SHERPA-MC-2.2.5-bp150.1.3

- Update to 2.2.5:
  * Interface changes
  - Add preliminary support for HepMC3 (this interface is still
    being tested)
  * New features
  - Improve decay showering off top quarks
  - Improve single-top support and add examples and documentation
    for its production channels
  - Improve hadronisation and hadron decays for heavy flavour
    production
  - Introduce a user hook framework which allows users to easily
    implement
    event generation afterburners, see Examples/Userhook
  - Add settings to output ME-only on-the-fly variations (along with
    the full ones):
    HEPMC_INCLUDE_ME_ONLY_VARIATIONS=1 (HepMC interface)
  - New cut-off for parton-shower reweighting:
    CSS_REWEIGHT_SCALE_CUTOFF=<pt_min>, default: 5 GeV2
    this is more generally applicable than using a cut-off on the
  number of reweighted emissions
  - Improve support for on-the-fly variations of clustering steps
    (reweight each step locally instead of doing it globally)
  * Bugfixes
  - Fix issue with the overweight treatment for negative weights
  - Add missing shower weight in some event outputs, e.g. when using
    `OUTPUT HepMC...[]`
  - Fix issue with the overweight threshold for on-the-fly variations
  - Fix issue with non-trivial beam spectra handling
  - Fix possible issues with electron-"PDF" initialization
  - Fix issue with the ISR-channel threshold for leptonic collider
  - Fix X/Y functions in METOOLS (this is relevant for hadron decays)
  - Fix crash in decay-channel sorting
  - Fix crash when using EVENT_TYPE=HadronDecay
  - Fix issue with CSS strong coupling values when using non-trivial
    scale factors
  - Fix issue with QSF variations for real emission calculations with
    COMIX
  - Fix crash when doing on-the-fly variations and invalid dipole
    kinematics occur
  - Fix issue in the interplay of the flux and the BIWeight for 1 -> n
    processes
  - Fix issue when using vertices with a vanishing coupling
  - Fix rare "invalid amplitude" crashes
  - Fix issues when adding limits for the number of t-channel propagators
    in matrix elements
  - Fix issue with checking PDF availability when enforcing jet vetoes
  - Fix several compilation issues
- Add patch to fix ARM (and PPC) builds:
  * SHERPA-MC-fix_arm_build.patch
- Refresh partially upstreamed patch:
  * SHERPA-MC-no-return-in-non-void-function.patch

- Upstream released a new 2.2.4 tarball, so refresh sources.

- Update to version 2.2.4:
  * Interface Changes:
  - New default parameters for hadronisation model to improve
    heavy flavour production fractions and meson/baryon ratios.
  - Updates to Recola interface with some new parameter (names),
    please refer to the manual.
  - Make STRICT_METS the default scale setter to guarantee full
    scale setting accuracy for unweighted events.
  * New features
  - Availability of DIRE NLO parton shower (cf.
    arXiv:1705.00742). This is not (yet) the default shower, in
    particular not in NLO-merged configurations.
  - Higher numerical stability in parton shower kinematics
    (re)construction.
  - Add OVERWEIGHT_THRESHOLD allowing user to cap maximal weight
    during event generation (cf. manual).
  - Avoid spuriously large weights in NLO-merged samples due to
    instable PDF ratios.
  - Add interface for hadron-level selectors.
  - Add Max_N_TChannels option in Amegic.
  - Add possibility to identify partons which originate in the
    hard decay handler.
  - Restructure gzip support such that it can be switched off at
    runtime (USE_GZIP).
  - Enable on-the-fly reweighting for calculations with the
    Higgs (Interference) add-on for HIGGS_INTERERENCE_MODE 1, 2,
    3 or 4.
  * Bugfixes
  - Bugfix in NLO-matched parton shower emission for non-trivial
    colour structures like in V+2j@NLO. This fixes an eta
    asymmetry which was visible at the ~2% level in such
    samples.
  - BSM models through UFO: fix running coupling behaviour and
    treatment of exotic color octets.
  - Bugfix in random generator for Poisson distributions with
    large mean values.
  - Bugfixes for segfaults and compiler warnings with gcc 7.1.
  - Bugfix for colour assignment in clustering for loop-induced
    merged processes.
  - Bugfix for decay showering in conjunction with NLO+PS
    matching.
- Change BuildRequires HepMC-devel to HepMC2-devel: not compatible
  with HepMC >= 3.0.
- Rename libSHERPA-MC-config to libSHERPA-MC0-config that follows
  the shared lib versioning. In addition the ld.so.conf.d file is
  also renamed to follow the shared lib version to make it
  possible to install multiple versions of the shared lib.
- Rebase SHERPA-MC-no-return-in-non-void-function.patch and add
  fixes for one more function showing the same issue in version
  2.2.4.
- GCC >= 6.0 required, build fails with older GCC for openSUSE <=
  1320.

