| 37ca4d6 | Axel Naumann | 23 August 2018, 14:57:30 UTC | Update ROOT version files to v6.14/04. | 23 August 2018, 14:57:30 UTC |
| 8c86427 | Axel Naumann | 23 August 2018, 14:11:55 UTC | Update release notes for v6.14/04. | 23 August 2018, 14:11:55 UTC |
| 3026016 | Guilherme Amadio | 23 August 2018, 07:50:37 UTC | Update release notes | 23 August 2018, 07:50:37 UTC |
| 68a7c0a | Guilherme Amadio | 30 July 2018, 07:33:22 UTC | Clear cache and set $DAVIX_FOUND when using builtin Davix | 23 August 2018, 07:04:58 UTC |
| 78b2d0a | Guilherme Amadio | 25 July 2018, 08:18:59 UTC | Require same version of CMake as ROOT in ROOTUseFile.cmake The minimum version needed by ROOT modules is necessary for projects including ROOT, since ROOT's modules use CMake 3.x constructs not supported by CMake 2.8. Fixes: ROOT-9493 | 23 August 2018, 07:04:14 UTC |
| 8800137 | Guilherme Amadio | 19 July 2018, 15:13:20 UTC | Use NEW behavior for CMake policy CMP0051 We need to move LLVM in ROOT to use the new behavior of policy CMP0051 as it will be removed soon and the warning below will become an error: CMake Deprecation Warning at interpreter/llvm/src/CMakeLists.txt:15 (cmake_policy): The OLD behavior for policy CMP0051 will be removed from a future version of CMake. The cmake-policies(7) manual explains that the OLD behaviors of all policies are deprecated and that a policy should be set to OLD only under specific short-term circumstances. Projects should be ported to the NEW behavior and not rely on setting a policy to OLD. | 23 August 2018, 07:03:34 UTC |
| 13a7be6 | Guilherme Amadio | 18 July 2018, 13:36:57 UTC | [GenVector] Do not use pow() to just square something | 23 August 2018, 07:03:21 UTC |
| 00f4ce1 | Philippe Canal | 22 August 2018, 13:30:37 UTC | Merge branch 'lz4-revert-6-14' of https://github.com/oshadura/root into v6-14-00-patches | 22 August 2018, 13:32:21 UTC |
| 8818b8a | Oksana Shadura | 20 August 2018, 18:33:15 UTC | Changing CMake settings back to default zlib | 22 August 2018, 11:16:08 UTC |
| e2d2441 | Oksana Shadura | 20 August 2018, 12:37:15 UTC | Revert "Changing default from ZLIB to LZ4 in case of illegal algorithm usage in ROOT" This reverts commit 8ac9554e05c146a4f8aefcb8024ef2781f764fb2. | 22 August 2018, 11:16:08 UTC |
| 7aa8445 | Oksana Shadura | 20 August 2018, 12:32:03 UTC | Revert "Final update to use LZ4 as a default for ROOT" This reverts commit 7cf40ecd4eb5eda52356946a1047d290e7916271. | 22 August 2018, 11:16:08 UTC |
| 4d05469 | Philippe Canal | 11 June 2018, 21:02:02 UTC | Avoid getting confused by out-of-range request to TTreeCache | 21 August 2018, 14:56:13 UTC |
| d95922c | Philippe Canal | 16 August 2018, 19:55:20 UTC | Additional fix for rule scheduling in TBranchElement. This is an additional fix for https://github.com/cms-sw/cmssw/pull/22594 If one of the class associated with a TBranchElement has a base class and/or sub-object class that has a rule that requires a cache (onfileObject) object to stage the original data *but* does not have branch in the TTree that also neeed the same cache object, we need to associate a cache object with the StreamerInfoAction sequence that needs it (associated with a higher level branch) and give it ownership of this cache object) | 21 August 2018, 14:31:27 UTC |
| e408e13 | Philippe Canal | 16 August 2018, 18:24:02 UTC | Avoid double search | 21 August 2018, 14:31:27 UTC |
| 66dd82f | Philippe Canal | 16 August 2018, 18:18:42 UTC | Improve debug priting of StreamerElement info | 21 August 2018, 14:31:27 UTC |
| f98930a | Philippe Canal | 16 August 2018, 18:18:10 UTC | [NFC] White space | 21 August 2018, 14:31:27 UTC |
| 3b016a8 | Giulio Eulisse | 20 August 2018, 19:52:58 UTC | Support unsigned integer types in RArrowDS | 20 August 2018, 21:49:08 UTC |
