ScienceTools: Jim presented the report. The latest release is now v9r11. See the report for a description of what's new. (Richard) Perhaps we should distribute IRFs along with the release. (Philippe) Yes, if we first test them more thoroughly.
FSSC Report: (John) No news
Documentation: (Chuck) sends this report:
I'm still working on trying to identify everybody in the pictures from the Collaboration Mtg. Masa also reports that he's working on some changes to the Pulsar Tools tutorials. He plans to replace the sample data files with those created by the Science Tools simulator tools.I'm also posting the blind link to the "GLEAM Application: Processing Real Data" overview in the hope that some of you will have an opportunity to browse through it sometime before next week's meeting and let me know if there are any corrections, comments, concerns, etc. If there are not, I plan to go ahead and take it live by linking it into the WB.
Beyond that, I plan to resume work on documenting the references and parameters for the Python Likelihood tutorial. The next edition of the Newsletter is also scheduled to be published next week.
GlastRelease: (Richard) invites Leon to describe the most recent rocky road to a new release. (Leon) By Sunday we thought everything was ok, but on Monday we discovered it wasn't. Tracy then found and fixed the last bug in Overlay. Heather tried to make a new release incorporating the fix but ran into problems with u30, where RM puts CMT builds, so she had to make a release by hand. There is still a problem with this release: it doesn't write the merit file header. Probably this will be straightforward to find and fix, but we do need to fix it. Then we can go back to the program:
(Leon) Might take a little more work before we're ready for backgrounds; we'll see.
u30 (Heather) Concerning u30: Redhat 9 builds were taking 8-9 hours [typically they take about 3.5 hours] and later on there were I/O errors. (Navid) It would take about half a day to move CMT builds to u35. He has not yet heard back from unix admins about what the problem with the disk might be. (Richard) Could this be due to contention or is it hardware? (Heather) Saw problems all day. (Tom) The load has been fairly constant over the last few days so the I/O errors are probably not due to contention.
Archiving status (Heather) Wilko had gotten all the data off of several disks. (Navid) Procedure has been: Wilko makes a list of archived data; Navid compares it with what the database thinks should be there. Tom will compare archived data with original. (Tom) Has done so; Anders says it looks good. Still need to confirm we're archiving everything we should be archiving. Karen is working the final piece of this project: a web page describing how to retrieve files from the archive.
astro revisited (Heather) A couple weeks ago we put a newer version of astro in GR used by the L1 pipeline. Has this fixed the problem with incompatible versions? (Anders) Yes; the versions for GR and ST are the same so there is no longer a problem. However, we shouldn't assume they always will be using the same version. (Warren) The only real problem is ft2Util. (Heather) would like to make a container for ft2Util; this is the only way to know for sure what version of astro was used to build, whether static or shared library is used.
New GR for pipeline (Anders) tried p7; it seems fine. Will try to put it into L1 this week. Need System Tests.
SCons RM (Navid) has solved Mac compile problems (again). ST now compiles on Mac except for some warnings. Neal [of SCCS] believes he's got Mac LSF problems fixed. Should be able to run builds on one of our Mac nodes, buildmac01. buildmac02 has been taken down to install leopard, the next OS version. Once this is done Navid will create externals for it, etc.
(Heather) What about auto-notification of build failures for ST? (Navid) not done yet; should be later this week.
(Navid) CHS build was a partial success; he needs to investigate further.
SCons/Windows and GoGui (Joanne) SCons now writes more or less usable project and solution files for her 2-package development installation. However Studio always thinks it needs to rebuild, whether or not it actually does. See the Windows log for details. At the bottom of the screenshot you can see the two "stop" buttons (one for SCons, one to stop the currently-running user application, if any) and a field where a specific build target may be entered.
Externals (Emmanuel) has been testing Fred as started up from Gleam. It's not working yet. He'll email error output to Heather and Joanne.
He has built several add-ons for python on Mac (version 2.5.1-gl1): numpy 1.2.0, 4Suite-XML 1.0.2, pyfits 1.3 and ipython 0.9.1.
previous | minutes index | <next |