54a757a | Axel Naumann | 03 March 2022, 06:39:52 UTC | "Update ROOT version files to v6.26/00." | 03 March 2022, 06:39:52 UTC |
f46a182 | Axel Naumann | 02 March 2022, 14:55:32 UTC | [relnotes] Add latecomer issue. | 03 March 2022, 06:39:09 UTC |
bd0718d | moneta | 02 March 2022, 10:46:06 UTC | [TMVA] Use a unique XML file for the trained model used in the test This should fixe some sporadic failures observed in the CI builds | 02 March 2022, 22:40:08 UTC |
1687a88 | Enrico Guiraud | 02 March 2022, 17:56:57 UTC | [DF] Have RColumnRegister own RLoopManager In turn it owns RDefines that need the RLoopManager to stay alive until they are deleted. Co-authored-by: Jonas Hahnfeld <jonas.hahnfeld@cern.ch> | 02 March 2022, 21:53:24 UTC |
7ca3a6b | Jonas Hahnfeld | 02 March 2022, 17:28:59 UTC | [DF] Delete column register before RLoopManager goes out of scope | 02 March 2022, 21:53:24 UTC |
2fc3000 | Enrico Guiraud | 02 March 2022, 17:21:48 UTC | [DF] RInterface: add missing defaulted move-assignment op | 02 March 2022, 21:53:24 UTC |
10eb8c0 | Axel Naumann | 02 March 2022, 11:39:50 UTC | [relnotes] Add v6.26/00 issue list. | 02 March 2022, 11:39:50 UTC |
fc3ed4a | Axel Naumann | 01 March 2022, 16:20:13 UTC | [cmake] Fix xrootd 5.4.1 sha256 hash. (cherry picked from commit db6a9d62f26ce735d010180ecee438137aa75857) | 01 March 2022, 16:21:11 UTC |
7394e69 | Enrico Guiraud | 01 March 2022, 08:58:44 UTC | [DF] Disable df107_SingleTopAnalysis on Mac+M1 There are jit relocation issues on that platform that cause the tutorial to crash, disable until fixed. | 01 March 2022, 14:50:32 UTC |
48fad69 | Enrico Guiraud | 01 March 2022, 08:45:21 UTC | [DF][NFC] Disable test with jitted exception logic on Mac+M1 Catching exceptions thrown by jitted code is not supported on that platform. | 01 March 2022, 14:50:08 UTC |
37dfb60 | Enrico Guiraud | 28 February 2022, 16:50:31 UTC | [DF][NFC] Disable a test with jitted exceptions on windows | 01 March 2022, 14:50:08 UTC |
98e0b02 | Enrico Guiraud | 18 February 2022, 11:44:33 UTC | [DF][NFC] Update docs: Vary+DefinePerSample is now supported | 01 March 2022, 14:50:08 UTC |
f9d9aed | Enrico Guiraud | 18 February 2022, 11:40:51 UTC | [DF] Fix Vary+DefinePerSample Columns defined via DefinePerSample cannot ever have a dependency on the systematic variation (they don't depend on any dataset column), so when a RDefinePerSample is asked to `MakeVariations` it just does nothing, and program logic should never end up requesting a varied value for a RDefinePerSample. The commit also adds a test for this case. | 01 March 2022, 14:50:08 UTC |
06c4ceb | Enrico Guiraud | 18 February 2022, 09:40:06 UTC | [DF] Add diagnostics for simultaneous variation of the same column It does not make sense, it should not be allowed. | 01 March 2022, 14:50:08 UTC |
ca698ea | Enrico Guiraud | 17 February 2022, 19:18:03 UTC | [DF] Remove useless protection against vector<bool> Value types of variations are always RVecs. | 01 March 2022, 14:50:08 UTC |
3ad3bad | Axel Naumann | 28 February 2022, 14:20:30 UTC | [cmake] Bump xrootd to 5.4.1: required for building with static libssl as we do on macOS. (cherry picked from commit 305493a709e5c4903cad6398e0e0294555f3b7de) | 01 March 2022, 14:46:30 UTC |
