Philippe Canal

> Done. I've also mentioned the ambiguity between `EProperty` and `::EProperty`, should we ever come across this code again. Thanks.

The failure is [execCheckClusterRange](https://lcgapp-services.cern.ch/root-jenkins/job/root-pullrequests-build/195807/testReport/projectroot.roottest.root.tree/fastcloning/roottest_root_tree_fastcloning_execCheckClusterRange/) is surprising.

The first step is to compare the result of `TFile::Map("forcomp");` for both files.

> It seems that the TMemFile::Write calls many times the GetClassSharedLibs function, and the resulting StreamerInfo is not exactly the same, see trace here: That's okay and it should not...

``` root -l 'openTree.C(0)' -b -q Processing openTree.C(0)... 7.31217e-322ºC root -l 'openTree.C(1)' -b -q Processing openTree.C(1)... 20.2826ºC ``` This is the expected result. `TBranch::SetAddress` is a lower level interface and...

This is a missing feature in `TClingCallbacks::tryFindROOTSpecialInternal` (see core/metacling/src/TClingCallbacks.cxx:733) and/or `TCling__GetObjectAddress` which assumes that the object retrieved inherits from `TObject` (as the left most base).

> There are a few GitHub and JIRA issues related to this feature that can also be closed if we completely remove it. Please list those here. Thanks.

> For the warning messages, I'm reusing one of the existing clang warning message (to not introduce more patches on top of the clang with custom error messages). Can you...

The errors are unrelated.

One of the parts/elements of the behavior is that: ``` cling::utils::Lookup::Tag(&fInterpreter->getSema(), name, dc); ``` with ``` name == "E" and dc == decl_for("B") ``` and with `B`: ``` NamespaceDecl 0x156e8b600...