e927453 | Lorenzo Moneta | 17 April 2020, 13:35:56 UTC | Fix the full training recurrent test by using a fixed seed instead of a random one Do not build cudnn full test for simple RNN when cudnn is not available | 17 April 2020, 14:39:36 UTC |
e92f9a6 | Vassil Vassilev | 16 April 2020, 14:04:14 UTC | [wincxxmodules] Enable proper installation of modulemaps; build the libc module. | 17 April 2020, 09:51:17 UTC |
fa1cd8f | Stefan Wunsch | 17 April 2020, 07:32:17 UTC | [PyROOT] Add libexpat to known libraries * Pulled in, e.g., by libpython3.8 on Ubuntu 20.04 | 17 April 2020, 08:53:49 UTC |
0df436c | Stefan Wunsch | 16 April 2020, 11:47:53 UTC | [RDF] Silence -Wunused-const-variable in sqlite test | 17 April 2020, 07:04:29 UTC |
ed5d82e | Vassil Vassilev | 12 April 2020, 13:19:39 UTC | [cxxmodules] Don't mount a custom modulemap file if a module.modulemap exists OSX has modularized sdk (libc++ included) and there we prefer the modulemaps that libraries ship with. However, libc++ can be also used on other Unix platforms and we should not try to mount the std.modulemap file for it. This patch fixes ROOT-10677. | 17 April 2020, 06:05:24 UTC |
39fe863 | Vassil Vassilev | 15 April 2020, 10:03:25 UTC | [cxxmodules] Preload the dependent modules if specified in rootcling. This should allow us to build non-standard modules such as boost. This patch should enable building the cmssw third-party modules for external dependencies. | 16 April 2020, 17:30:46 UTC |
60814b1 | Axel Naumann | 16 April 2020, 11:59:09 UTC | [cling] Add test for ROOT-10689. | 16 April 2020, 16:12:00 UTC |
573ddb0 | Axel Naumann | 16 April 2020, 10:23:25 UTC | [cling] Preserve AnonStructIds of mangling ctx: Across transactions, lambda types ($_0 etc) must keep their mangling, instead of resetting the counter and recreating the AnonStructIds. Enable that by swapping old and new builder cxxabi mangle context. Fixes ROOT-10689. | 16 April 2020, 16:12:00 UTC |
f9595cd | Axel Naumann | 16 April 2020, 07:04:08 UTC | Shadow only prompt definitions: This prevents an assert (isClingShadowNamespace(D->getDeclContext()) && "D not in a __cling_N5xxx namespace?"), function hideDecl, file /build/jenkins/night/LABEL/mac1014/SPEC/soversion/V/master/root/interpreter/cling/lib/Interpreter/DefinitionShadower.cpp, line 61 | 16 April 2020, 16:11:42 UTC |
929c8e2 | Alexander Voigt | 16 April 2020, 10:40:38 UTC | correcting incorrectly rounded coefficient in dilogarithm expansion in terms of Chebyshev polynomials. Compare with [Yudell L. Luke: Mathematical functions and their approximations, Academic Press Inc., New York 1975, p.67, Table 3.12]: 0.42996693560813697204 0.40975987533077105847 -0.01858843665014591965 0.00145751084062267855 -0.00014304184442340049 0.00001588415541879553 -0.00000190784959386583 0.00000024195180854165 -0.00000003193341274252 0.00000000434545062677 -0.00000000060578480118 0.00000000008612097799 -0.00000000001244331660 0.00000000000182255696 -0.00000000000027006766 0.00000000000004042209 -0.00000000000000610325 0.00000000000000092863 -0.00000000000000014226 0.00000000000000002193 -0.00000000000000000340 0.00000000000000000053 -0.00000000000000000008 0.00000000000000000001 On IEEE-conforming platforms this change does not have any numerical effect, because the two floating point numbers double x = 0.40975987533077105; double y = 0.40975987533077106; have the same binary representation at double precision. | 16 April 2020, 15:27:24 UTC |