7642f3e | Axel Naumann | 01 February 2022, 15:03:10 UTC | [io] Enable auto schema conversion for vector<ARITH_TYPE>: Despite root claiming to be running a conversion, it actually does not. If one looks in ReadObjectAny, there is some code dealing with conversions, but the actual streamer call is ``` TClass *clOnfile = 0; ... clRef->Streamer( obj, *this, clOnfile ); ``` where there are no other assignments to clOnfile. So the fact that there is a conversion does not get passed to the streamer. It also seems excessive to generate an Info message each time there is a conversion. Co-authored-by: Scott Snyder <snyder@bnl.gov> (cherry picked from commit ea79ffcb1006761f70f1499d7f25db5846210528) | 24 February 2022, 15:13:04 UTC |
6b6ff5e | Axel Naumann | 01 February 2022, 15:37:15 UTC | [io] Add test against ROOT-8367. Co-authored-by: Scott Snyder <snyder@bnl.gov> (cherry picked from commit 4cb2968b56235177c5a0d43a40240c49e1c069fa) | 24 February 2022, 15:10:51 UTC |
8efc1ea | Axel Naumann | 24 February 2022, 15:04:54 UTC | [relnotes] Formatting. | 24 February 2022, 15:04:54 UTC |
252cf52 | Axel Naumann | 22 February 2022, 17:37:06 UTC | [relnotes] Remove empty sections; move PyROOT up; bump dates. | 22 February 2022, 17:37:06 UTC |
6f5c8a1 | Axel Naumann | 22 February 2022, 17:26:40 UTC | [relnotes] Formatting. | 22 February 2022, 17:26:40 UTC |
fb3bf27 | Axel Naumann | 22 February 2022, 13:37:25 UTC | [relnotes] Mention cling opt; cmake 3.16; C++14; rm memstat, alien depr. Co-authored-by: Javier Lopez-Gomez <javier.lopez.gomez@cern.ch> | 22 February 2022, 13:37:25 UTC |
6b8531c | Philippe Canal | 15 February 2022, 23:31:26 UTC | TBranchElement: no drilling through new members. This resolved the problem seen at: https://github.com/cms-sw/cmssw/issues/36908#issuecomment-1036397481 and fix #9899. The problem is the rules are applied to a data member nested inside an object nested inside an STL collection that is a new data member of the class reco::HaloClusterCandidateHCAL, since it is a new member compared to the layout on file, none of the objects; from the new member down to the object on which the rules need to be run) are actually streamed and the code gathering the information to run the rule got a bit lost ; it is likely (I am checking as we speak) that in previous release the rule was not even attempted to be run ... which might actually be the desired behavior in this specific case. The solution is to have GatherArtificialElements stop drilling through data members which are not stored in the existing TTree. | 18 February 2022, 16:17:50 UTC |
436df37 | Philippe Canal | 25 January 2022, 23:58:47 UTC | TBranchElement::InitInfo improve doc | 18 February 2022, 16:17:50 UTC |
2ef62b7 | Philippe Canal | 10 November 2021, 21:58:09 UTC | Add missing init of TIsAProxy::fNextLastSlot. (thanks valgrind for pointing this out) | 18 February 2022, 16:17:50 UTC |
41fa3f1 | Sergey Linev | 18 February 2022, 12:47:58 UTC | [jsroot] small fixes in TGeo drawing in TCanvas | 18 February 2022, 12:50:06 UTC |
a9ebd7c | Enrico Guiraud | 17 February 2022, 19:16:54 UTC | [DF] Add cross-checks for jitted expression type vs column type | 18 February 2022, 09:43:37 UTC |
2d899fe | Sergey Linev | 17 February 2022, 14:57:34 UTC | Fix TMultiGraph::RecursiveRemove histogram and functions should be handled as well | 18 February 2022, 09:39:31 UTC |
066e610 | Jonas Rembser | 17 February 2022, 13:24:35 UTC | [RF] Remove the check is CUDA works from `RooBatchCompute.cu` Testing to see if CUDA works by actually allocating memory in the initialization of RooBatch compute is not necessary so far. The point of this test was to verify that CUDA works on the systm to dynamically figure out whether the CUDA library should be used or not. However, this doesn't need to be figured out by RooFit in the current design, as the user has to specify `BatchMode("cuda")` or `BatchMode("cpu")` himself. | 17 February 2022, 20:45:28 UTC |
