| 657c62c | ferdymercury | 09 April 2024, 10:17:45 UTC | [tree] prevent crash in LoadCurrentDim Fixes https://its.cern.ch/jira/browse/ROOT-8269 Fixes https://its.cern.ch/jira/browse/ROOT-7993 | 12 April 2024, 19:48:09 UTC |
| 688a8af | Jonas Hahnfeld | 05 April 2024, 10:04:02 UTC | [ntuple] Use CommitSealedPagesV in RNTupleMerger This uses a bit more memory because all sealed pages of a cluster must be kept in memory, but is up to a factor 2x faster. (cherry picked from commit 05203aeefeaaeaa8efe1ca162fa6477d2b638c8a) | 12 April 2024, 13:05:38 UTC |
| 20d0f57 | Jonas Hahnfeld | 02 April 2024, 14:25:52 UTC | [ntuple] Implement incremental merging This fixes the test case with TBufferMerger. (cherry picked from commit 731890e612d1d29b804652428ccd8422e20d792d) | 12 April 2024, 13:05:38 UTC |
| 3e7cd9b | Jonas Hahnfeld | 02 April 2024, 13:19:53 UTC | [ntuple] Simplify RNTupleMerger::CollectColumns Take an RNTupleDescriptor directly and check columns iff the map is not empty. (cherry picked from commit 5e6ba2ae7029bb6ece5ce379aaea41b0db5fae3a) | 12 April 2024, 13:05:38 UTC |
| 66a55e3 | Jonas Hahnfeld | 02 April 2024, 14:24:47 UTC | [ntuple] Add RPageSink::GetDescriptor() (cherry picked from commit f1ab84526028a67d71e84908c0163fc084eed4c5) | 12 April 2024, 13:05:38 UTC |
| ec22949 | Jonas Hahnfeld | 02 April 2024, 12:09:07 UTC | [ntuple] Add RPageSink::IsInitialized() Keep track if Init() was called, and throw an error if it is called again. Requires renaming the current Init() method to InitImpl(). (cherry picked from commit d21642b6bd58dad509bff130e2c48869b59de1f0) | 12 April 2024, 13:05:38 UTC |
| 2e53ed2 | Danilo Piparo | 04 April 2024, 10:48:01 UTC | [core] Replace sprintf with snprintf which is more secure. | 11 April 2024, 06:03:50 UTC |
| 9eb4cd2 | Vincenzo Eduardo Padulano | 10 April 2024, 12:55:55 UTC | [tmva] Warn when using RBDT and xgboost Representing the current situation at https://github.com/root-project/root/issues/15197 | 10 April 2024, 15:56:02 UTC |
| 846b4db | ferdymercury | 10 April 2024, 08:10:27 UTC | [hadd] fix processing of indirect file when multiprocess is enabled (#14913) Fixes https://github.com/root-project/root/issues/14910 The full list of input files to merge is now created beforehand by accumulating both the direct and indirect ones (e.g. those found listed in file(s) signaled by the `@` syntax at the command line). Afterwards, the list is used either in the sequential or parallel merge depending on user input. --------- Co-authored-by: Philippe Canal <pcanal@fnal.gov> Co-authored-by: Vincenzo Eduardo Padulano <vincenzo.eduardo.padulano@cern.ch> | 10 April 2024, 14:44:26 UTC |
| 0802bcb | ferdymercury | 19 December 2023, 08:37:58 UTC | [skip-ci] document reasonable max bin content for th1f th2f th3f th1d th2d th3d Fixes https://github.com/root-project/root/issues/6671 clarify where the limit comes from in the docu use exact mantissa bit limit do not do safety checks in code, it just wastes CPU cycle because: 2/std::numeric_limits<float>::epsilon()+1 == 2/std::numeric_limits<float>::epsilon() yields true | 10 April 2024, 09:18:21 UTC |
| f0a585c | Mattias Ellert | 25 March 2022, 10:13:12 UTC | Byte swap values read from the protobuf raw data stream on big endian | 10 April 2024, 01:00:37 UTC |
