Matti Kortelainen
Matti Kortelainen
@cmsbuild, please close
assign geometry, visualization
> @makortel > Do I need to open an issue to add unit tests? > > ( PS: I'm on vacation starting tomorrow until Tuesday) I'm not sure I understand...
The case @smuzaffar quoted above is actually CPU-only code (even if it uses Alpaka), so the problem is not specific to CUDA/ROCm. The full build log shows a crash also...
Would it be helpful to repeat the test with debug symbols enabled for cling (and ROOT)?
> > I think this mostly happens when cuda code is envolved > > Why we put cuda in dictionaries? We have classes (data products) that use CUDA (or ROCm)...
The `DataFormats/Histograms` dictionary generation fails with assertion failure ``` rootcling: /data/cmsbld/jenkins/workspace/ib-run-pr-tests/testBuildDir/BUILD/el8_amd64_gcc12/lcg/root/6.33.01-48fd78c793bc73bd3f87499c5ca7bf90/root-6.33.01/interpreter/cling/lib/Utils/AST.cpp:1561: clang::NestedNameSpecifier* cling::utils::CreateNestedNameSpecifierForScopeOf(const clang::ASTContext&, const clang::Decl*, bool, const clang::ClassTemplateSpecializationDecl*): Assertion `Spec && "Cannot determine type's scope"' failed. #0 0x00001529359d3acc llvm::sys::PrintStackTrace(llvm::raw_ostream&,...
> That looks like a bug in llvm... I wonder on what code it happens. I see there is a debug llvm build already. What's the way nowadays to debug...
The `DataFormats/Histograms` failure seems to occur when the `rootcling` is processing [`MEtoEDM::MEtoEDMObject`](https://github.com/cms-sw/cmssw/blob/CMSSW_14_2_X_2024-09-03-2300/DataFormats/Histograms/src/classes_def.xml#L47)