0535905 | Jonas Rembser | 17 February 2022, 11:47:15 UTC | [RF] Defer loading of `RooBatchCompute` libs to first RooFitDriver use Initializing RooBatchCompute takes over a second when CUDA is also available on the system. It is unreasonable to do this everytime the RooFit library is loaded. For example, the time of small unit tests was completely dominated by the RooBatchCompute initialization time before this commit. This change entailed another change in `Buffers.cxx` to avoid static initialization problems. | 17 February 2022, 20:45:28 UTC |
d3ea674 | Enrico Guiraud | 17 February 2022, 14:32:21 UTC | [DF] Add a section on systematics to RDF's user guide | 17 February 2022, 18:21:17 UTC |
67867c0 | Enrico Guiraud | 17 February 2022, 10:02:02 UTC | [DF][NFC] Improve documentation of systematic variations support | 17 February 2022, 18:21:17 UTC |
dcbfa43 | Enrico Guiraud | 17 February 2022, 15:14:02 UTC | [DF] Add a missing include in RRangeBase.hxx | 17 February 2022, 16:27:52 UTC |
ab3ae7f | Enrico Guiraud | 17 February 2022, 11:32:34 UTC | [DF] Add support for varied Ranges Range calls can depend on systematic variations by means of upstream filters. | 17 February 2022, 16:27:52 UTC |
afb3083 | Enrico Guiraud | 17 February 2022, 11:37:33 UTC | [DF] Move a data member to the non-template base class We prefer keeping class templates slim. | 17 February 2022, 16:27:52 UTC |
49e0614 | Enrico Guiraud | 17 February 2022, 11:22:19 UTC | [DF] Only create varied filters if the RFilter depends on the variation Before this patch, we were needlessly creating varied RFilter clones even for variations on which the RFilter node did not depend. The varied filters were therefore exact clones of the nominal. With this patch we just use the nominal RFilters instead. | 17 February 2022, 16:27:52 UTC |
b64bea2 | Enrico Guiraud | 17 February 2022, 11:18:31 UTC | [DF] Compare pointers to base rather than pointers to void | 17 February 2022, 16:27:52 UTC |
aeec4b6 | Enrico Guiraud | 17 February 2022, 10:50:08 UTC | [DF] Error out rather than silently returning a nullptr We should never call the base class implementation of GetVariedFilter, but at the same time it makes no sense to implement this method for the head node (RLoopManager), on which we never call this. Let's make it clear that we never expect the base classs implementation to be called with the appropriate error message, rather than (potentially, in case of logical bugs in RDataFrame) silently returning a nullptr. | 17 February 2022, 16:27:52 UTC |
5310299 | Enrico Guiraud | 17 February 2022, 10:46:22 UTC | [DF] Simplify GetVariedFilter Program logic guarantees we never call it for the "nominal" case. | 17 February 2022, 16:27:52 UTC |
91028ef | Enrico Guiraud | 16 February 2022, 16:24:19 UTC | [NFC][relnotes] Mention syst. variations support in the release notes | 17 February 2022, 09:24:51 UTC |
bd3bc42 | Enrico Guiraud | 17 February 2022, 09:24:02 UTC | [DF] Add support for systematic variations The feature branch is particularly large and not all commits compile due to some sequencing in the introduction of new types, hence the merge commit. | 17 February 2022, 09:24:02 UTC |
4e9ab8d | Enrico Guiraud | 16 February 2022, 10:41:46 UTC | [DF] Add GetVariations method to query registered variations | 17 February 2022, 09:23:21 UTC |