| adf5498 | Jonas Rembser | 09 April 2024, 22:22:56 UTC | [TMVA] Avoid xgboost-cppyy symbol clashing problem in tmva101_Training XGBoost has to be imported before ROOT to avoid crashes because of clashing std::regexp symbols that are exported by cppyy. See also: https://github.com/wlav/cppyy/issues/227 | 10 April 2024, 01:00:37 UTC |
| 6103caa | Jonas Rembser | 09 April 2024, 22:18:45 UTC | [TMVA] Disable `rbdt` test Disabled because RBDT doesn't support the imbalanced tree structure of XGBoost models. | 10 April 2024, 01:00:37 UTC |
| 9f871c8 | Jonas Rembser | 09 April 2024, 21:41:38 UTC | [TMVA] Fix failing `rbdt_xgboost` test * avoid warnings with opened file that is not closed * don't assume the number of features is in the `_features_count` attribute (that one doesn't exist with xgboost 2.0) * support the `"reg:squarederror"` target, which is the default regression target in xgboost 2.0 | 10 April 2024, 01:00:37 UTC |
| 87a5152 | Jonas Rembser | 09 April 2024, 15:48:16 UTC | [CI] Add `xgboost` and `scikit-learn` to requirements This should be added so we can test RBDT in the CI. | 10 April 2024, 01:00:37 UTC |
| 4cac104 | Jonas Rembser | 09 April 2024, 13:56:42 UTC | [PyROOT][RDF] Support conversion of `bool` columns to NumPy arrays The `bool` columns in RDF are special, because the Take action returns a `std::vector<bool>`, which has an implementation-depended memory layout for space optimization. Therefore, I suggest supporting taking `bool` columns as `unsigned char` with `Take()`, such that in `RDataFrameAsNumpy` the values can be directly taken as bytes. This avoids superfluous copying in the code, and keeps the special logic in the pythonization side minimal. Closes #8639. | 10 April 2024, 01:00:37 UTC |
| d7b4074 | Vincenzo Eduardo Padulano | 08 April 2024, 13:05:48 UTC | [tmva] Adapt test to class layout changes | 10 April 2024, 01:00:37 UTC |
| c7fd1a1 | Vincenzo Eduardo Padulano | 08 April 2024, 12:24:55 UTC | [tmva][NFC] Apply clang-format to changed files | 10 April 2024, 01:00:37 UTC |
| bb996a0 | Vincenzo Eduardo Padulano | 08 April 2024, 12:17:16 UTC | [tmva] Cleanup RModel classes layout * Apply rule of five * Use correct `ClassDef` macro only when necessary * Remove relationship to `TObject` from `RModel_Base` * Properly implement virtual hierarchy for `RModel_GNNBase` and derived | 10 April 2024, 01:00:37 UTC |
| 0b78b8b | Jonas Rembser | 06 April 2024, 12:05:41 UTC | [tmva][sofie] Fix issues with RModel streamer * remove wrong multiplication of type size in `CastPersistentToShared` * restructure `InitializeTensor` to only provide read-only accessors * avoid memory leak of persistified data when reading back from file * reset `fSize` member in `CastSharedToPersistent` to avoid wrong size info when persistifying the same model twice (otherwise, the multiplication with the element size in bytes would happen twice, resulting in overruns when copying the transient data) Extends on #15162. | 10 April 2024, 01:00:37 UTC |
| e4ecfd5 | ferdymercury | 02 April 2024, 13:40:31 UTC | [hist] add bounds and overflow checking See https://root-forum.cern.ch/t/tprincipal-fails-for-a-very-large-number-of-entries/58795 | 10 April 2024, 01:00:37 UTC |
| 8227c2e | ferdymercury | 02 April 2024, 13:28:18 UTC | [skip-ci,hist] improve doc wrt covariance matrix format | 10 April 2024, 01:00:37 UTC |
| fecc7db | Jonas Rembser | 08 April 2024, 01:59:21 UTC | [CMake] Don't mention C++17 requirement for `root7` and `webgui` flags ROOT requires C++17 anyway, so this is redundant and can be confusing. | 10 April 2024, 01:00:37 UTC |
