Core Week: (Heather) See her extensive minutes for the kick-off meeting on Oct. 20 and for the SCons discussion later in the week. Main issues discussed include
See details on these and others below in Core Talk.
Overlay: (Tracy, Leon) Plans and status are all described in detail in this Confluence page and in the presentation Leon gave at C & A meeting yesterday. Executive summary is: we're very close to being able to try this out. Some ACD code is still needed. Ideally this would be written by Eric Charles, but he is not available this week.
Other GR: (Tracy) Claudia reports problems running GlastRelease in Lyon. The crash looks very much like other crashes which have been traced to problems in initialization, but in this case all the usual suspects have been exonerated.
ScienceTools: No one was present to give a report, but you can read it for yourself.
Documentation: (Chuck) kindly sent the following:
RM Navid sends the following:
I had a few productive meetings with Karen to discuss the Release Manager frontend. The RM frontend made some considerable progress since it was a lot easier to explain the database tables in person that over email. Karen expects we can have something useful and working within a few weeks.
Concerning the backend of the RM, I've now compiled Qt for the mac and have successfully compiled the RM code against Qt on the mac. I need to test the code and have the SLAC administrators put our two mac machines on lsf. Once that's done I'll be able to submit SCons builds for mac. On the windows side the RM has been setup to submit jobs for compilation but because windows is so much different than the other platforms it needs extra work that I haven't had a chance to do yet. As a result, jobs are currently getting submitted for windows builds but they all fail.
SCons and Windows (Heather summarizing Core Week discussion) The project and solution files built by SCons are not adequate for Windows developers. Our plan is to have SCons instead invoke customized Python code. The work Toby did some time ago will provide a starting point. Tracy will help to incorporate it.
GR and Externals (Emmanuel) He has gotten past the obf hurdle and has about 9 GR packages left which don't compile, such as GCRCalib. Most likely he will need to collaborate with package owners. See GlastRelease-scons build status page for details.
GoGui (Joanne) The Core Week demo largely followed the organization of the GoGui how-to documentation. There were a couple suggestions:
#2 was already on the GoGui to-do list; #1 has now been added.
CHS (Heather) belatedly — but still in time for Core Week — realized that the CHS container also needs to build under SCons. Predictably Bryson got right on it and managed to get most of it done, meanwhile implementing some improvements to our SCons machinery which we may want to port back for GlastRelease and ScienceTools. (Joanne) See Bryson's additions to the SCons Discussion page (scroll down about half-way) for details. There are three packages in the CHS container which he didn't attempt to build since they had no SConscript or xxLib.py files: LATC_vrfy, fswDecipher, and MOOT. The first two are only in there because MOOT uses them and all three are more or less in my bailiwick. fswDecipher was straightforward to convert. As Bryson mentions, LATC_vrfy is awkward because it has to access fsw files which are not really meant for export. Some of the functionality of LATC_vrfy is already available in the current LATC parser (rewritten by Owen Saxton some time ago, but well after MOOT was in operation); it would simplify matters considerably if we could dispense with LATC_vrfy altogether. MOOT presents no special problems other than its dependency on LATC_vrfy. [(Joanne) After the meeting I discussed this with Owen. LATC parser does not now quite do everything MOOT needs, but he is willing to upgrade it so it will. At that point some limited but crucial pieces of MOOT would have to be carefully rewritten.]
FluxSvc, PointInfoAlg and so forth Toby will review the situation. Expected action is to back out some changes to FluxSvc.
Fred (Heather) Fred as distributed does not work with Gleam on a RHEL4 machine. Joanne has builds of fox and Fred which do [or at least did as of last December]; if Heather can confirm they work for her as well, we can distribute these versions.
WIRED (Heather) There are three issues here:
previous | minutes index | next |