6501a7c | Enrico Guiraud | 16 February 2022, 08:49:27 UTC | [DF] Detach varied histograms from the current directory This is consistent with what we do for all other histograms produced by RDataFrame and it fixes a double free bug with histograms that are deleted by PyROOT at the end of the ROOT session and only afterwars by the shared pointers that contain them. | 17 February 2022, 09:23:21 UTC |
29440e7 | Enrico Guiraud | 15 February 2022, 16:16:20 UTC | [PyROOT] Do not hide functions in namespace ROOT.RDF.Experimental Instead of overwriting ROOT.RDF.Experimental, just add the Distributed submodule to it. Co-authored-by: Vincenzo Eduardo Padulano <v.e.padulano@gmail.com> | 17 February 2022, 09:23:20 UTC |
96e2287 | Enrico Guiraud | 15 February 2022, 11:24:47 UTC | [DF][NFC] Add missing docs to some members | 17 February 2022, 09:23:20 UTC |
cc5518c | Enrico Guiraud | 15 February 2022, 09:53:46 UTC | [DF][NFC] No need to default-initialize a std::array We set its values in the ctor body. | 17 February 2022, 09:23:19 UTC |
cb9b0ea | Enrico Guiraud | 15 February 2022, 09:53:21 UTC | [DF][NFC] Improve formatting of MakeColumnReaders | 17 February 2022, 09:23:19 UTC |
c518d3b | Enrico Guiraud | 15 February 2022, 09:58:19 UTC | [DF] Prefer range-based to explicit loop Co-authored-by: Axel Naumann <Axel.Naumann@cern.ch> | 17 February 2022, 09:23:18 UTC |
8f7cd83 | Enrico Guiraud | 15 February 2022, 09:49:23 UTC | [DF][NFC] Use a clearer template parameter name Co-authored-by: Axel Naumann <Axel.Naumann@cern.ch> | 17 February 2022, 09:23:18 UTC |
8856e06 | Enrico Guiraud | 15 February 2022, 09:24:17 UTC | [DF][NFC] Update a comment Co-authored-by: Axel Naumann <Axel.Naumann@cern.ch> | 17 February 2022, 09:23:17 UTC |
78d4419 | Enrico Guiraud | 07 February 2022, 16:45:08 UTC | [DF][NFC] Add doxygen docs for jitted Vary overloads | 17 February 2022, 09:23:17 UTC |
3d7c655 | Enrico Guiraud | 07 February 2022, 16:18:21 UTC | [DF] Try to make windows happy | 17 February 2022, 09:23:16 UTC |
d14e30a | Enrico Guiraud | 07 February 2022, 11:47:53 UTC | [DF] Fix checks for wrong output of a Vary expression The new logic also fixes bugs in the diagnostics of two cases: - multiple columns varied simultaneously - a single varied column of RVec type Test coverage for these cases was improved. | 17 February 2022, 09:23:15 UTC |
e34c874 | Enrico Guiraud | 02 February 2022, 17:17:00 UTC | [DF] Add tests for some Vary edge cases | 17 February 2022, 09:23:15 UTC |
f674f7a | Enrico Guiraud | 01 February 2022, 16:03:14 UTC | [DF] Add tests for jitted Vary | 17 February 2022, 09:23:14 UTC |
5efce65 | Enrico Guiraud | 01 February 2022, 16:03:03 UTC | [DF] Add support for jitted Vary calls | 17 February 2022, 09:23:14 UTC |
a28c5dc | Enrico Guiraud | 01 February 2022, 16:02:31 UTC | [DF] Test multiple variations on the same column | 17 February 2022, 09:23:14 UTC |
7dfc9a0 | Enrico Guiraud | 26 January 2022, 14:20:31 UTC | [DF][NFC] Document that helper's Initialize method must reset state This is a requirement in case the result of the action is varied. | 17 February 2022, 09:23:13 UTC |
826f82c | Enrico Guiraud | 24 January 2022, 18:11:07 UTC | [DF] Fix jitted filters depending on systematic variations | 17 February 2022, 09:23:13 UTC |
9fd11c7 | Enrico Guiraud | 24 January 2022, 17:47:28 UTC | [DF] Fix jitted Defines depending on a variation RJittedDefine now stores and propagates information about the variations it depends on. | 17 February 2022, 09:23:12 UTC |