| 72c816a | ferdymercury | 08 April 2024, 01:51:56 UTC | [hist] Use virtual THnBase::CreateIter to generalize Fixes https://github.com/root-project/root/issues/13962 | 08 April 2024, 06:35:28 UTC |
| 8273c93 | ferdymercury | 06 April 2024, 08:24:34 UTC | [core] allow also normal shared libraries without the ACLiC suffix with a main function equal to the shared library name as suggested by pcanal [core] mention user shared lib alternative | 08 April 2024, 06:34:47 UTC |
| 0382cb8 | ferdymercury | 24 February 2024, 17:13:16 UTC | [cling] Allow running named macro from shared library As it was the case with CINT and was documented in users guide. See https://root-forum.cern.ch/t/cannot-run-shared-object-in-batch-mode/58213 Fixes https://github.com/root-project/root/issues/14772 [core] document run-shared-object behavior [core] only handle special sharedLibrary when running, not when loading | 08 April 2024, 06:34:47 UTC |
| 4e30f26 | Jonas Rembser | 26 February 2024, 13:01:49 UTC | Remove the deprecated Pythia 6 interface The Pythia 6 interface was deprecated in ROOT 6.30 and scheduled for removal in 6.32. As we are now in the 6.32 development cycle, everything related to `pythia6` can be removed. | 05 April 2024, 19:30:00 UTC |
| 6f4b749 | Jonas Rembser | 05 April 2024, 13:21:19 UTC | [CMake] Fix ordering problem when checking and setting `cuda` flag In the ROOT build system, `CheckCompiler` is included before `RootBuildOptions`. The problem with this is that when building with `-Dall=ON`, the `cuda` flag is only set in `RootBuildOptions`, which is too late to find the CUDA compiler in `CheckCompiler`. This results in issues like #15139: ``` CMake Error in roofit/batchcompute/CMakeLists.txt: Exporting the target "RooBatchCompute_CUDA" is not allowed since its linker language cannot be determined ``` This commit suggests a solution to this problem: just always enable the CUDA language in CMake if it is found, and only set the `cuda` flag with `-Dall=ON` if the CUDA language is enabled. | 05 April 2024, 19:30:00 UTC |
| b47683e | Vincenzo Eduardo Padulano | 04 April 2024, 13:58:35 UTC | [pyroot] Use raw string to allow escaped character in docstring To avoid warnings such as `_rvec.py:95: SyntaxWarning: invalid escape sequence '\p'` | 05 April 2024, 19:30:00 UTC |
| a990e92 | Jonas Rembser | 05 April 2024, 01:10:06 UTC | [CMake] CMake dependency on Python: use targets Closes #6753. | 05 April 2024, 19:30:00 UTC |
| 2c6d93b | ferdymercury | 03 February 2024, 12:01:27 UTC | [math] Clarify meaning of TFoam::GetIntNorm parameters Fixes https://github.com/root-project/root/issues/14528 | 05 April 2024, 19:30:00 UTC |
| 988b7c4 | Danilo Piparo | 03 April 2024, 21:55:52 UTC | [hist] Add test for TSpline precision when saved as macro. | 05 April 2024, 09:29:14 UTC |
| 1159b6e | ferdymercury | 27 March 2024, 14:42:11 UTC | [hist] Improve precision of TSpline SaveAs Fixes https://github.com/root-project/root/issues/12091 | 05 April 2024, 09:29:14 UTC |
| b560c3c | ferdymercury | 05 April 2024, 00:16:21 UTC | [io] do not use single-file-merge shortcut if filters are specified Fixes https://github.com/root-project/root/issues/13359 | 05 April 2024, 09:27:23 UTC |
| 78960f1 | ferdymercury | 05 April 2024, 00:20:39 UTC | [gui] Do not reset FitPanel params when ADD on top of prefit function Fixes https://github.com/root-project/root/issues/11021 | 05 April 2024, 08:57:37 UTC |
