LHCb Core Software Meeting

Date and Location

24 January 2007
10:30 - 12:30
CERN (2-R-030)


Andres (phone), Florence, Gloria, Hubert, Marco Ca., Marco Cl. (minutes), Matt, Olivier, Philippe





  • LCG 49 released
    • LCG 50 will be with the new COOL (2.0.0), later a version with the new CORAL

Software Releases

Gaudi (Hubert)

  • Released v19r1 only on Linux
    • problem with the dictionary and DataStreamTool on Windows. It needs a fix in Reflex, but a temporary work-around is in place.
  • Applied 3 patches by Atlas to the HEAD, including a fix in MsgStream::operator<<.
    Marco Ca. pointed out that that operator should not be used because we have implementations for ostream::operator<<. It should be clarified why MsgStream::operator<< is there and if it can be removed for the correspondent ostream::operator<<.
  • Another patch from Atlas is pending. It is supposed to fix a problem in the distruction of algorithms, but it is not clear. To be investigated a bit further.

Marco Ca.: Next versions of Gaudi should be driven by bug-fixes and the needed version of the libraries used for the conditions database (even if it does not depend on COOL, we need a new version of Gaudi to pick up a new version of LCGCMT).

LHCb, Boole, Brunel (Marco Ca.)

  • Everything released on the new Gaudi, but not yet announced because the instructions about how to work with the Install Area are not ready.
    • After some discussion we identified 4 use cases:
      • Normal User
        • work on one version of a project
        • upgrade to a new version
      • Advanced user
        • work on a project used by other projects and override the released version with the local one when working on a higher level (e.g. change LHCb and share the chenges with Boole and Brunel)
        • work with many copies of the same version of a project (e.g. bug-fix on one side and developement on the other)
    • Not to delay further the release, only the first use case will be addressed in the annuncement. All the others will be addressed in more detail on a TWiki page.

  • Thomas discovered a serious bug in the CloneKiller code (not all clones were killed).
    The Physics Planning Group have to study the problem. Anyway it does not seem feasible to restart the production and the effect shouldn't be dramatic.
  • New version of Moore/HLTApp will be released soon. Niko wants to prepare a stand-alone disk with the HLT code and some data in MDF format to easily benchmark CPUs.

Gauss (Gloria)

  • Compiled a version of Gauss with the new Gaudi featuring
    • few backward compatible changes (factories)
    • new things in the generators
    • complex changes on Giga (for the new plugin manager)
  • There will be two releases in parallel:
    • v25r8 with the backward compatible changes and the news in the generators
    • v30r1 with everything

Round Table

The change in the name of the main package of Gaudi (Gaudi -> GaudiRelease) broke the documentation web page. As a preliminary work-around, a symbolic link will be created, then the web pages scripts will be updated.
Marco Ca.
On monday, lxplus will switch to what currently is lx64slc4, so we have to give recipes about how to work.
  • DC06: just connect to lxplus and build (CMTCONFIG will be slc4_ia32_gcc34). This binaries should work on the GRID as long as they do not need library like GFAL or LFC, otherwise one can follow the recipe for DC04 (build SLC3 binaries)
  • DC04: for the moment, connect to lx32slc3 and build, later we will provide a script to provide the build-time environment to build SLC3 binaries on SLC4 machines (it will probably be SetupProject).
Marco Cl.
  • Almost ready to move XmlDDDB to SQLite
    • Moved the scripts to a dedicated directory and modified them to generate different filenames.
    • Added 2 scripts: one to extract a self consitent subset of XML files and one to rename the files and get rid of the directories used for versioning (e.g. v6 in Rich1 and v200511 in Velo)
    • Committed DC06 compatible changes in the Velo (Mark Tobin)
  • After some discussion, we agreed on the following next steps:
    • Tag XmlDDDB as v30r15
    • Rename the files in XmlDDDB and commit (possibly tag it as v40r0)
    • Prepare a new package to contain the requirements and options to use the SQLite files (it will also contain the script to generate the files)
      • We will have two databases one for XmlDDDB (almost static) and one for XmlConditions (time varying). A third one will be needed for on-line conditions.
      • The initial tag will be DC06 (that can be upgraded to DC06-v# if needed)
    • A new module will be added to CVS to act as a back up of the SQLite files (the script generating the SQLite files will start from that CVS module). CVS will allow to maintain a set of files with different versions.
    • New conditions (XML files) will be added to both the back up CVS and SQLite files

-- MarcoClemencic - 24 Jan 2007

Edit | Attach | Watch | Print version | History: r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r1 - 2007-01-25 - MarcoClemencic
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LHCb All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback