Skip to content
Snippets Groups Projects
This project is mirrored from https://gitlab.cern.ch/atlas/athena.git. Pull mirroring updated .
  1. Feb 17, 2025
  2. Feb 16, 2025
  3. Feb 15, 2025
  4. Feb 14, 2025
  5. Feb 13, 2025
  6. Feb 12, 2025
  7. Feb 11, 2025
  8. Feb 07, 2025
    • Will Buttinger's avatar
      Update SGInputLoader.cxx - promote transient providerless loads of objects to being an error · 43c326af
      Will Buttinger authored and Duc Ta's avatar Duc Ta committed
      Update SGInputLoader.cxx - promote transient provider-free loads of objects to an error, as usually indicative of missing output declaration
      43c326af
    • Scott Snyder's avatar
      CxxUtils: Add set_unaligned functions for writing potentially unaligned data. · 126f669a
      Scott Snyder authored and Tadej Novak's avatar Tadej Novak committed
      CxxUtils: Add set_unaligned functions for writing potentially unaligned data.
      
      Add set_unaligned functions, analogous to the existing get_unaligned functions,
      for writing possibly-unaligned data without invoking undefined behavior.
      126f669a
    • Scott Snyder's avatar
      AthenaKernel: Additional dictionary entry. · 92dccb65
      Scott Snyder authored and Tadej Novak's avatar Tadej Novak committed
      AthenaKernel: Additional dictionary entry.
      
      Generate a dictionary for implements<IProxyDict>.
      Fixes errors seen in dev3LCG when loading TEvent (which derives
      from this).
      92dccb65
    • Scott Snyder's avatar
      CxxUtils: Avoid spurious division-by-zero FPEs with clang19. · e7567bda
      Scott Snyder authored and Tadej Novak's avatar Tadej Novak committed
      CxxUtils: Avoid spurious division-by-zero FPEs with clang19.
      
      Use CXXUTILS_TRAPPING_FP to tell clang to assume that FP division may trap.
      e7567bda
    • James Catmore's avatar
      Adding a new D2AOD format to allow command-line skimming of PHYSLITE · 7143fa0a
      James Catmore authored and Tadej Novak's avatar Tadej Novak committed
      Draft: Adding a new D2AOD format to allow command-line skimming of PHYSLITE
      
      This MR adds a new D2AOD format (D2AOD_SKIM) to the derivatioon framework to allow users to skim PHYSLITE via the command line, rather than the current situation where they must set up their own formats. Usage is:
      ```
      Derivation_tf.py --CA --inputDAOD_PHYSLITEFile DAOD_PHYSLITE.33081115._000042.pool.root.1 --outputD2AODFile test.pool.root --formats SKIM --skimmingString "count(AnalysisMuons.pt > (1 * GeV)) >= 1" --extraDataForDynamicConsumers xAOD::MuonContainer/AnalysisMuons
      ```
      The last argument should be a list of all containers touched by the skiming expression. The skimming string itself is the usual ExpressionEvaluation string, so identical to the strings used in the derivation framework. Per-event content should be identical to PHYSLITE.
      
      This would not be used in official production due to the generic format name. In cases where official production of a PHYSLITE skim were needed, one would need to define a new format as ususal. This could, however, be the basis of a user skimming service, if we decide to have such a thing in the future.
      7143fa0a
    • Peter Van Gemmeren's avatar
      Remove query from EventSelector · 9289c3b3
      Peter Van Gemmeren authored and Tadej Novak's avatar Tadej Novak committed
      Remove query from EventSelector
      9289c3b3
  9. Feb 06, 2025
  10. Feb 05, 2025
    • James Catmore's avatar
      Adding a WritexAOD example to AthExBasics · 897ae611
      James Catmore authored and Tadej Novak's avatar Tadej Novak committed
      Adding a WritexAOD example to AthExBasics
      
      We realised during the preparation of the Athena tutorial material that there isn't an example in AthenaExamples showing how an xAOD container should be created and written. This MR provides such an example, as a counterpart to the existing ReadxAOD example.
      
      This needs to be given a *very* close review given that this is supposed to be a template. In particular I'd like to check that everyone approves of the way this is done in the C++, including the use of the new operator... can this be avoided given our coding guidelines?
      897ae611
  11. Feb 04, 2025
  12. Feb 03, 2025
  13. Jan 31, 2025
  14. Jan 30, 2025
  15. Jan 29, 2025
    • Scott Snyder's avatar
      CxxUtils: Don't try to use the x87 fsincos instruction any more. · 323d99c3
      Scott Snyder authored and Johannes Elmsheuser's avatar Johannes Elmsheuser committed
      CxxUtils: Don't try to use the x87 fsincos instruction any more.
      
      Since a long time, we've had CxxUtils::sincos, which calculates sin and cos
      using the x87 fsincos instruction.  At the time, this was significantly
      better than the alternatives.  However, with current compilers on x86_64,
      it is in fact faster to use the sincos() library function than the x87 instruction.
      So remove the inlining bit.
      sincos() is still a GNU extension, so don't remove this entirely from CxxUtils ---
      this allow us to keep the #ifdefs in one place.
      323d99c3
  16. Jan 28, 2025
  17. Jan 27, 2025
Loading