| 271199b | Danilo Piparo | 04 April 2024, 11:22:01 UTC | [ci] Specify v6-32-00-patches branch in the root-ci.yaml | 04 April 2024, 11:22:01 UTC |
| d34aeee | ferdymercury | 30 March 2024, 09:39:25 UTC | [hist] prevent nullptr access | 04 April 2024, 10:08:19 UTC |
| eb14d32 | ferdymercury | 11 March 2024, 09:08:16 UTC | [skip-ci] cross-reference markdown docu with link as well as human-readable as requested by pcanal | 04 April 2024, 10:02:20 UTC |
| 30f1c8a | ferdymercury | 11 March 2024, 09:00:55 UTC | [skip-ci] [io] Collapse data format details of TFile class. Supersedes https://github.com/root-project/root/pull/9616 | 04 April 2024, 10:02:20 UTC |
| ba91344 | Danilo Piparo | 04 April 2024, 07:10:41 UTC | [ci] Remove macOS 12 from the tested platforms | 04 April 2024, 08:05:18 UTC |
| 6e27bd1 | Danilo Piparo | 04 April 2024, 06:44:44 UTC | [doc] Update reference guide build procedure | 04 April 2024, 06:44:44 UTC |
| 43bf8e8 | Danilo Piparo | 04 April 2024, 06:37:50 UTC | Update ROOT version file | 04 April 2024, 06:37:50 UTC |
| 7784644 | ferdymercury | 27 March 2024, 10:06:56 UTC | [hist] fix similar bug in TSpline5_saveas | 03 April 2024, 20:07:33 UTC |
| 05eec86 | ferdymercury | 27 March 2024, 10:02:17 UTC | [hist] fix bug in TSpline3 save-as code Fixes https://github.com/root-project/root/issues/13441 | 03 April 2024, 20:07:33 UTC |
| 71c812a | tmadlener | 03 April 2024, 08:29:38 UTC | Require only one template argument for usage Defaulting the rest to the first in case functions are used where argument types are not deduced, e.g. in RDataFrames | 03 April 2024, 19:43:29 UTC |
| 5c6f43a | tmadlener | 02 April 2024, 13:35:02 UTC | Replace static_asserts with gtest macros | 03 April 2024, 19:43:29 UTC |
| 2efaa3e | tmadlener | 02 April 2024, 11:12:09 UTC | Add basic tests for InvariantMass functionality | 03 April 2024, 19:43:29 UTC |
| 750d66a | tmadlener | 02 April 2024, 10:48:26 UTC | Add basic tests for DeltaR | 03 April 2024, 19:43:29 UTC |
| 8d9ae94 | tmadlener | 02 April 2024, 09:24:27 UTC | Add basic tests for DeltaPhi | 03 April 2024, 19:43:29 UTC |
| 5a744c3 | tmadlener | 28 March 2024, 12:16:09 UTC | Make inv mass calculations accept different floating point types | 03 April 2024, 19:43:29 UTC |
| 37c9476 | tmadlener | 28 March 2024, 11:53:21 UTC | Allow to use DeltaR with different floating point types | 03 April 2024, 19:43:29 UTC |
| 4200741 | tmadlener | 28 March 2024, 11:44:19 UTC | Make DeltaPhi work with different floating point types | 03 April 2024, 19:43:29 UTC |
| 0facd4a | Bertrand Bellenot | 03 April 2024, 13:08:53 UTC | [CI][windows] Disable failing tests on Jenkins | 03 April 2024, 15:30:43 UTC |
| 3b8a908 | Bertrand Bellenot | 03 April 2024, 14:59:02 UTC | [cmake][windows] Don't install system DLLs (#15114) * [cmake][windows] Don't install system DLLs Fix Jira issue ROOT-10926 https://its.cern.ch/jira/browse/ROOT-10926 * [skip-ci] Update cmake/modules/RootCPack.cmake Co-authored-by: ferdymercury <ferdymercury@users.noreply.github.com> --------- Co-authored-by: ferdymercury <ferdymercury@users.noreply.github.com> | 03 April 2024, 14:59:02 UTC |
| 6e67978 | Danilo Piparo | 03 April 2024, 09:59:22 UTC | [skip-ci][ci] Skip matrix overrides in case of workflow_dispatch events, which is what is triggered when releasing. | 03 April 2024, 10:00:33 UTC |
| 8d2ce49 | Andrei Rabusov | 02 April 2024, 09:55:06 UTC | [minuit2] make letter case consistent | 03 April 2024, 07:54:33 UTC |
| bea7e2d | Jonas Rembser | 02 April 2024, 10:18:03 UTC | [PyROOT] Remove `PyzCppHelpers` Most of these helpers were single-use, so it's better to inline them in the implementations. | 03 April 2024, 07:48:47 UTC |
| c85ae33 | Jonas Rembser | 02 April 2024, 09:38:58 UTC | [PyROOT] Fix infinite loop in `TMemoryRegulator::ClearProxiedObjects()` There is an infinite loop in `PyROOT::TMemoryRegulator::ClearProxiedObjects()`: ```C++ while (!fObjectMap.empty()) { auto elem = fObjectMap.begin(); auto cppobj = elem->first; auto klassid = elem->second; auto pyclass = CreateScopeProxy(klassid); auto pyobj = (CPPInstance *)MemoryRegulator::RetrievePyObject(cppobj, pyclass); if (pyobj && (pyobj->fFlags & CPPInstance::kIsOwner)) { ... } else { // Non-owning proxy, just unregister to clean tables. // The proxy deletion by Python will have no effect on C++, so all good MemoryRegulator::UnregisterPyObject(pyobj, pyclass); } } ``` In the second code branch, the object is not removed from the `fObjMap`, if `UnregisterPyObject` doesn't call the unregister hook because it quits early. This can happen if there is not C++ object corresponding to the regulated python object, causing an infinite loop because `fObjectMap` never gets empty. Although this seems like an obvious logic error, this was only noticed after the cppyy upgrade. Probably the code path was not hit before. Addresses https://github.com/root-project/root/issues/15085#issuecomment-2031363812. | 03 April 2024, 07:48:47 UTC |
| 1c67cd3 | ferdymercury | 02 April 2024, 16:04:27 UTC | [skip-ci,math] Specify that returned weight is not normalized Fixes https://github.com/root-project/root/issues/14765 | 03 April 2024, 05:07:57 UTC |
| 845a003 | Olivier Couet | 02 April 2024, 15:20:32 UTC | implement missing IsInside() (#15074) * implement missing IsInside() * Add missing `<CR>`s * clang format * More clang format * Final adjustment | 02 April 2024, 15:20:32 UTC |
| ef6b35e | Jakob Blomer | 02 April 2024, 12:09:59 UTC | [ntuple] fix CyclicCollectionProxy inheritance (tests) Co-authored-by: Jonas Hahnfeld <hahnjo@hahnjo.de> | 02 April 2024, 13:32:43 UTC |
| 41701b1 | Jakob Blomer | 02 April 2024, 12:09:14 UTC | [NFC][ntuple] fix typo in code comment Co-authored-by: Jonas Hahnfeld <hahnjo@hahnjo.de> | 02 April 2024, 13:32:43 UTC |
| f102515 | Jakob Blomer | 26 March 2024, 10:25:48 UTC | [ntuple] fix detection of cyclic collection proxies | 02 April 2024, 13:32:43 UTC |
| 9950a44 | Jonas Hahnfeld | 02 April 2024, 09:04:23 UTC | [llvm-project] Bump to tag ROOT-llvm16-20240402-01 | 02 April 2024, 11:11:39 UTC |
| 07021ea | Jonas Hahnfeld | 18 March 2024, 14:34:08 UTC | [clang] Reuse ContextRAII for ContextAndScopeRAII | 02 April 2024, 11:11:39 UTC |
| b702391 | Jonas Hahnfeld | 18 March 2024, 13:56:27 UTC | [clang] Correctly RAII reset two "new" Sema fields FunctionScopesStart and InventedParameterInfosStart were introduced by upstream commit https://github.com/llvm/llvm-project/commit/5c845c1c50ac89a6f12557d1571678f3d1432478 and are handled in ContextRAII, but not our ContextAndScopeRAII. | 02 April 2024, 11:11:39 UTC |
| 8227e95 | Jonas Rembser | 30 March 2024, 16:18:14 UTC | [RF] Fixup work around deleted move constructor in `std::span` backport This is a followup to #15089, where I missed a second change. Closes #15087. | 01 April 2024, 12:07:40 UTC |