- Update to version 2.2.3:
  * Support for the Recola one-loop generator for QCD corrections.
- Add SHERPA-MC-no-return-in-non-void-function.patch -- Fix a
  non-void (bool) function that was not returning any data to
  return "true".

- Shared library package must only contain time-unique names;
  split off libSHERPA-MC-config.

- use individual libboost-*-devel packages instead of boost-devel

- Update to version 2.2.1:
  * Incoporation of approximate NLO EW corrections in MEPS@NLO
    simulation as described in arXiv:1511.08692
  * On-the-fly renormalisation and factorisation scale, PDF and
    alpha_s variations in the matrix elements for LO, LOPS, NLOPS,
    MEPS, MENLOPS and MEPS@NLO simulations as described in
    arXiv:1606.08753
- Changes from version 2.2.0: See
  https://sherpa.hepforge.org/trac/wiki/SherpaDownloads/Sherpa-2.2.0.
- Drop gcc6_char_cast.patch; incorporated upstream.

- RPM group classification
- run install_info in preun to avoid errors of missing files
- Shared library subpackage is supposed to include the number

-  new gcc6_char_cast.patch

- License tag in spefile corrected, should be GPL-2.0+ and
  GPL-3.0.

- Update to version 2.1.1 (Bug-fix release):
  + A bug related to MC@NLO has been fixed. This will affect the
    total cross section in heavy quark production at the level of
    few permille.
  + A bug related to parton showering in associated Higgs boson
    production has been fixed.
  + A problem related to SQLite database locking on IBM BG/Q
    systems has been solved.
  + A bug related to CS subtraction with Comix has been fixed.
    This will eliminate possible large weights in events
    containing extremely soft gluons.
  + A bug related to the QED corrections on the hard interaction
    accidentally being switched off has been fixed.
  + A bug related to the contents of the HepMC::PdfInfo? object
    in MC@NLO event generation has been fixed.
  + A bug in the underlying event simulation, presenting as a final
    state particle asymmetry, has been fixed. An update of the
    underlying event parameters was necessary, invalidating MPI
    grids generated with previous versions.
  + The Python interface to access Sherpa's matrix elements has
    been updated and supplemented with a C++ interface.
  + A bug in the CS subtraction in DRED has been fixed.
- Changes from version 2.1.0:
  + Introduced SQLite as the preferred storage manager
  + Spin correlated decay chains (Comix only), includes all 1->2
    and 1->3 decays allowed by the model, iterated to arbitrary
    decay chain length, all decay steps are dressed with QCD
    parton showers and QED corrections
  + MC@NLO H events with Comix, speeds up initialisation and
    runtime
- Packaging changes:
  + Disable rivet bindings until SHERPA-MC builds with latest
    Rivet >= 2.1.2
  + BuildRequires on sqlite3-devel introduced since version 2.1.0
  + Package new binaries init_nlo.sh, make2scons, plot_graphs.sh.

