April/May priorities: (Richard) See them all in Confluence. Will be updated after the Service Challenge meeting tomorrow [Wednesday].
Filter bits (Heather) We have the new DFI and a new EventHandler class, providing a list of handlers. This information makes its way to the TDS. Still to do: get it into ROOT output. (Richard) and also into merit; Leon is working on this part. (Anders) Note this involves changes in Bryson's 1/2 pipe code which have to be cleared by the Flight Operations CCB.
GR v14 validation (Richard) v14r8 validation runs are complete and look good. (Anders) There is one problem: v14r8 can't process real data. (Richard) That is fixed in v14r9. v14r9 looks like a good candidate for branching (thereby releasing the main branch for, e.g., Pass 7 development).
(Joanne) Has a couple candidate updates. One is support for Tracker alignment calibrations; the other has to do with recovering (LATC) intent information. Leon has not yet had a chance to try out the alignment calibrations. The intent queries were requested by Eric C., primarily for use in stand-alone programs. (Heather) leans towards including these things in the spirit of keeping packages up to date; should be harmless if unused. (Leon) was under the impression that tracker alignment was not a top priority. Richard) Do we have inter-tower alignment data? (Leon) Yes. It's not perfect but it does improve things somewhat. We also have intra-tower data, which is more important. (Anders) Eric G. has asked for these calibrations. (Richard) Bill wanted a muon run for alignment purposes; does not look like this will happen.
(Richard) Is there anything else which might go into L1 branch? (Heather) knows of only one more thing: updates Johann and Claudia are working on for GCRCalib. There is one part of those updates which is believed to be important to include in L1 processing. They are not yet finished, and need a way to test their updates. Heather will provide them with a branch, and we may be asking Richard (in Tom's absence) to start up a couple of runs for them to take a look at.
Pass 6 reprocessing (Richard, Heather) The problem described last week (presumably connected to all the beating on xrootd for these particular jobs) was reproducible by xrootd support and they have come out with a fix involving the client side, hence it needs to be incorporated into ROOT. There is no tagged version of ROOT (yet) with this fix, so we have had to build it locally just for this task. It's not needed for L1 processing.
There is another issue with these runs: some input files have no events because of SAA, in fact for some of the the runs none of the input files have any events. ROOT doesn't like this! Heather has addressed both problems. One can now ask to look for these conditions and handle them appropriately with job options (default is old behavior).
Science Tools : Jim went through the Science Tools Update for May 13th.
Data handling: (Dan) New release of data client has tagging of folders and groups with metadata.
Documentation: (Chuck) The May newsletter is out, as well as pictures from the recent field trip to the Alpine Inn. Work on How-to-fix continues. He is in the process of adding information from Bryson and Philip on FASTCopy injest. Tony has also made a major contribution in the form of a new monitoring tool, which may be incorporated into Nagios. Most importantly, he has identified and incorporated every resource he thinks must be monitored, and has created tests for each.
Skimmer (David C.) is refactoring the Skimmer: most binaries have been merged into a single one and many duplicate operations are being removed one by one. Once done, the next step will be the full support of CELs as input (currently, only the file list is taken into account).
Several issues raised by Tom and Wilko seem linked to xrootd failuers. A fix has been developed by the xrootd authors [see Pass 6 Reprocessing above], which should be available in the next ROOT 5.18 patch. [Thanks to David for substantially augmenting my puny notes.]
Event display (Heather) What is happening with Riccardo and DataMind? (Richard) is trying to find 3 months' worth of money to fund the effort. The International Finance Committee has that much unspent money, but we are not the only ones asking for it. In any case, many of Bill's complaints have to do with fillers, not exactly Riccardo's turf. Leon is looking into this. (Leon) believes Bill wants to be able to replay events, which would require structural changes to Fred. Leon and Joe Perl are continuing to investigate using the Ace/Tao server in place of OmniOrb, but it is difficult to build properly on Windows. Another option is to try the latest release of OmniOrb.
SCons on Mac (Emmanuel) He managed to work around last week's problem with building ScienceTools by finding the correct magic set of compiler options; he's now working on the next roadblock: multiple symbols definitions complaints. (Richard) Was Eric Winter able to help with last week's problem? (Emmanuel) He's building only a subset of the ST packages, not including the one which was causing the problem.
New RM, SCons (Navid) SCons has been switched to use 0.98.3. The bug I had discovered with 0.98.0 was fixed by the SCons folks between that release and 0.98.3. So we can use SCons 0.98.3 without any change to our SConscript files. I am now using the GetBuildFailure() function of SCons to get a list of all the targets that failed. This is more accurate than my previous way of parsing the SCons output for gcc/vc++ errors. I have concentrated on building 64bit libraries of our external libraries for ST on SCons. Once that's done I think we'll have 64bit builds of ST shortly. The major problem holding this work back is ROOT being in flux right now because of xrootd. After 64bit linux libraries are working, I'm debating if I should make Intel MAC builds happen for ST or if I should concentrate on the tag collector. [Thanks to Navid for this report.]
previous | minutes index | next |