[clang] [Clang][Diagnostics] Update select uses in DiagnosticXKinds.td to use enum_select (PR #130868)

[Erich Keane via cfe-commits]

================
@@ -131,17 +131,21 @@ def note_constexpr_past_end : Note<
   "dereferenced pointer past the end of %select{|subobject of }0"
   "%select{temporary|%2}1 is not a constant expression">;
 def note_constexpr_past_end_subobject : Note<
-  "cannot %select{access base class of|access derived class of|access field of|"
-  "access array element of|ERROR|"
-  "access real component of|access imaginary component of}0 "
-  "pointer past the end of object">;
+  "cannot %enum_select<AccessType>{"
+    "%BaseClassAccess{access base class of}|%DerivedClassAccess{access derived class of}|"
+    "%FieldAccess{access field of}|%ArrayElementAccess{access array element of}|"
+    "%ErrorAccess{ERROR}|"
+    "%RealComponentAccess{access real component of}|%ImaginaryComponentAccess{access imaginary component of}"
+  "}0 pointer past the end of object">;
 def note_non_null_attribute_failed : Note<
   "null passed to a callee that requires a non-null argument">;
 def note_constexpr_null_subobject : Note<
-  "cannot %select{access base class of|access derived class of|access field of|"
-  "access array element of|perform pointer arithmetic on|"
-  "access real component of|"
-  "access imaginary component of}0 null pointer">;
+  "cannot %enum_select<NullSubobjectAccess>{"
----------------
erichkeane wrote:

>We recently discussed allowing ‘redeclaring’ an %enum_select iff the enumerators and text are the same across all uses of %enum_select with the same enumeration. I think it might additionally make sense to allow just e.g. %enum_select<T>, that is e.g. cannot %enum_select<AccessType>0 null pointer, to reference an existing %enum_select. That should be fairly simple to implement; I finally have some more time on my hands, so I could come up w/ a patch for that later today.

Hmm... pulling in the same text as well is interesting...  What we discussed was pulling in the same `%enum_select` iif the enumerators were the same across all uses, not if the text itself was.  I think ensuring the text is the same isn't necessary, and is actually problematic.

The re-use of the same text like that is interesting....  I'd not thought of doing something like that (without the curleys), but I'm a touch afraid it harms readability of the diagnostic.

https://github.com/llvm/llvm-project/pull/130868