8d7989a | Axel Naumann | 18 March 2020, 15:18:51 UTC | [readme] Replace stale README/README by developer info. | 16 April 2020, 14:49:52 UTC |
5bc0acc | Axel Naumann | 11 March 2020, 14:39:30 UTC | [cmake] Do not expose -Werror to external software. | 16 April 2020, 14:49:52 UTC |
0f0121d | Axel Naumann | 19 December 2019, 11:55:27 UTC | [cmake] Introduce `-Ddev=On` with preferred developer settings: Reduce build times: - use lld or gold or good-old-ld, in that order of preference; - split debug info to accelerate rebuilding and relinking; - do not rebuild shared libraries just because a shared lib they depend on has changed; - generate the gdb-index to accelerate root startup in gdb. Turn on -Werror to not overlook warnings. | 16 April 2020, 14:49:52 UTC |
8076501 | Axel Naumann | 19 December 2019, 07:30:06 UTC | [cmake] Remove set-to-self set() invocations. | 16 April 2020, 14:49:52 UTC |
f2268bb | Lorenzo Moneta | 16 April 2020, 09:20:53 UTC | - Add possibility to specify in input RNN string the restGateAfter option | 16 April 2020, 14:38:11 UTC |
2bc6810 | Lorenzo Moneta | 15 April 2020, 21:25:35 UTC | - Fix the full recurrent test (training, i.e. forward + backward) for all type of RNN layers Use a single test program code ( RNN/testFullRNN.h) and remove teh ones in LSTM and GRU - Fix a problem found in updating weights using gradient and learning rate in case of Cudnn - remove an empty file (TestRecurrentForwardPassCuda.cxx) | 16 April 2020, 14:38:11 UTC |
1020d91 | Lorenzo Moneta | 15 April 2020, 07:25:30 UTC | - implement the backpropagation for CPU for teh GRU in case the option reset gate after is on, this means that the candidate gate output is computed as : c(t) = tanh ( W * x(t) + r(t) * ( U * h(t-1) ) + B ) instead of the vanilla GRU implmentation : c(t) = tanh ( W * x(t) + U * ( r(t) * h(t-1) ) + B ) | 16 April 2020, 14:38:11 UTC |
d617127 | Lorenzo Moneta | 14 April 2020, 15:59:38 UTC | - Add the resetGateAfter option in GRU. The Cuddnn implementation of GRU is not the vanilla GRU as described for example in https://en.wikipedia.org/wiki/Gated_recurrent_unit#Fully_gated_unit When computing the candidate gate the multiplication of the candidate weight state is done only with the previous state and the multiplication with the reset state is done afterwards. See https://docs.nvidia.com/deeplearning/sdk/cudnn-api/index.html#cudnnRNNMode_t We use in Cudnn this mode with a single input bias - Ad test for comparing CPU and GPU implementations for all recurrent layer types | 16 April 2020, 14:38:11 UTC |
7ef2bd0 | Olivier Couet | 16 April 2020, 14:20:32 UTC | Improve TMathText Doc (#5372) | 16 April 2020, 14:20:32 UTC |
8492f0a | Michael Wilkinson | 16 April 2020, 14:19:54 UTC | Clarify that TMathText cannot be used in PDF files (#5374) * [TPad][docs] bring Print and SaveAs documentation into better alignment * [TPad][docs] add notes that cannot save TMathText as PDF properly | 16 April 2020, 14:19:54 UTC |
edf1b2c | Olivier Couet | 16 April 2020, 07:16:17 UTC | Avoid non ASCII characters in CREDITS (#5362) * Avoid non ASCII characters in CREDITS * one more non ASCII removed | 16 April 2020, 07:16:17 UTC |
25ed5d7 | Enric Tejedor Saavedra | 15 April 2020, 14:23:01 UTC | [Exp PyROOT] Fix location of _rdf_utils.py | 16 April 2020, 06:32:38 UTC |
f81353f | Enric Tejedor Saavedra | 15 April 2020, 12:55:12 UTC | [Exp PyROOT] Fix naming | 16 April 2020, 06:32:38 UTC |
60209e8 | Enric Tejedor Saavedra | 15 April 2020, 12:53:43 UTC | [Exp PyROOT] Fix error in comment | 16 April 2020, 06:32:38 UTC |
0cc0dc3 | Enric Tejedor Saavedra | 15 April 2020, 12:50:12 UTC | [Exp PyROOT] Ensure Python source files are installed before compiling them Otherwise the compilation fails, since it relies on the Python sources being present in the installation directory. | 16 April 2020, 06:32:38 UTC |
0227cac | Enric Tejedor Saavedra | 15 April 2020, 10:49:24 UTC | [Notebooks] Add installation of libJupyROOTX_Y | 16 April 2020, 06:32:38 UTC |
6a4f1c1 | Enrico Guiraud | 08 April 2020, 09:12:20 UTC | [DF][NFC] Fix typo in comment | 15 April 2020, 15:09:55 UTC |
b92caa0 | Enrico Guiraud | 10 April 2020, 13:17:49 UTC | [DF] Fix wrong "forwarding" in RCustomColumn/RFilter ctors F&& was actually an rvalue reference there. Prefer copy+move over rvalue reference + useless std::forward. | 15 April 2020, 15:09:55 UTC |
3a586fd | Enrico Guiraud | 10 April 2020, 13:17:09 UTC | [DF] Forward rather than move Filter/Define lambdas The move assumed that we always pass lambda temporaries to JitFilterHelper and JitDefineHelper, but that might/will change in the future. | 15 April 2020, 15:09:55 UTC |
98afb6f | Axel Naumann | 14 April 2020, 11:54:29 UTC | [cmake] Bump builtin openssl to 1.1.1f. | 15 April 2020, 11:14:08 UTC |
e7df844 | Enrico Guiraud | 14 April 2020, 21:52:02 UTC | [DF][NFC] Explicitly delete RCustomColumnBase's move ctor It was already implicitly deleted due to the const data members. | 15 April 2020, 08:33:17 UTC |
92dc58d | Enrico Guiraud | 14 April 2020, 21:51:34 UTC | [DF][NFC] Add docstring to RCustomColumnBase data member | 15 April 2020, 08:33:17 UTC |
9be32e3 | Enrico Guiraud | 14 April 2020, 21:35:34 UTC | [DF][NFC] Prefer std::strcmp to strcmp, add missing header | 15 April 2020, 07:46:53 UTC |
171c192 | Enrico Guiraud | 14 April 2020, 17:13:23 UTC | [DF] Remove now unused RLoopManager::GetID and related logic | 15 April 2020, 07:46:53 UTC |
6e2c749 | Enrico Guiraud | 14 April 2020, 17:09:08 UTC | [DF] Always use namespace __rdf for jitting Before this commit, aliases of defined columns were jitted in namespace __rdfX with X monotonously increasing. That is strictly unnecessary, as every alias already contains a unique ID. This commit makes RDF more consistent (everything is jitted in the same namespace) and simplifies some logic and a few signatures. | 15 April 2020, 07:46:53 UTC |
7e71a1d | Enrico Guiraud | 11 April 2020, 16:03:56 UTC | [DF][NFC] Generate jitted code with newlines in the correct places | 14 April 2020, 11:30:45 UTC |
fa68946 | Enric Tejedor Saavedra | 09 April 2020, 09:08:45 UTC | [Exp PyROOT] Check correct python version of libcppyy_backend It is possible that a Python process spawns another Python process that uses a different version of Python, and the second process inherits the parent's environment (seen in a notebook that had a %bash cell which used one of the Python command line tools of ROOT). This is why we need to remove the setting of CPPYY_BACKEND_LIBRARY if it was done by us, to prevent any interference with subprocesses. If the variable was defined by the user, we error out if the current Python version is not the same as the one of the user-defined variable. | 14 April 2020, 10:13:18 UTC |
c9cedcc | Enric Tejedor Saavedra | 27 March 2020, 15:59:57 UTC | [Exp PyROOT][ROOT-8011] Turn libJupyROOT into a C extension module As discussed in ROOT-8011, TPython can't work with the ROOT C++ kernel for Jupyter because the current implementation relies on a plain shared library to run the cell code and capture its output and error. In particular, a crash happens when the library, libJupyROOT, tries to get the __main__ module in TPython's initialization. Turning libJupyROOT into a C extension module makes TPython's initialization succeed, since Python is fully operational. | 14 April 2020, 10:13:18 UTC |
4f99c51 | Stephan Hageboeck | 06 April 2020, 15:58:30 UTC | [Exp PyROOT] Replay "Use pythonised RooWorkspace.Import() in tutorials." Replay 710a13e7b659ecc2eec11d635a512859c9055dca, which got lost in 8c64905e0a12e577eac23ee61323d3736999bef3. | 14 April 2020, 06:43:53 UTC |
9118451 | Vassil Vassilev | 13 April 2020, 17:07:03 UTC | Fix compilation error due to a change merged before. This is an urgent fix to adjust root-project/root@c7ca01f to root-project/root@6b70c86 | 13 April 2020, 19:12:13 UTC |
3ba9e43 | Philippe Canal | 07 April 2020, 22:37:49 UTC | Avoid having to modify TStreamerSTL when streaming it. This fix a "corruption" of the state of TStreamerSTL when using multi-thread. Fix ROOT-10648. See related 777f6b750e657e5c70bdf07095214d59c320e985 and 274c7302ebf95d81255db6d8a02afccb387d3c9e | 13 April 2020, 18:17:36 UTC |
c7ca01f | shahoian | 04 April 2020, 13:24:21 UTC | Avoid {using namespace std} in global scope of dictionary [ROOT-10661] In order to avoid {namespace std} interfering with following include files, when generating a dictionary file add the {using namespace std;} only after GenerateNecessaryIncludes method is called. Since many legacy root header files do require this global {using namespace std;} to complete the tests, old behaviour is kept by default and the global {using namespace std} can be avoided by adding -noGlobalUsingStd to rootcling invocation. Allow passing ACLiC RootCling flags via .rootrc (Vassil) | 13 April 2020, 12:53:03 UTC |
af06a7c | Enrico Guiraud | 10 April 2020, 12:48:45 UTC | [DF] Do not move a RInterface data member in Count The logic here was plain wrong, calling Count() should not modify the state of RInterface. It only happened to work because RActionBase's constructor forgot to actually move-construct its data member (also fixed). | 13 April 2020, 10:48:02 UTC |
6b70c86 | Pratyush Das | 10 April 2020, 09:54:35 UTC | Simplify dictionary stream splitting logic Fixes failing clang-tidy travis instance | 12 April 2020, 11:14:45 UTC |
acfe6dc | Vassil Vassilev | 09 April 2020, 13:06:13 UTC | [math] Always specify the parameter name used in the assert statement. The intent of this code seemed to be avoiding a CINT issue. Now cling can handle such constructs. This fixes an issue with runtime_cxxmodules on OSX sdk where even when NDEBUG is defined the assert still checks the identifier. | 11 April 2020, 17:43:11 UTC |