| f5f253b | Oksana Shadura | 16 August 2018, 16:23:40 UTC | Recovery of correct CMake invocation of rootcling We are fixing a correct invocation accidentally removed in ROOT_GENERATE_DICTIONARY() in '[cxxmodules] Move away from environment variable' commit. | 17 August 2018, 21:24:34 UTC |
| 3650268 | Giulio Eulisse | 17 August 2018, 13:08:31 UTC | v6-14-00-patches: Fix typo in test. | 17 August 2018, 14:21:59 UTC |
| 18df8d0 | Giulio Eulisse | 16 August 2018, 14:27:39 UTC | Fix typo in RArrowDS The name of the type is Long_t, not Long32_t. | 17 August 2018, 09:16:08 UTC |
| 777103a | Guilherme Amadio | 09 August 2018, 12:12:31 UTC | Revert "Don't install intermediate static lib minicern" This reverts commit 1915926942b8f4f525a6f1460d09df9528e1cc66. The minicern library is used by ATLAS. | 09 August 2018, 12:14:39 UTC |
| a8cc688 | Andrei Gheata | 07 August 2018, 13:49:42 UTC | Consistency for copy constructors and assignment. (cherry picked from commit 9999fb552c400538d0250c436c4983cefa34f216) | 08 August 2018, 11:10:50 UTC |
| b5fde2a | Andrei Gheata | 05 July 2018, 10:13:46 UTC | Added protection for ComputeLradTsaiFactor when Z=0 (cherry picked from commit 4f477ee4fd4a4ee2fb2f941fd2279522dec08bb2) | 07 August 2018, 13:51:13 UTC |
| b553c5d | Philippe Canal | 27 July 2018, 23:17:28 UTC | Improve thread scability of TRef. Use Read/Write lock where relevant. Cache the last result of TProcessID::IsValid and TProcessID::GetProcessWithUID as most often the same PID will be used for most of the process lifetime. This addresses: https://root-forum.cern.ch/t/copying-trefs-and-accessing-tref-data-from-multiple-threads/29417 | 02 August 2018, 12:28:46 UTC |
| 68186c5 | Andrei Gheata | 01 August 2018, 09:16:56 UTC | Extended return value of operator* to TGeoHMatrix. Added Multiply taking const reference. (cherry picked from commit af55b3e542e647d7b91a23d0880f69021e02a882) | 02 August 2018, 09:11:06 UTC |
| 3eefc91 | Mattias Ellert | 30 June 2018, 19:27:57 UTC | Fixes for Python 3.7 * PyROOT_PyUnicode_AsString changed return type from char* to const char* * Using _PyObject_GC_TRACK results in linking error: undefined reference to `_PyGC_generation0' The python documentation says this about _PyObject_GC_TRACK: "A macro version of PyObject_GC_Track(). It should not be used for extension modules." So it should not be used. Calling PyObject_GC_Track instead avoids the undefined symbol. I also removed the #ifndef R__WIN32 since the comment indicates that this was added to avoid the same issue on windows. Calling PyObject_GC_Track instead of using the _PyObject_GC_TRACK macro should fix the undefined symbol probelem on windows too. | 31 July 2018, 17:54:54 UTC |
| 7bc0338 | Enrico Guiraud | 26 July 2018, 16:21:52 UTC | [DF] Add a test for ROOT-9555: Reduce on a bool column | 30 July 2018, 09:26:52 UTC |
| 29324b8 | Enrico Guiraud | 26 July 2018, 16:15:14 UTC | [DF] Use a common type alias for the container of action helper results A centralized type alias for this type makes it easy to change it later and it makes it possible to avoid instantiating std::vector<bool>: the latter benefit resolves ROOT-9555. | 30 July 2018, 09:26:52 UTC |
| 2773d4c | Axel Naumann | 27 July 2018, 12:58:59 UTC | [RelNotes] Formatting. | 27 July 2018, 12:58:59 UTC |
| 23f7fc7 | Axel Naumann | 27 July 2018, 12:51:32 UTC | [RelNotes] Mention "late" v6.14/02 Jira tickets. | 27 July 2018, 12:51:32 UTC |
| 9577c42 | Axel Naumann | 10 July 2018, 20:41:45 UTC | Make TBits::op~() const-correct [ROOT-9529]. (cherry picked from commit dfb1a7b3877f48a3b4a0b2531e450a5f8cb8e807) | 27 July 2018, 10:50:14 UTC |
| 31a34f1 | Axel Naumann | 27 July 2018, 10:30:04 UTC | [RelNotes] Fix Jira links, update date. (cherry picked from commit 4abdf1857b07fbb4699dd8533ae5dd8b641f625e) | 27 July 2018, 10:30:49 UTC |
| 56dd3cc | Axel Naumann | 27 July 2018, 08:52:29 UTC | Update ROOT version files to v6.14/03. | 27 July 2018, 08:52:29 UTC |
| de9d0f3 | Axel Naumann | 27 July 2018, 08:47:09 UTC | Update ROOT version files to v6.14/02. | 27 July 2018, 08:47:09 UTC |
| 6665707 | Axel Naumann | 27 July 2018, 08:45:27 UTC | [RelNotes] Update for 6.14/02. | 27 July 2018, 08:45:27 UTC |
| 1eaf56c | Guilherme Amadio | 26 July 2018, 13:30:01 UTC | Revert "Do not include internal PostgreSQL header" This reverts commit 6fa43c88b40e058f0c1dfeed9f25f852ee57a6ea. (cherry picked from commit 20999049250ff5083e5d1770d9d8a6b55cb2784e) | 27 July 2018, 07:00:55 UTC |
| bd8ea90 | Andrei Gheata | 25 July 2018, 16:25:29 UTC | Fixed missing Inverse in the base class of TGeoMatrix. Removed GetInverse. (cherry picked from commit cfd7435fd2223e3822b9a6e08c54a839034245da) | 26 July 2018, 07:40:09 UTC |
| d33580d | Andrei Gheata | 25 July 2018, 15:07:17 UTC | Thread safety for matrix operations. Consistent use of operators across matrix types with test macro. (cherry picked from commit dc035e754bc3de6329dcc52e91d5440f49ed064d) | 26 July 2018, 07:39:47 UTC |
| 0302af9 | Andrei Gheata | 18 July 2018, 12:15:24 UTC | Added thread safe method GetInverse to compute the inverse matrix for all transformations (cherry picked from commit 36a3c2a5ee1852e6bc20d4273e9a2281043bd70d) | 23 July 2018, 09:23:40 UTC |
| 6749fe1 | Guilherme Amadio | 16 July 2018, 09:38:53 UTC | Use all-keyword style in target_link_libraries() When using builtin_openssl=ON, CMake erroneously exports the builtin static libraries in ROOTConfig-targets.cmake, which causes problems with dependent projects. In order to avoid this, we need Net and RHTTP to use SSL as a private dependency. Since CMake requires target_link_libraries() to either be all-plain (as before) or all-keyword (required to allow PRIVATE linking), we need to move the build system to use all-keyword linking only. Fixes: ROOT-9532 | 18 July 2018, 09:57:21 UTC |
| 0ab08ce | Bertrand Bellenot | 13 July 2018, 09:28:07 UTC | Add support for Python 3 on Windows This fixes the issue reported on the forum at: https://root-forum.cern.ch/t/windows-10-source-installation-failing-due-to-python-3-bindings/29771 | 13 July 2018, 09:28:07 UTC |
| e041fdc | Enrico Guiraud | 09 July 2018, 15:18:38 UTC | [DF] Make RResultPtr move-constructible, add a test for it | 12 July 2018, 08:58:41 UTC |
| a28c5e2 | Enrico Guiraud | 09 July 2018, 15:17:47 UTC | [DF] Make RResultPtr members non-const to allow copy/move assignment Fixes ROOT-9526. | 12 July 2018, 08:58:41 UTC |
| 430204e | Enrico Guiraud | 05 July 2018, 15:50:18 UTC | [DF] Add test for validation of custom column names | 06 July 2018, 21:34:51 UTC |
| 23a8b89 | Enrico Guiraud | 05 July 2018, 17:14:58 UTC | [DF] Throw if name of defined column is not a valid C++ var name | 06 July 2018, 21:34:51 UTC |
| c070e85 | Philippe Canal | 03 July 2018, 21:28:47 UTC | In GatherArtificialElements do not drill through transient data member. They are guaranteed to not have induced any branches and often do not have dictionary information (or have members/bases that do not). Fixes ROOT-9513 And see https://github.com/cms-sw/cmssw/pull/22594 | 06 July 2018, 14:24:33 UTC |
| c85208f | Philippe Canal | 05 July 2018, 15:55:10 UTC | Update release notes | 06 July 2018, 14:24:33 UTC |
| 7ae4f28 | Guilherme Amadio | 05 July 2018, 17:34:28 UTC | Add ROOTDataFrame to ${ROOT_LIBRARIES} | 05 July 2018, 18:18:33 UTC |
| b9dc06b | Sergey Linev | 04 July 2018, 07:37:14 UTC | xml: no need to reassign curr variable - not a dummy now | 05 July 2018, 10:30:22 UTC |
| db0b330 | Sergey Linev | 04 July 2018, 07:34:47 UTC | xml: use variable reference in ExpandStream Introduce dummy for it | 05 July 2018, 10:30:22 UTC |
| 340fd6a | Sergey Linev | 03 July 2018, 08:51:53 UTC | xml: fix in bugfix | 05 July 2018, 10:30:22 UTC |
| 54b1739 | Sergey Linev | 03 July 2018, 08:40:28 UTC | xml: use only offset from current position in LocateValue If current pointer changes due to reallocation offset can be used as is | 05 July 2018, 10:30:22 UTC |
| c954498 | Sergey Linev | 03 July 2018, 08:31:58 UTC | xml: when reallocate buffer, also optionally recalculate current position | 05 July 2018, 10:30:22 UTC |
| c997552 | Olivier Couet | 29 June 2018, 11:25:34 UTC | The error on angle position along the radius doesn't depend on the angle's units (rad, deg, grad). Fix; https://sft.its.cern.ch/jira/browse/ROOT-8476 | 05 July 2018, 09:19:51 UTC |
| 6c42e08 | Bertrand Bellenot | 03 July 2018, 10:05:42 UTC | Implement splash screen version 6 on Windows | 03 July 2018, 10:05:42 UTC |
| 77868d9 | shahoian | 29 June 2018, 08:51:22 UTC | Fix in addition of template fwd declaration in rootmap | 02 July 2018, 14:59:13 UTC |
| cdd253c | Axel Naumann | 26 June 2018, 07:43:15 UTC | [cling] Check file is C_User before hitting disk. Thanks, Vassil! (cherry picked from commit 81f10a37507b14392d7de2558835c5a32b51c2eb) | 02 July 2018, 14:57:08 UTC |
| 80d6982 | Axel Naumann | 27 June 2018, 06:20:44 UTC | [rootrc] Replace Root.ZipMode by Root.CompressionAlgorithm, passing it to R__SetZipMode(). (cherry picked from commit eed8115038f8aac656e517b5f3f8efca475bfbdc) | 02 July 2018, 14:57:08 UTC |
| 5a58b8b | Axel Naumann | 23 June 2018, 09:12:19 UTC | [cling] Only check pointers for Decls from writable dirs (ROOT-9377). (cherry picked from commit fd11b390f3ea0223d272b71bde67b3727ccf1cd5) | 02 July 2018, 14:57:08 UTC |
| 46db07f | Axel Naumann | 22 June 2018, 14:51:01 UTC | [cling] Move NullDerefProtectionTransformer into unnamed namesp. (cherry picked from commit 4d78470dd2e1a3f39ce712a237103c7672ae6f0c) | 02 July 2018, 14:57:08 UTC |
| 2dc421e | Axel Naumann | 04 June 2018, 09:42:14 UTC | RConfig.h uses things defined in RConfigure.h: include it! This fixes warnings due to undefined R__SIZEDDELETE for C++17 builds. (cherry picked from commit c2877de6fe2d61ce7c68e8a5bb7dcb54b4110701) | 02 July 2018, 14:57:08 UTC |
| 7733b72 | Axel Naumann | 24 May 2018, 07:25:20 UTC | [cling] Unroll the pointer-check cache loop. (cherry picked from commit d468346e419d4bb21bac30cc719013128298fe79) | 02 July 2018, 14:57:08 UTC |
| 90b6504 | Bertrand Bellenot | 02 July 2018, 13:11:44 UTC | Fix line width and maker index issues on Windows | 02 July 2018, 13:11:44 UTC |
| d029b6f | Xavier Valls | 29 June 2018, 07:15:07 UTC | fix wrong early return in hadd | 29 June 2018, 13:14:34 UTC |
| ac2986f | Philippe Canal | 27 June 2018, 18:19:10 UTC | Remove obsolete TBranchElement::fIDs | 28 June 2018, 13:00:46 UTC |
| d5b0b95 | Philippe Canal | 27 June 2018, 17:36:19 UTC | Apply new ActionSequence creation to Fill | 28 June 2018, 13:00:46 UTC |
| 5537490 | Philippe Canal | 27 June 2018, 17:32:48 UTC | Factor out ActionSequence creation | 28 June 2018, 13:00:46 UTC |
| f5bc4ea | Philippe Canal | 27 June 2018, 17:06:36 UTC | Remove obsolete (commented-out) code fragement | 28 June 2018, 13:00:46 UTC |
| aafd9b3 | Philippe Canal | 26 June 2018, 19:50:12 UTC | Insure we call SetRead/FillSequence after sub-branches have been added. This is need to be able to distinguish split node from unsplit node. | 28 June 2018, 13:00:46 UTC |
| 6252fa8 | Philippe Canal | 20 June 2018, 15:25:29 UTC | Use same normalization for branch, element name and later searches. | 28 June 2018, 13:00:46 UTC |