2a6a180 | Enrico Guiraud | 24 January 2022, 17:42:07 UTC | [DF] Fix jitted actions that depend on systematic variations For this to work we need RJittedAction to propagate dependencies on systematic variations to VariationsFor. | 17 February 2022, 09:23:12 UTC |
dafcc35 | Enrico Guiraud | 21 October 2021, 08:45:23 UTC | [DF] Add test suite for Vary | 17 February 2022, 09:23:11 UTC |
13ceb53 | Enrico Guiraud | 10 January 2022, 17:10:58 UTC | [DF] Silence spurious compiler warnings The compiler either does not see that we are initializing every element of `doesVariationApply` in the for loop (and complains) or it sees `i < sizeof...(ColTypes)` and complains that the comparison is never true when the template parameter pack ColTypes is empty. This patch fixes those warnings. | 17 February 2022, 09:23:11 UTC |
c7430ad | Enrico Guiraud | 10 January 2022, 16:04:30 UTC | [DF] Teach RDF to propagate systematic variations through Filters They now produce different results in different "universes" as needed. | 17 February 2022, 09:23:10 UTC |
94368a9 | Enrico Guiraud | 16 December 2021, 15:56:36 UTC | [DF] Teach RDF to propagate systematic variations through Defines | 17 February 2022, 09:23:10 UTC |
959ed9c | Enrico Guiraud | 08 November 2021, 16:10:22 UTC | [DF] Add ROOT::RDF::Experimental::VariationsFor | 17 February 2022, 09:23:09 UTC |
17f5f6d | Enrico Guiraud | 16 November 2021, 09:19:50 UTC | [DF] Add RAction::MakeVariedAction This lets us produce RVariedActions from RActions in a type-erased manner. | 17 February 2022, 09:23:09 UTC |
19fe1d1 | Enrico Guiraud | 16 November 2021, 09:18:44 UTC | [DF] Teach most action helpers to handle systematic variations The only requirement is that they implement the `MakeNew` method so the helper can be "cloned" for each variation. Support for variation of Report, Display and Snapshot is still missing. | 17 February 2022, 09:23:08 UTC |
c8dbb29 | Enrico Guiraud | 16 November 2021, 09:16:38 UTC | [DF] Add support for an optional MakeNew method in action helpers This will be needed to create action helpers for each of the systematic variations from the original (nominal) helper. Concrete implementations for the different action helpers will be added in future commits. | 17 February 2022, 09:23:08 UTC |
de1785c | Enrico Guiraud | 16 November 2021, 09:10:35 UTC | [DF] Add RVariedAction An RAction that produces N results, one per systematic variation. | 17 February 2022, 09:23:07 UTC |
de4c1e9 | Enrico Guiraud | 14 December 2021, 12:00:50 UTC | [DF] Teach MakeColumnReaders to handle systematic variations | 17 February 2022, 09:23:07 UTC |
9afd75f | Enrico Guiraud | 11 November 2021, 14:33:53 UTC | [DF] Add RVariationReader Like RDefineReader, but it reads a column value for a specific column, slot and variation. | 17 February 2022, 09:23:06 UTC |
03d13ab | Enrico Guiraud | 05 November 2021, 16:00:12 UTC | [DF] Make CheckForDefinition refer to Vary as well as Redefine Both methods need to check the definition of the column. | 17 February 2022, 09:23:06 UTC |
f98c365 | Enrico Guiraud | 29 October 2021, 14:14:19 UTC | [DF] Add RInterface::Vary | 17 February 2022, 09:23:05 UTC |
8e3a9f5 | Enrico Guiraud | 10 November 2021, 17:23:11 UTC | [DF] Have RRange, RFilter store the variations they depend on In the case of ranges, their result never actually depends on a variation, but they still need to propagate variations that filters depend on down the computation graph, because results will depend on variations that filters depend on. | 17 February 2022, 09:23:05 UTC |