| 62a7ea3 | Vincenzo Eduardo Padulano | 28 March 2024, 14:44:31 UTC | [df] Correctly clone AsNumpyResult in distributed execution | 28 March 2024, 23:06:59 UTC |
| 1edca86 | Vincenzo Eduardo Padulano | 28 March 2024, 14:43:20 UTC | [df] Enable cloning of AsNumpyResult | 28 March 2024, 23:06:59 UTC |
| dcebd46 | Jonas Rembser | 28 March 2024, 16:17:58 UTC | [RF] Work around deleted move constructor in `std::span` backport Closes #15087. | 28 March 2024, 16:51:10 UTC |
| a27e60a | Jonas Rembser | 26 March 2024, 23:53:59 UTC | [RF] Improve parameter management in `RooEvaluatorWrapper` Only register actual observables as value servers in the `RooEvaluatorWrapper` this avoids that spurious variables appear in user code, which were only meant to be used internally by the new CPU evaluation backend. To make the management of the parameter easier, impose the new reasonable constraint that when resetting the data via `setData()`, the new dataset needs to have the same structure as the old one. It was confirmed with an ATLAS model provided by Will that this commit is fixing the problem of spurious variables. | 28 March 2024, 16:49:58 UTC |
| f4f6467 | Jonas Rembser | 28 March 2024, 16:48:54 UTC | [PyROOT] Replace `TDirectory.__getattr__` with `__getitem__` The new recommended way to get objects from a `TFile` or any `TDirectory` in general is now via `__getitem__`: ```python tree = my_file["my_tree"] # instead of my_file.my_tree ``` This is more consistent with other Python collections (like dictionaries), makes sure that member functions can't be confused with branch names, and easily allows you to use string variables as keys. The old pythonization with the `__getattr__` syntax still works, but emits a deprecation warning and will be removed from ROOT 6.34. | 28 March 2024, 16:48:54 UTC |
| 211d1e5 | Jonas Hahnfeld | 27 March 2024, 16:31:49 UTC | [tree] Add regression test for writing long strings | 28 March 2024, 15:22:05 UTC |
| 638593e | scott snyder | 06 March 2024, 03:06:50 UTC | Fix writing long strings. In TBufferFile::WriteFastArrayString, we had ``` if (n < 255) { *this << (UChar_t)n; } else { *this << (UChar_t)255; *this << n; } ``` A recent commit changed the type of the n parameter from Int_t to Long64_t. This is effectively an incompatible change in the on-disk format, but only for strings which are at least 255 characters long. Further, ReadFastArrayString is still reading an Int_t, so this version of ROOT cannot read files that it writes. Resolve by changing WriteFastArrayString to explicitly write an Int_t. Also move the bounds check on the length to before writing anything into the buffer. Fixes a failure seen in the unit tests of the ATLAS EventInfo package in the dev3LCG build. | 28 March 2024, 15:22:05 UTC |
| ac73d30 | Jakob Blomer | 28 March 2024, 07:58:13 UTC | [ntuple] speed-up temperatures tutorial | 28 March 2024, 12:53:32 UTC |
| 727181e | Jonas Rembser | 27 March 2024, 11:35:05 UTC | [cppyy] Fix `missing-field-initializers` with Python 3.12 | 28 March 2024, 10:04:57 UTC |
| d8330ab | ferdymercury | 26 March 2024, 14:18:05 UTC | [hadd] avoid large RAM in hadd with multiple subdirs Fixes https://github.com/root-project/root/issues/13531 | 28 March 2024, 00:12:48 UTC |