| c565081 | Philippe Canal | 26 June 2018, 21:07:08 UTC | Properly set fOnfileObject for the parent | 28 June 2018, 13:00:46 UTC |
| 40beb64 | Philippe Canal | 19 June 2018, 00:46:09 UTC | Fix for https://github.com/cms-sw/cmssw/pull/22594 Instead of assigning rule that do not correspond to any specific branches (for example rules setting transients or whole objects) to an hypothetical 'previous' branch, we switch to assign those orphan rules to either the non-data bearing branches (split single objects or bases) or the collection node (since the splitting completely flatten the hiearchy in this case). This requires enhancing the list of IDs from a single list of element index in the 'sole' current StreamerInfo to a nested list of elements that carries along the sub-StreamerInfo information. | 28 June 2018, 13:00:46 UTC |
| 059ce7a | Philippe Canal | 26 June 2018, 15:07:38 UTC | Warn about problematic base class StreamerElement only once | 28 June 2018, 13:00:46 UTC |
| ae39537 | Philippe Canal | 19 June 2018, 01:42:30 UTC | Avoid double write | 28 June 2018, 13:00:46 UTC |
| 6b2e236 | Philippe Canal | 09 June 2018, 02:52:12 UTC | [NFC] White space | 28 June 2018, 13:00:46 UTC |
| ec01e78 | Enrico Guiraud | 26 June 2018, 13:09:31 UTC | [DF] Remove obsolete part from error message | 27 June 2018, 16:37:04 UTC |
| 772a3d2 | Guilherme Amadio | 27 June 2018, 13:08:07 UTC | Do not exclude etc/http directory from installation This was an unintended side-effect of a previous commit: 9b4d0d8dccbd48d21b0f7c79b6eaf94428f691b1. | 27 June 2018, 14:18:07 UTC |
| 382e8f2 | Enrico Guiraud | 24 June 2018, 15:10:59 UTC | [DF] Avoid TChain::AddClone in Snapshot when unnecessary Since output trees are recreated for each task, and each task only processes one tree cluster (and never crosses file boundaries), we don't need to add the output trees as clones of the input tress unless the input trees have friends (which might cross file boundaries) even if the main tree does not. Usage of AddClone here is undesirable in the general case because it generates many (safe) warnings printed at screen during the event loop, see ROOT-9487. | 26 June 2018, 07:55:05 UTC |
| 74034ba | moneta | 21 June 2018, 14:49:18 UTC | Re-add computation of correlation matrix for every data set that it is input in TMVA. This should be probably changed to be an option, because in case of large dataset (many features and many events) the time spent could be too large | 25 June 2018, 14:21:31 UTC |
| 5bb362b | Enrico Guiraud | 22 June 2018, 16:53:03 UTC | [DF] Use stacks rather than single output files in Snapshot MT This resolves ROOT-9456, "[Jenkins][DF] Sporadic failures in test_snapshotNFiles". The issue is due to interleaved TBB task execution in the same thread. When nested task spawning is present (i.e. when TBB tasks spawn other TBB tasks and the thread pool contains at least three threads), the TBB task stealing mechanism makes it possible that a thread starts executing a new instance of the "parent" tasks before it finishes the execution of the previous parent task (this happens when the thread is waiting for a "stolen" child task to finish, so it steals another threads' parent task). In the case of a multi-thread snapshot, one needs to write at least two branches for the multi-thread branch writing to kick in and have nested task spawning. The crash was caused by a new parent task overwriting the output TTree of the previous parent task in a given thread. When the previous parent task resumed execution, it ended up trying to flush out the contents of the deleted output TTree. The solution adopted is to use thread-local stacks of output TTrees (as opposed to single output TTrees), that tasks push and pop when they start and finish respectively. | 24 June 2018, 08:08:06 UTC |