3398f6d | Olivier Couet | 09 April 2020, 15:23:17 UTC | [Cocoa] Fix compilation problem on MacOsX 10.13 | 10 April 2020, 20:06:43 UTC |
e20506e | Vassil Vassilev | 18 February 2020, 19:26:00 UTC | [cxxmodules] isAbstract can trigger deserialization. | 10 April 2020, 15:44:58 UTC |
71efe47 | Stephan Hageboeck | 09 April 2020, 13:31:37 UTC | [RF] Fix ROOT-10670. Due to changes somewhere in math, the RooMinimizer in ROOT 6.20 falls back to Minuit after using minimize() with a different minimiser. minim.minimize("Minuit2"); minim.hesse(); // Users expected minuit2 here, but now minuit The usual behaviour was restored by remembering the choice of minimiser in `minimize`, which apparently happened behind the scenes in math in the past. | 09 April 2020, 17:21:55 UTC |
23b6cbc | Stephan Hageboeck | 09 April 2020, 13:19:32 UTC | [RF] Fix ROOT-7921. When using the RooCustomizer factory interface, deep trees couldn't be reused in subsequent expressions. | 09 April 2020, 17:21:55 UTC |
0178b88 | Sergey Linev | 09 April 2020, 16:40:01 UTC | [json] properly handle error when reading object (#5330) * [json] properly handle error when reading object If provided pointer type does not match with type read from JSON check inheritance. It can be that pointer type is parent class. In case of error return - do not try to read data for wrong object Co-Authored-By: Philippe Canal <pcanal@fnal.gov> | 09 April 2020, 16:40:01 UTC |
8c64905 | Olivier Couet | 09 April 2020, 13:48:24 UTC | [Roofit Tutorials] consistent headers formatting (#5335) | 09 April 2020, 13:48:24 UTC |
710a13e | Stephan Hageboeck | 06 April 2020, 15:58:30 UTC | [Exp PyROOT] Use pythonised RooWorkspace.Import() in tutorials. | 09 April 2020, 10:53:14 UTC |
e783924 | Stephan Hageboeck | 06 April 2020, 15:39:40 UTC | [Exp PyROOT] Add pythonisation for RooWorkspace. RooWorkspace.import() cannot be called from Python. This adds a pythonisation called RooWorkspace.Import() instead. A unit test was added. | 09 April 2020, 10:53:14 UTC |
31a8023 | alja | 09 April 2020, 03:49:36 UTC | Enable highlight in REveBox and REveBoxProjected | 09 April 2020, 08:00:18 UTC |
22bc120 | Lorenzo Moneta | 08 April 2020, 21:29:31 UTC | Fix also CPU Forward pass test for LSTM layer | 09 April 2020, 07:55:08 UTC |
8e039f0 | Lorenzo Moneta | 08 April 2020, 09:34:03 UTC | Improve and fix rnn forward tests for simple RNN cells and GRU This commit fixes failures observed in incremental builds with debug mode | 09 April 2020, 07:55:08 UTC |
67366ce | Vassil Vassilev | 08 April 2020, 10:21:11 UTC | [genvector] Do not warn if end is unused. This fixes a nightly build issue on OSX with -Druntime_cxxmodules=On by default. There rootcling still needs to parse the assert statement even if NDEBUG is defined. | 09 April 2020, 06:55:00 UTC |
653a18b | alja | 08 April 2020, 18:54:09 UTC | Add missing bbox initalization in REveBoxProjected | 08 April 2020, 20:41:54 UTC |
2eb0093 | Stephan Hageboeck | 07 April 2020, 16:32:55 UTC | [RF] Update tutorial for asymptotic errors. Properly typeset documentation, and shorten output of rf611 tutorial. Make use of feature fixed in previous commit. | 08 April 2020, 18:52:53 UTC |
66d7297 | Stephan Hageboeck | 07 April 2020, 15:22:19 UTC | [RF] Fix ROOT-10668. When asymptotically correct errors are used, the current implementation breaks if variables don't have identical name and title. This is solved by always using the name. | 08 April 2020, 18:52:53 UTC |
6e17597 | Philippe Canal | 07 April 2020, 16:08:42 UTC | Renable TFile::ShrinkCacheFileDir. Use the non static TString::Form to actually modify the cmd string. Use || to actually delete the file when the .ROOT.cachefile does not exist | 08 April 2020, 17:31:20 UTC |
006755a | Stephan Hageboeck | 08 April 2020, 10:08:19 UTC | [RF] Fix tutorial rf212. | 08 April 2020, 13:11:31 UTC |
17c2dc9 | Stephan Hageboeck | 07 April 2020, 14:44:36 UTC | [RF][Docs] Update documentation of RooWorkspace and RooFactoryWSTool. | 08 April 2020, 13:11:31 UTC |
3ba91f9 | Stephan Hageboeck | 07 April 2020, 12:56:12 UTC | [RF] Improve errors and warnings of RooAddPdf/RooAddModel. Improve documentation, warning messages, and replace ineffective asserts by throws. | 08 April 2020, 13:11:31 UTC |
9d2830e | Stephan Hageboeck | 07 April 2020, 12:55:33 UTC | [RF] Modernise RooRecursive fraction, update documentation. | 08 April 2020, 13:11:31 UTC |
4687344 | Stephan Hageboeck | 06 April 2020, 18:17:50 UTC | [RF][Docs] Add section about memory when reading workspaces. | 08 April 2020, 13:11:31 UTC |
62e5626 | Stephan Hageboeck | 26 March 2020, 12:58:29 UTC | [RF] Improve documentation and warnings. Edit docstrings and warnings in Roo{AbsReal,FunctorBinding,RealVar}. | 08 April 2020, 13:11:31 UTC |
52f0ead | Stephan Hageboeck | 16 March 2020, 11:14:09 UTC | [RF] Document constructors of RooRealVar. | 08 April 2020, 13:11:31 UTC |
c931732 | Stefan Wunsch | 08 April 2020, 07:11:58 UTC | [PyROOT] Add libxxhash to known libraries after import ROOT | 08 April 2020, 10:47:26 UTC |
07d54b2 | Massimiliano Galli | 08 April 2020, 07:42:52 UTC | [CMake] PyMVA depends on python-dev too, switch off if not found | 08 April 2020, 09:15:37 UTC |
f62fb2a | Massimiliano Galli | 07 April 2020, 09:05:03 UTC | [CMake] python-dev required only for PyROOT python-dev is required to build PyROOT with multiple Python versions. The reason is that the python related variables are set only if python-dev was found. On the other hand, python-dev is not a requirement for ROOT per se, since it just requires interpreter. With this change we relax the condition on python-dev found and force pyroot and pyroot_exp to off it python-dev was not found. | 08 April 2020, 09:15:37 UTC |
78433bd | Vassil Vassilev | 08 April 2020, 06:58:00 UTC | [math] Do not warn if len is unused. This fixes a nightly build issue on OSX with -Druntime_cxxmodules=On by default. There rootcling still needs to parse the assert statement even if NDEBUG is defined. | 08 April 2020, 08:46:21 UTC |
56038f8 | Enrico Guiraud | 07 April 2020, 09:57:56 UTC | [DF][NFC] Bring aliases in scope rather than re-defining them | 08 April 2020, 07:19:59 UTC |
2a446bf | Enrico Guiraud | 07 April 2020, 09:45:13 UTC | [DF][NFC] Remove unused include | 08 April 2020, 07:19:59 UTC |
5901099 | Enrico Guiraud | 07 April 2020, 09:09:35 UTC | [DF][NFC] Remove obsolete comment | 08 April 2020, 07:19:59 UTC |
6287d8b | Vassil Vassilev | 02 April 2020, 17:19:53 UTC | [travis] Adjust auto clang-format command. | 08 April 2020, 05:07:50 UTC |
9b6df8c | Vassil Vassilev | 02 April 2020, 15:07:35 UTC | Resolve dependencies and relink if explicit linking is on. This patch fixes ROOT-10637. | 08 April 2020, 05:07:50 UTC |
b246f83 | Vassil Vassilev | 30 March 2020, 10:45:56 UTC | [dyld] Blacklist libcppyy_backend* from symbol resolution. We host both libcppyy_backend2.7 and libcppyy_backend3 which are python2 and python3 compatible libraries. We must not resolve symbols from there as we can choose the wrong one. The symbols are managed by libROOTPython | 08 April 2020, 05:07:50 UTC |
8fdd5c3 | Vassil Vassilev | 27 March 2020, 11:44:56 UTC | [core] Delegate loading of the rootmap file to GetSharedLibDeps. NFC. Instead of calling GetSharedLibDeps and then externally figuring out if a rootmap file exists, load it again and call again GetSharedLibDeps, this patch does it in one call. This improves the encapsulation and should keep code which uses TCling-details in TCling. | 08 April 2020, 05:07:50 UTC |
35af5c8 | Vassil Vassilev | 28 March 2020, 15:07:00 UTC | [core] Use the resolved library path. GetSharedLibDeps may need to resolve to the absolute path, however, if we give only the file stem GetSharedLibDeps cannot support features such as resolving of relative path to $ROOTSYS/lib/. For example, `cd /home/`; root.exe -e 'gSystem->Load("../test/libEvent.so")'. `TSystem::Load` will call an interface passing the relative path which will resolve to $ROOTSYS/lib/../test/libEvent.so, however, if GetSharedLibDeps calls the same interface it will give the wrong results. This patch fixes this behavior and optimizes GetSharedLibDeps when we have already the full path. | 08 April 2020, 05:07:50 UTC |
02233d5 | Pratyush Das | 03 April 2020, 19:36:47 UTC | [metacling]Refactor Property into base class of TClingTypeInfo and TClingDataMemberInfo by using pass by reference. | 07 April 2020, 16:04:04 UTC |
d7c80b4 | Pratyush Das | 04 April 2020, 09:40:29 UTC | [metacling]Make Name const by making fNameCache mutable. This allows us to preserve the semantic constness of the interfaces. | 07 April 2020, 16:04:04 UTC |
36a9a1d | Pratyush Das | 03 April 2020, 19:39:53 UTC | [metacling]isArray: Add bitwise operator in Property The interface went out of sync due to code duplication. TClingDataMemberInfo handled isArray but TClingTypeInfo, TClingMethodArgInfo, TClingMethodInfo and TClingTypedefInfo did not. | 07 April 2020, 16:04:04 UTC |
1295c19 | Vassil Vassilev | 06 April 2020, 08:23:08 UTC | Fix a typo. Co-Authored-By: Stephan Hageboeck <stephan.hageboeck@cern.ch> | 07 April 2020, 14:58:37 UTC |
fc4dc5d | Vassil Vassilev | 05 April 2020, 08:57:54 UTC | Do not special case solaris. Ancient rootcint had to make a special case for it because it was done too early (see root-project/root@6e3d084). The code in rootcling has changed significantly and we do it after inclusion of header files which also may have had such a construct. | 07 April 2020, 14:58:37 UTC |
6f0c31e | Vassil Vassilev | 07 April 2020, 11:15:10 UTC | [cxxmodules] Make C++ modules default on OSX for c++11. | 07 April 2020, 13:58:02 UTC |
1e6f8bd | Olivier Couet | 07 April 2020, 13:37:26 UTC | [Graphics] avoid double delete in TRatioPlot dtor (#5312) | 07 April 2020, 13:37:26 UTC |
fbbd72f | Stefan Wunsch | 03 April 2020, 07:30:36 UTC | [PyROOT] Make namespace pythonizations lazy * Remove dependency on VecOps, RDataFrame and TMVA for import ROOT * Inject namespace overload via property in the ROOT facade * Take care that we have proper error handling if, e.g., TMVA is not there | 07 April 2020, 12:04:49 UTC |