| 7a4b12f | Jonas Rembser | 27 March 2024, 15:14:41 UTC | [RF] Fix dirty state propagation for offsetting with new CPU backend The logic that determined the offset hiding of not was coded inside the `RooNLLVarNew` evaluation function so far. This caused trouble, because a change in the global `RooAbsReal::hideOffset()` state did not mark the NLL as dirty. Therefore, it was unpredictable if the offset was actually hidden or not. This commit suggests an improved logic: * Reducer nodes like the NLL always register a value and an offset to the `EvalContext` * The evaluator decides whether to subtract the offset or not * A change in `hideOffset()` makes the evaluator wrapper set all reducer nodes to dirty A new unit test to cover this was also implemented | 27 March 2024, 17:18:52 UTC |
| 7365f92 | Jonas Rembser | 25 October 2023, 14:47:46 UTC | [RF] Improve the signature of vectorized evaluation function So far, the signature for the function that is called for the vectorized evaluation was this one: ```c++ void RooAbsReal::computeBatch(double* output, size_t nEvents, RooFit::Detail::DataMap const& dataMap) const ``` This commit is suggesting a new signature: ```c++ void doEval(RooFit::EvalContext & ctx) const; ``` The idea is to make the signature as short as possible, so it doesn't have to be changed anymore if more information needs to be passed. That's why the only parameter is now an `EvalContext` object, reminicint of the old `RunContext` object that fulfilled this task in the very first implementation of the BatchMode by Stephan. The name is now simply `doEval`, because the overloaded term "Batch" should be dropped. It needed to be something with "evaluate", because there is also `RooAbsReal::evaluate()` and we are talking about "evaluation backends". The motivation to change this interface now is because I want to write a documentation for developers (like CMS combine mainteiners) on how to use these new interfaces. And if they start to use it, the interfaces should not change anymore. Than's why I'm doing this change now, which I had in mind already for some time. | 27 March 2024, 15:24:58 UTC |
| 7a4af49 | Bertrand Bellenot | 27 March 2024, 08:43:44 UTC | [cmake][tests] Fix and re-enable tests on Windows - Export a couple of missing symbols (x86) - Re-enable several tests working now with LLVM 16 and VS 2022 - Reorganize `llvm13_broken_tests` and `win_broken_tests` for clarity | 27 March 2024, 12:05:12 UTC |
| 60f7275 | Jerry Ling | 26 March 2024, 07:26:53 UTC | [NFC][ntuple] clarify in spec that encoding is page-wise * [RNTuple] describe how exactly `Index` columns are "relative to the cluster" page-by-page follow up of #14949 * mention encoding is page-wise | 26 March 2024, 07:26:53 UTC |
| d93a89a | Jonas Rembser | 26 March 2024, 06:44:11 UTC | [RF] Also veto Python version of TestNonCentral when `mathmore=OFF` | 26 March 2024, 07:06:44 UTC |
| 2a82db0 | Jonas Rembser | 25 March 2024, 21:48:28 UTC | [RF] Always reference data in workspace in `rs101_limitexample` Otherwise, there is no clear ownership model, which might result in PyROOT crashes at cleanup. | 25 March 2024, 23:14:11 UTC |
| dfff4b1 | Jonas Rembser | 25 March 2024, 17:28:08 UTC | [RF] No manual `del` in RooFit/RooStats tutorials The new cppyy has improved memory management with lifelines, so the manual calls to `del` to get the right deletion order are not necessary anymore. Closes #10891. | 25 March 2024, 23:14:11 UTC |