- Initial version.

- Add Git-Clone URL

- Add SHERPA-MC-ignore-distutils-deprecation-warning.patch:
  Disable deprecation warning when importing distutils from
  python3.10, so that autotools checks can still work.

- Make sure to provide/obsolete the python- subpkg to support
  migration

- Update to version 2.2.12:
  * Significant speed-up of (partially) unweighted event
    generation with on-the-fly variations enabled, e.g. for many
    PDF variations or implicitly when EWvirt associated
    contributions are evaluated !410
  * New option to perform (significantly faster) uncorrelated
    mc@nlo matching after unweighting, enabled by flag
    NLO_CSS_PSMODE: 2 !410
  * Further performance improvements !410
  * New MCFM interface (bd3e25ad, TODO: @shoeche to provide more
    info) !409
  * CSS improvements (efcffda4, TODO: @shoeche to provide more
    info) !410
  * New output options steered through BATCH_MODE !410 (see Manual
    for additional information):
  - 8 prints the current process name,
  - 16 prints the time elapsed/left in unformatted seconds
  * Runtime issues: Fix compatibility issue when using binary
    relocation after installation !453
  * Improve compatibility with autoconf 2.70 or later (in
    particular when building process libraries without scons) !396
  * Fix MadLoop 5 interface add-on installation
  * Fix Recola interface add-on compilation !410

- Update to version 2.2.11:
  * Improvements in ME+PS merging for unordered scale
    configurations !381
  * New NLO_SUBTRACTION_SCHEME option for improved behavior of the
    parton shower in processes with large higher-order
    contributions from t-channel gluon exchange !382
  * Additional power corrections to better match soft limit !394
  * Implement csin2_thetaW for SMEFTsim !319
  * Exponentiation version of EWvirt corrections !348
  * Phase space enhancing with function for unweighted event
    generation (Enhance_Function) !357 !379
  * Rivet interface improvements: zipped YODA files (!374) and
    split pos/neg weights (!376)
  * Runtime issues
  - Bugfix in Comix for models with vertices with negative
    coupling orders !327
  - Bugfixes for crashes with new SPECIAL_TAU_SPIN_CORRELATIONS
    option !335 !362
  - Bugfix for crash with spin correlations in ttll setups !388
  - Bugfixes for non-standard KFACTOR treatments !352
  * Compilation issues
  - Fix build problem with very new glibc !328
  - Updated swig requirement for python interface to 2.0.12 !342
  - Improvements in HepMC3 linking !380
  - MCFM interface fixes !389
- Drop SHERPA-MC-build-with-recent-glibc.patch: incorporated
  upstream.
- Add BuildRequires: libtool needed for building swig-based
  bindings for updated version.

- Add SHERPA-MC-build-with-recent-glibc.patch to fix builds
  against recent glibc; patch taken from upstream git.

- Drop post(un) install-info scripts for devel pkg, these are
  handled by rpm file triggers; also drop the appropriate
  dependencies.

- Update to version 2.2.10:
  * Further improvements for Rivet3 and HepMC3 compatibility
    [!321, !314].
  * Bugfix for crashes with new heavy-flavour shower evolution
    schemes [!313].
  * Bugfix for crash in ReweightWithoutSubevents [!318].
  * Bugfix for documentation signal blob with pre-NLO+PS
    kinematics [!320].
  * Bugfix for crash in setups with neutron beams [!312].
  * More robust discovery of HZTool and HepMC3 ROOT IO libraries
    [!317, !322, !323].
- Changes from version 2.2.9: See
  https://gitlab.com/sherpa-team/sherpa/-/tags/v2.2.9.
- Build with Rivet support (BuildRequires: Rivet-devel).
- Fix some incorrect Provides and Obsoletes in subpackages.
- Execute hashbang cleanup on source files instead of acting on
  installed files in %{buildroot}.
- Run spec-cleaner for minor formatting cleanups.