7ed0c4d | Stefan Wunsch | 31 March 2020, 08:51:57 UTC | [PyROOT] Add test checking loaded libraries after import | 07 April 2020, 12:04:49 UTC |
66916b6 | moneta | 23 March 2020, 10:04:00 UTC | Various improvements in the documentation for TH1::Fit, TGraph/TGraph2D/TMultiGraph::Fit. - The documentation for the TGrah's EX0 fit option is clarified as suggested in https://root-forum.cern.ch/t/fitting-tgraphasymerrors-difference-between-w-and-ex0/38522 - Expalin better the meaning of the documentation of the W and WW fit option for TH1 and TGraph's In additon fix in the code the case of using the fit options W or WW together with the option P. The option P is Pearson chi2fit and it does not make sense to ignore errors in that case. Add a warning message to signal the user this case | 07 April 2020, 09:00:02 UTC |
ab7f6e1 | Massimiliano Galli | 06 April 2020, 15:09:44 UTC | [Exp PyROOT] Change JupyROOT.utils to JupyROOT.helpers.utils | 07 April 2020, 06:51:55 UTC |
ad9c998 | Massimiliano Galli | 06 April 2020, 14:50:46 UTC | [Exp PyROOT] Change libcppyy.CPPOverload._threaded to libcppyy.CPPOverload.__release_gil__ | 07 April 2020, 06:51:55 UTC |
9af3726 | Enrico Guiraud | 06 April 2020, 21:03:14 UTC | [DF][NFC] Use GetThreadPoolSize instead of GetImplicitMTPoolSize | 06 April 2020, 21:42:15 UTC |
3c37f0a | Enrico Guiraud | 06 April 2020, 16:20:22 UTC | [IO] Remove now unused TFile::fgRwLock | 06 April 2020, 17:39:38 UTC |
180b266 | Enrico Guiraud | 06 April 2020, 15:06:12 UTC | [IMT] Remove now unused TParTreeProcessingRAII and related functions | 06 April 2020, 17:39:38 UTC |
1259ce4 | Enrico Guiraud | 06 April 2020, 15:06:34 UTC | [TreeProcMT] No need to use TParTreeProcessingRAII anymore The code it was protecting is now always protected by gCoreMutex. | 06 April 2020, 17:39:38 UTC |
901c0e5 | Enrico Guiraud | 06 April 2020, 15:02:14 UTC | [IO] Always protect MT usage of TFile Currently, some sections of TFile::ReadProcessID are only protected from concurrent access if Internal::IsParTreeProcessingEnabled() is true. The only entity that enables ParTreeProcessing in ROOT is TTreeProcessorMT::Process. With this patch, concurrent access to TFile::ReadProcessID is always protected. | 06 April 2020, 17:39:38 UTC |
b753284 | Enrico Guiraud | 03 April 2020, 13:35:36 UTC | [DOCS] Add deprecation of GetImplicitMTPoolSize to release notes | 06 April 2020, 17:00:45 UTC |
a73a2f9 | Enrico Guiraud | 03 April 2020, 13:27:18 UTC | [MT] Deprecate GetImplicitMTPoolSize in favor of GetThreadPoolSize TThreadExecutor does not activate implicit multi-threading (ROOT::IsImplicitMTEnabled() is false after constructing a TThreadExecutor) but it does change the size of the thread-pool, which is not only used by implicit multi-threading features, but also by TThreadExecutor and TTreeProcessorMT. So the thread pool is not the "ImplicitMTPool" but it's really ROOT's one and only pool of threads, so we prefer ROOT::GetThreadPoolSize to GetImplicitMTPoolSize. | 06 April 2020, 17:00:45 UTC |
7b9e833 | Enrico Guiraud | 03 April 2020, 13:25:40 UTC | [MT] Add ROOT::GetThreadPoolSize ...and implement GetImplicitMTPoolSize in terms of it. | 06 April 2020, 17:00:45 UTC |