| 4dcc0fd | jolly-chen | 25 March 2024, 20:40:52 UTC | [RF] Convert 4 RooStat tutorials to PyROOT Contributes to https://github.com/root-project/root/issues/8758 Co-authored-by: Jonas Rembser <jonas.rembser@cern.ch> | 25 March 2024, 20:40:52 UTC |
| bbd5eb9 | Jakob Blomer | 25 March 2024, 15:22:19 UTC | [ntuple] forbid user classes with virtual inheritance | 25 March 2024, 20:40:17 UTC |
| 4998acf | Jakob Blomer | 25 March 2024, 13:12:55 UTC | [ntuple] add unit test for virtual inheritance | 25 March 2024, 20:40:17 UTC |
| 8823081 | Philippe Canal | 25 March 2024, 15:29:38 UTC | Add note of the link between TObject and k630forwardCompatibility. | 25 March 2024, 19:19:14 UTC |
| d7281c0 | Philippe Canal | 20 March 2024, 20:58:29 UTC | [io] Add rootrc key TFile.v630forwardCompatibility The default is no. | 25 March 2024, 19:19:14 UTC |
| 6045f02 | Philippe Canal | 18 March 2024, 21:51:58 UTC | Allow producing forward compatible file for fBits value. Calling `file->SetBit(TFile::k630forwardCompatibility);` will request to file to store the current value of the bit kIsOnHeap and kNotDeleted for reading in older version that were not setting their value based on the actual state of the read into object | 25 March 2024, 19:19:14 UTC |
| f3668d9 | Jonas Rembser | 21 March 2024, 15:04:15 UTC | [RF] Re-implement `MakeNumpyDataFrame` with no interpreter calls Re-implement `MakeNumpyDataFrame` with no interpreter calls, and avoid a explicit lifetime management of Python objects in the RDNumpyDS helper. Like this, the former `RNumpyDS.hxx` helper could be moved to `tree/dataframe`, as it doesn't import `Python.h` anymore. The `RNumpyDS` also got renamed to `RVecDS`, because it takes RVecs and not NumPy arrays directly. | 25 March 2024, 17:02:37 UTC |
| 00cd9fc | Danilo Piparo | 24 March 2024, 15:14:26 UTC | [ci] Contextualize cmake warning about builtin_afterimage=OFF since the warning is not needed in case the graphics system, cocoa or x11, is OFF. Fixes #15027 | 25 March 2024, 14:23:23 UTC |
| 6e3f5d3 | Vassil Vassilev | 25 March 2024, 08:06:36 UTC | [runtime_cxxmodules] Allow for changing the modulemap name to match the overlay. Fixes #14964 | 25 March 2024, 13:49:15 UTC |
| ac661a8 | Danilo Piparo | 24 March 2024, 14:55:26 UTC | [DF][NFC] Mention AsRNode in the progress bar documentation fixes #15040 | 25 March 2024, 11:10:45 UTC |
| 239cafe | ferdymercury | 22 March 2024, 15:19:45 UTC | Clarify wording Thanks vepadulano! Co-authored-by: Vincenzo Eduardo Padulano <v.e.padulano@gmail.com> | 23 March 2024, 07:22:48 UTC |
| 5db7764 | ferdymercury | 20 March 2024, 12:33:30 UTC | [hist] explain better what computeIntegral does Fixes https://its.cern.ch/jira/browse/ROOT-8274 | 23 March 2024, 07:22:48 UTC |
| a29e81c | ferdymercury | 20 March 2024, 09:44:10 UTC | [skip-ci] brief is for file, not for namespace In https://root.cern.ch/doc/master/namespaces.html the description of the ROOT namespace was picked form this file brief. Try to fix it this way. | 22 March 2024, 13:17:25 UTC |
| 6b6ef47 | ferdymercury | 21 March 2024, 09:34:49 UTC | fix while | 22 March 2024, 13:16:28 UTC |
| 63dbab1 | ferdymercury | 21 March 2024, 09:22:24 UTC | improve error messages and simplify for loop vepadulanos suggestions | 22 March 2024, 13:16:28 UTC |
| 6f894da | ferdymercury | 20 March 2024, 13:52:06 UTC | fix include | 22 March 2024, 13:16:28 UTC |