| ee3efc6 | Enrico Guiraud | 20 June 2018, 12:17:34 UTC | [DF] Avoid concurrent destruction of input and output trees At the beginning of a task a new input TChain replaced the old input TChain for that _thread_ in TTreeProcessorMT, causing a destruction of the previous _thread-local_ TTree. At the same time, the initialization of a new task _in a different thread_ could delete the output TTree of a multi-thread Snapshot action, which was listed in the fClones of the input TTree. This race caused the input ~TTree to access an output tree just after its deletion, in TTree::CopyAddresses. The root cause of this issue is that TTreeProcessorMT uses thread-local variables while RDataFrame uses processing slots which threads can exchange between themselves. | 22 June 2018, 09:28:10 UTC |
| 80dbda0 | Enrico Guiraud | 14 June 2018, 15:54:18 UTC | [DF] Test Snapshot does not write out empty tree buffers | 22 June 2018, 09:28:10 UTC |
| 9dc4ea4 | Enrico Guiraud | 14 June 2018, 17:06:16 UTC | [DF] Do not flush buffers with empty trees from MT Snapshot With very few entries passing a given RDataFrame cutflow, it can happen that a TBB task ends up processing zero "good" entries. In that scenario, it could happen that the first TTree header flushed to the output file from a Snapshot action was for a tree with no entries and no branches. This effectively set the number of branches to zero for all the content written to file, even if following buffers contained well-formed TTrees. With this commit, if the output tree contains zero entries, we avoid flushing it to file. And since the next task will remove this TTree from the output directory, we can be sure that the buffer with the malformed TTree will never be written out. This fixes ROOT-9471. | 22 June 2018, 09:28:10 UTC |
| b7daa09 | Enrico Guiraud | 14 June 2018, 16:54:36 UTC | [DF] Give SnapshotHelperMT ownership of its output trees We used to heap-allocate one TTree per TBB task and never deallocate them, for fear of interfering with ROOT's garbage collection. At the same time, though, we deregistered said trees from ROOT's gc mechanism with `ResetBit(kMustCleanup)`. As a result I'm not sure the output trees were ever deleted. This commit lets SnapshotHelperMT take ownership of the output trees so that when a new task comes along the previous output tree in use by that slot is deleted. This has the desired side-effect of deregistering it from the output directory (the TBufferMergerFile) to which it belongs. This is desirable because we call `Write` on these output files, which in turn would call `Write` on the current output tree __and all previous output trees__ if they were still registered with the directory -- this is not a problem in terms of final output (`Write` on old output trees is a no-op as there is no pending buffer to flush) but is wasteful. The deregistration of obsolete per-task output trees from the thread-local output directories also sets the stage for fixing ROOT-9471. Finally, output trees are re-added to the (now thread-safe) list of cleanups to make sure that output trees are correctly deregistered from the input tree at their destruction, lest the input tree tries to access the deleted output trees in its dtor, as they are registered in its list of clones. | 22 June 2018, 09:28:10 UTC |
| d1bb93f | Enrico Guiraud | 14 June 2018, 15:11:26 UTC | [DF][NFC] Add comment on the ownership of the output trees | 22 June 2018, 09:28:10 UTC |
| 0cd0e17 | Enrico Guiraud | 20 June 2018, 14:18:16 UTC | [DF] Call ActionHelper::FinalizeTask at the end of each task | 22 June 2018, 09:28:10 UTC |
