Present: Joanne Bogart, Toby Burnett, Jim Chiang, Seth Digel, Richard Dubois, Warren Focke, Berrie Giebels, Navid Golpayegani, Traudl Hansl-Kozanecka, Heather Kelly, Julie McEnery, Pat Nolan, Chuck Patterson, James Peachey, Dirk Petry, Leon Rochester, Alex Schlessinger, Tom Stephens, Tracy Usher
For this and other reasons, Toby would like to move to v1r16, which various people in various situations have been using successfully for some time. What about demonstrating it works (or doesn't) with a special RM run? (Heather) This has to wait until we have an acceptable EngineeringModel tag. Then, if that tag builds and runs successfully on both platforms with v1r16, it should be ok to move.
To eliminate the complexity of test program versus library dependencies, Toby made a container package for GRB. (This temporarily broke the EngineeringModel build, now fixed.)
(Leon) Why even have a separate EngineeringModel? There are no signs of a stable build to date. (Richard) True, but it's not needed yet, either. Our experience with EM1 was that ongoing changes to GlastRelease caused I&T much grief. This way we will be able to isolate I&T from our ongoing GlastRelease development. There may be a better way, but we haven't found it yet.
(Julie) Why are three identical executables (EM1, EM2, LatIntegration) built? The only difference among these packages is in their job options.Will the executables ever differ? What should System Tests use? (Heather) For now yes, the executables are identical. She would like to reserve the possibility of differences in the future. Would be nice to run the test for a couple different geometries. (Julie) This is being done now.
(Joanne) A footnote and a plea: There is a relation between Toby's two bad news items ("Joanne is very annoyed", "I spent a lot of time that could have been much better spent"). I was simultaneously grateful for the fix, and, yes, annoyed that modifications beyond those needed to fix the immediate problem were made to a file which is my responsibility, even though I concur with the intent behind the changes. Working on the project together with our somewhat complementary expertise when I returned a couple days later would have eliminated the annoyance and saved Toby some time. Please, let's keep to the policy that, except for urgent items, no changes should be made to a package by a non-owner without the owner's knowledge and consent.
System tests: (Julie, Navid) See status and future plans. Highlights include
(Toby) What about other extensions, for example analysis of PSFs? (Julie) Yes, they plan to incorporate PSF analysis, but not ability to, for example, set cuts from the Web interface. They do make the ntuples available from the interface; users can then do private analyses on this data.
External library conventions: (Navid) The conventions used by the Installer are laid out here. (Richard) Can we expect a roll-out of the Windows external library installer before the Windows RM is ready? (Navid) Yes, it should be possible to do external libraries independent of Windows RM, but not the full binary installer.
SSC: (James) They're in their second build cycle, on schedule.
Flight integration news for the week:
previous | minutes index | next |
J. Bogart Last Modified: 01-Jun-2010 15:47:24 -0700