c4cca0d | Enrico Guiraud | 09 November 2021, 21:15:34 UTC | [DF] Have RDefineBase store the variations its result depends on | 17 February 2022, 09:23:04 UTC |
c3f6161 | Enrico Guiraud | 09 November 2021, 20:47:38 UTC | [DF] Have RActionBase store the variations its result depends on | 17 February 2022, 09:23:04 UTC |
a20f6e8 | Enrico Guiraud | 29 October 2021, 13:43:31 UTC | [DF] Make code in RColumnRegister more uniform Different names and patterns were used for fVariations and fColRegister, let's pick the most readable. Also fix some comments. | 17 February 2022, 09:23:03 UTC |
bad18da | Enrico Guiraud | 29 October 2021, 13:31:15 UTC | [DF] Teach RColumnRegister to handle systematic variations They are treated similarly to defines. In particular, make it possible to registered RVariation objects in RDFBookedColumns and teach RDFBookedColumns to evaluate which systematic variations are associated to a given set of columns. This includes systematic variations directly applied to the columns and also those that the columns depend on indirectly because they are Defined columns that depend on varied columns. | 17 February 2022, 09:23:03 UTC |
301d31a | Enrico Guiraud | 29 October 2021, 08:36:48 UTC | [DF] Add RVariation, RVariationBase classes RVariation is to systematic variations what RDefine is to a Define'd RDF column. It stores all information about the systematic variations (including the user-defined callable to be used to evaluate them) and it can be used to evaluate the varied columns. | 17 February 2022, 09:23:03 UTC |
16e3ffe | Enrico Guiraud | 08 November 2021, 15:14:51 UTC | [DF] Add RResultMap Like RResultPtr, but a key/value map-like container of multple results. It will be useful to store all variations of a result when systematic variations are enabled. | 17 February 2022, 09:23:02 UTC |
476ede7 | Enrico Guiraud | 10 November 2021, 17:27:34 UTC | [DF] Create RJittedFilter inside BookFilterJit This solves a sequencing problem: to track the dependncy on systematic variations, we will need to know which columns the filter depends on (which is evaluated inside BookFilterJit) in order to construct the RJittedFilter, which will propagate that information to downstream nodes. | 17 February 2022, 09:23:02 UTC |
928af41 | Enrico Guiraud | 09 November 2021, 21:16:48 UTC | [DF] Add a Union(vector1, vector2) helper function | 17 February 2022, 09:23:01 UTC |
28bb0b0 | Enrico Guiraud | 16 November 2021, 15:55:12 UTC | [DF] Add IsRVec to Utils.hxx | 17 February 2022, 09:23:01 UTC |
ece66a1 | Enrico Guiraud | 14 December 2021, 11:41:56 UTC | [DF] Move IsStrInVec out of anonymous namespace We want to use it in several source files. | 17 February 2022, 09:23:00 UTC |
4b4274d | Enrico Guiraud | 26 January 2022, 12:22:28 UTC | [DF] Make internal naming convention more uniform Prefer calling the previous nodes of in the computation graph and their types fPrevNode/PrevNode instead of fPrevData and PrevDataFrame. | 17 February 2022, 09:22:58 UTC |
74c66b2 | Jonas Rembser | 16 February 2022, 14:35:02 UTC | [RF][RelNotes] Deprecated string fit options in RooFit The string-based fit options are officially deprecated with this commit, with the intention to finally remove them in the next release (ROOT v6.28). This will give us more flexibility when refactoring the RooMinimizer to integrate new features like the batch mode, the new likelihood classes, or automatic differentiation. | 16 February 2022, 20:19:01 UTC |
