ScienceTools: Jim presented the report. He still doesn't have adequate access to a Windows platform for fixing Windows-specific problems with ScienceTools.
FSSC: (Eric W.) is working on the v9r14 port; it's going smoothly. John is working on the Data Server.
Documentation: (Chuck) sends the following:
Since I last reported, there have also been a number of changes to the workbook. These include a warning regarding hazards of running multiple simultaneous batch jobs on SLAC Public and the need for a pfile directory to be made for each batch job being run on SLAC Public in order to avoid the .par file swamping effect that can bring down a server. Jim's updates to the Likelihood Tutorial, including the BandFunction formula used to model GRB spectra have also been incorporated. As for pyLikelihood, Richard, Jim, and I met the week before last to review the status of the tutorial and identified the steps and action items needed to bring it up-to-date. I've also posted an updated version of the pyLikelihood Usage Notes to the web. You can find links to all of these changes in the Change Log, which is linked in from the Workbook's "splash" page.
Before closing, I should also mention the discovery that quite a few changes are needed to bring the Science Tool Help files current with SLAC's data portal.
Pass7 reprocessing: (Tom) Merit reprocessing was submitted last weekend. All but 11 runs (out of more than 4000) completed successfully. One problem, skimmer-related, is responsible for 10 of the 11 failures. The next step is to generate FT1 for photons and electrons and LS1. This task is in final development and testing; it should be ready to go in a day or so. Warren is in charge of the remaining piece: generating FT2 files. (Warren) started processing yesterday; about 300 runs are left to go.
(Tom) Anders and others he has recruited have done and are doing extensive validation which has been very helpful.
See also status in Confluence.
GlastRelease: (Heather) v17r31 is the tag being used for reprocessing. There is now a v17r31p1 tag along a branch, of no interest for reprocessing, containing new code from Eric Charles. He and Anders have been checking it out. (Anders) So far, so good. (Heather) There will be more tags in this series as Eric adds more new stuff.
Core Week will be the last week in July. (If you have objections, get back to Heather ASAP.)
skimmer (Heather) David Chamont has been alerted to various problems Leon has encountered and has helped with some.
SCons RM (Navid) was working on a reature of RM to erase old builds and inadvertantly erased the entire RM home directory! SCCS makes nightly back-ups of AFS so he was able to recover almost everything; however that directory includes links to directories in nfs which also got erased, among them the builds of QT upon which the new RM depends. He has rebuilt QT for rhel4 and as of meeting time was nearly done with rhel5.
A recent tag of astro, incorporated in ScienceTools v9r14, should take care of the last of Mac build errors. v9r14 should build correctly on Mac.
SCCS has not yet figured out how to get around the afs token problem with batch jobs on the Mac. They suggest that, in the meantime, a possible workaround would be for them to put a host-based acl on glastrm's home directory. Anyone able to log onto bldmac01 would have read access to the directory. Navid will confer with Richard.
SCons & Windows (Joanne) thought it would save some work if st_graph built a shared rather than a static library. James Peachey, package owner, was agreeable. But in the end it appears it's best to leave st_graph (and a couple statically-built libraries it depends on) alone; the cure is much worse than the disease.
SCons endgame (Heather) is working on a list. Ultimately would like to add names and approximate ETA for remaining tasks.
Command-line environment (Toby) and other Windows users are in the habit of invoking science tools from a dos window after doing CMT setup so that environment is set up properly. He's concerned this mode of working might not be supported with SCons. (Joanne) It is more or less; a wrinkle or two remain to be worked out. Facilities to address this issue include
To use any of these tools with a pre-built installation, user must have already defined GLAST_EXT and INST_DIR.
Fred/FOX (Emmanuel) v0r100p1 is up and running on Windows. It differs from operation on other platforms in that the external library FOX is not really used; Fred can use its own copy of FOX which is included with the Fred distribution. He has modified the FOX interface package requirements file to be essentially a no-op for Windows.
previous | minutes index | next |