from periodic trigger overlay event into the ntuple, then see to it that various pieces of reconstruction and perhaps analysis code which should pay attention to these quantities do in fact use them. It will take some care to make sure this doesn't meanwhile break something else. May want to wait on making larger event collections until this is in.
ACD mini-review: (Heather) It took place last week. Those who were not present and have any interest in, e.g., TDS structures or just the ACD itself are urged to look through the comprehensive documentation (scroll to bottom to see all the child pages) on their own. See also — and add to, if you are so inclined — Eric's list of proposed tests.
u30 (Heather) Navid recently cleaned out u30, but it's back up around 92% full. (Navid) is looking into how much space is taken by SCons builds. Perhaps can recover some space that way, but not enough long-term.
RM and friends (Navid) reports:
The mac builds in SCons still suffer from the problem that lsf frequently decides that it will not submit a job to actually run. It simply marks the job as running and keeps it that way for several hours. Eventually it just stops the job without having ever actually executed it.
The RM for windows in SCons has been turned back on for testing but due to the Windows outage and now the u30 problems I've had to put this problem away again while working on a fix for u30.
The RMViewer has a new feature that shows you what the status for the Release Manager is. I'll work on expanding it to show all Release Manager status information as well as allow you to modify the information.
CHS builds have been enabled but none has occurred yet because there have been no new tags. He could initiate one manually.
SCons & Windows (Joanne) lurched forward and back over the past week; net result was a little bit of progress. She can use the vs_revamp development branch of SCons code both from the command line and from GoGui, but as of yesterday afternoon couldn't get any further than that because of upheavals in SLAC central Windows services. See latest entry in the log for more detail.
ROOT Mac build (Emmanuel) Almost there. He just needs to fix id names and then test.
cfitsio (Emmanuel) rebuilt and reinstalled the cfitsio patch, v3060A, for rhel3, rhel4 and Mac (tiger-i386-32-bit) machines. The first time he had inadvertantly built static libraries; this time they are shared.
previous | minutes index | next |