[meta] Return exec id of TRef, TRefArray and derived data members

[dpiparo]

Instead of all data members the type of which has a name beginning with "TRef".

Fixes ROOT-7052

This PR builds on top of the solid work done here #14930.

[ferdymercury]

Thanks a lot! Optional idea: could you add and updates fClassObject. on line 287 ?

[github-actions]

Test Results

    18 files      18 suites   4d 4h 30m 21s ⏱️
 2 731 tests  2 731 ✅ 0 💤 0 ❌
47 768 runs  47 768 ✅ 0 💤 0 ❌

Results for commit 62a9c3a.

♻️ This comment has been updated with latest results.

[dpiparo]

Yes, I focussed on the functional change before. I'll make that part of the new version of the PR after the review.

[ferdymercury]

@olifre Do you happen to still have a reproducer/test for this issue of yours with a class CBTRef deriving from TRef, that you could share?

[olifre]

@olifre Do you happen to still have a reproducer/test for this issue of yours with a class CBTRef deriving from TRef, that you could share?

Sadly, I have no example at hand 😢 . As far as I remember, I think I tried this in the context of a larger framework with classes inheriting from TRef and then did not continue to create a minimal reproducer, sorry.

[dpiparo]

Just added a test.

[pcanal]

Why the name TRefFoo?

[pcanal]

Nevermind this the point of the test :)

Suggested change

	"class TRefFoo {int i;};\n"
	"// We are testing that this is not confused for a ROOT TRef or TRefArray
	"class TRefFoo {int i;};\n"

[dpiparo]

:)

[pcanal]

Thanks.