739b18d | Enrico Guiraud | 16 February 2022, 16:24:43 UTC | [relnotes][NFC] Fix a couple of typos | 16 February 2022, 16:40:44 UTC |
a26bc46 | Enrico Guiraud | 16 February 2022, 11:26:10 UTC | [NFC][relnotes] Sync with master | 16 February 2022, 13:03:53 UTC |
38e1ac1 | Jonas Rembser | 15 February 2022, 15:25:28 UTC | [RF][PyROOT] Prohibit resetting RooWorkspace import functions in PyROOT Many people pythonized the RooWorkspace themselves, by adding a new attribute `_import` that calls getattr(self, "import") under the hood. However, `_import` is now the reference to the original cppyy overload, and resetting it with a wrapper around `import` would cause infinite recursions! We prevent resetting any import-related function here, which results in a clearer error to the user than an infinite call stack involving the internal pythonization code. | 16 February 2022, 12:56:41 UTC |
cf9646d | Jonas Rembser | 14 February 2022, 15:40:27 UTC | [RelNotes] Deprecate RooHashTable, RooNameSet, RooSetPair, and RooList | 16 February 2022, 12:56:41 UTC |
56709fd | Javier Lopez-Gomez | 16 February 2022, 12:29:56 UTC | [relnotes][skip-ci] Add missing information about RNTuple and `latest-stable` branch | 16 February 2022, 12:53:32 UTC |
2259628 | moneta | 08 February 2022, 18:43:58 UTC | Fix class GoFTest using pre-defined distributions (1 sample test) Before GoFTest was using the data computed parameters when using the pre-defined Gaussian, Exponential or LogNormal This is a different test and not a standard AD or KS test. It was a Lilliefors test, see https://en.wikipedia.org/wiki/Lilliefors_test. Now add possibility of passing distribution parameters when using the predefined distributions. Use same convention as R for distribution parameters Update test, stressGofTest for the new fixes and add the computation of the KS and AD test using R when the R interface is available in ROOT | 15 February 2022, 15:23:01 UTC |
62438fb | moneta | 04 February 2022, 18:24:37 UTC | Fix Kolmogorov-Smirnov 1 sample test (comparison data with a model) As reported in #9636 the line computing the distance between the model cdf and the empirical distribution was not correct. It is now fixed as in the original implementation. | 15 February 2022, 15:23:01 UTC |
a36c685 | Sergey Linev | 15 February 2022, 13:26:16 UTC | [jsroot] release 6.3.4 with several fixes | 15 February 2022, 13:45:37 UTC |
fd661c2 | Enrico Guiraud | 14 February 2022, 18:28:02 UTC | [DF] Clean up dummy files after test | 15 February 2022, 13:23:01 UTC |
2736b59 | Enrico Guiraud | 14 February 2022, 18:19:31 UTC | [DF] Prefer plain std::string to std::stringstream when possible | 15 February 2022, 13:23:01 UTC |
0fd981b | Enrico Guiraud | 14 February 2022, 18:12:12 UTC | [DF] Rely on RAII to close files, delete their contents This way we don't have to remember to close the files manually (in one case we closed a file twice by mistake), and lifetime management of the file contents is clearer: they are never deleted by Close or after the file has been closd, as they go out of scope before the file is destroyed and closed. | 15 February 2022, 13:23:01 UTC |
3ba6e71 | Sergey Linev | 14 February 2022, 09:43:21 UTC | [rootfit] fix compiler warnings in LikelihoodGradientWrapper.cxx | 14 February 2022, 12:26:08 UTC |
a7d634d | Vincenzo Eduardo Padulano | 10 February 2022, 23:09:28 UTC | [relnotes] Add HistoND to distributed RDataFrame | 14 February 2022, 08:59:14 UTC |