GR updates: (Heather) Currently the latest GR release with Sys Test results is v15r42. v15v40 is production (used for L1).
FSW B1-1-2 and B1-1-3 are due out this week. One or the other (TBD in a day or so) will go into effect on Sept. 29. (Tracy) B1-1-2 has only minor changes and bug fixes. It has new pedestals hence there will be subtle changes in energy evaluation. B1-1-3 has changes to the gamma filter, among them a reordering of status bits. This affects not only OBF but everything downstream, including Sys Tests. The other significant (and non-backwards-compatible) change is to DFI. (Heather) This affects the half-pipe and many packages downstream.
ScienceTools: Seth went over the weekly report.
RHEL5 (Heather) We need to build externals for RHEL5. (Navid) There is a small amount of additional work needed to get RM to make RHEL5 builds, just some entries to be added to its MySQL database.
Reprocessing (Tom) Have reprocessed 20 runs from digis repeatedly (for validation of ACD geometry) and there have been no problems. Experiment #2 was to reprocess 433 runs from merit (there is currently also a need for input digis for this task, but it's a false dependency stemming from some code which could be rewritten to avoid it). Of these 2 failed for reasons not relevant to this discussion. However Seth noted an oddity: some of the runs are split across 2 batch machines, and in this case the second part doesn't get handled properly. Tom is hopeful the problem will be found and fixed soon; Heather evinced interest.
New spacecraft alignment numbers have been generated but have not yet made it to the calibration file archive. When the new calibration has been registered and the change approved we'll need to reprocess with this calibration.
Ghost tracks (Heather) Leon and Tracy have made a bunch of tags. Will wait to integrate them into GR until new FSW is in.
New RM, SCons (Navid) [sent the following summary. Thank you, Navid! ed.]
I've finished the new RM and it is now independent of the old RM. This means that LATEST builds in the new RM don't match with those of the old RM. The reason is that the timing is different. The old RM was limited to a max of 1 hour between submissions of LATEST builds while the new RM has a new max of 24 hours. (Here "max" refers to period on how long the RM can keep track of the past. If no changes happen until max period a new build isn't submitted but RM will lose track of the time. The reason for this is that the old RM daemon only ran for an hour before shutting down and letting a new instance of itself take its place. So if a LATEST was set to be built every 2 hours, after an hour the instance would stop without submitting anything because the 2 hour min was never reached. The new instance would then start the timer from 0 and never reach 2 hours again.) It is also now capable of keeping LATEST builds timings between the packages separate. So GR is set to 4 hours while ST is set to 1 hour or less. I've turned my attention currently to creating wrapper scripts for the applications.
(Emmanuel) Needs upgrades to external libraries to make further progress on GlastRelease. See also a summary of status of various 3rd party Python packages.
(Joanne) has added ability to run a program within GoGui on either
Linux or Windows (with some caveats), to do cvs diff
on a file
(Linux and Windows), and to start up a terminal window with some of the
infrastructure needed to start the process in the environment of the
currently-selected package (Linux only so far).
See the
GoGui Confluence
page for more details, including screen shots, and for a new subsection
of the To Do list: SCons customization.
Core Week (Heather) Last call for objections to week of Oct. 20th
previous | minutes index | next |