| dab4f6c | Enrico Guiraud | 20 June 2018, 14:04:54 UTC | [DF] Decorate every action helper with an empty FinalizeTask method | 22 June 2018, 09:28:10 UTC |
| a1da74d | Enrico Guiraud | 20 June 2018, 14:04:08 UTC | [DF] Make all action helpers inherit from RActionImpl This is in line with what we require of user-defined action helpers in Book. | 22 June 2018, 09:28:10 UTC |
| ab66a8e | Enrico Guiraud | 20 June 2018, 13:10:03 UTC | [DF] Rename action helpers' InitSlot to InitTask In line with the naming we used for the interface required by `RInterface::Book` for user-defined action helpers. There are several other places where `InitSlot` is used to indicate a function that is executed to initialize a slot at the beginning of a task. Those might be renamed as well, since `InitTask` makes it much more clear that the function is executed multiple times per slot. | 22 June 2018, 09:28:10 UTC |
| e7aedf4 | Enrico Guiraud | 12 June 2018, 08:35:35 UTC | [DF][TOREVERT] Disable datasource_root gtest while it's being debugged | 21 June 2018, 08:21:10 UTC |
| 7768f57 | Axel Naumann | 19 June 2018, 09:30:12 UTC | [R__LOCKGUARD] Add parens around mutex expression (ROOT-9479). (cherry picked from commit e6b8a3cfc66d1d6e07c50b080251d32a7c64177f) | 20 June 2018, 06:12:33 UTC |
| de18b2c | moneta | 22 May 2018, 07:41:53 UTC | Speed up TMVA regression tutorial by switching off by default the old MLP and GA methods. Add instead DNN when TMVA supports CPU. | 18 June 2018, 10:44:01 UTC |
| fcbd86d | Guilherme Amadio | 18 June 2018, 09:30:10 UTC | Add VDT to list of builtin targets so it builds before ROOT The dependency of Vdt::Vdt imported target should have been automatically added by CMake due to it being listed in the BUILD_BYPRODUCTS of the ExternalProject_Add() command, but not all versions of CMake work, so it is necessary to build it early by force. Targets listed in ROOT_BUILTIN_TARGETS get added as a dependency of the move_headers target, which is reasonable since they often provide headers without which ROOT cannot be built in any case (e.g. vdt/vdtMath.h). | 18 June 2018, 09:30:10 UTC |
| 128d948 | Danilo Piparo | 15 June 2018, 09:00:56 UTC | Make header standalone, encapsulate better sha related routines. | 15 June 2018, 11:38:36 UTC |
| 8a3cef4 | Danilo Piparo | 15 June 2018, 07:25:45 UTC | [DF] Look for branches names only once and cache the result rather than re-creating the full set every time it's needed. This is a performance optimisation. The routine for finding branches names is not lightweight, especially when dealing with trees with tens of branches. The names of the branches is now cached in the RInterface and its ownership shared among the RInterface<T> instances. | 15 June 2018, 11:38:29 UTC |
| 914899f | Enrico Guiraud | 14 June 2018, 17:22:47 UTC | [DF] Clean after test cast | 15 June 2018, 06:49:11 UTC |
| 720d314 | Danilo Piparo | 14 June 2018, 06:51:33 UTC | [DF] Add tests for non-triggered lazy snapshot in the sequential and MT cases | 15 June 2018, 06:49:11 UTC |
| 0342898 | Danilo Piparo | 14 June 2018, 06:50:13 UTC | [DF] Warn the user in case a lazy snapshot was booked but not triggered both in the sequential and MT case. The sequential case has also been fixed since a non-triggered lazy snapshot caused a crash. | 15 June 2018, 06:49:11 UTC |
| 9c50f47 | Whitney Armstrong | 16 May 2018, 14:10:19 UTC | Fixed use of TString to compile with gcc 8.1 (#2012) Fixed use of TString to compile with gcc 8.1 (cherry picked from commit 5c1921a4be42c195c4490521de35774f5f1a2aa0) | 15 June 2018, 06:21:32 UTC |
| 96ffa90 | moneta | 14 June 2018, 09:35:41 UTC | Update contributed people | 14 June 2018, 09:35:41 UTC |
