C++: Fix performance of leap year queries

The `sameBaseType` predicate was fundamentally quadratic, and this blew
up on large C++ code bases. Replacing it with calls to `Type.stripType`
fixes performance and does not affect the qltests. It looks like
`sameBaseType` was used purely an ad hoc heuristic, so I'm not worried
about the slight semantic difference between `sameBaseType` and
`stripType`.

Author: jbj

cpp/ql/src/semmle/code/cpp/commons/StructLikeClass.qll

@@ -59,7 +59,7 @@ predicate setter(MemberVariable v, MemberFunction f, Class c) {
       v2 != v
     ) and
     f.getNumberOfParameters() = 1 and
-    sameBaseType(f.getParameter(0).getType(), v.getType())
+    f.getParameter(0).getType().stripType() = v.getType().stripType()
 }
 
 /**
@@ -77,25 +77,5 @@ predicate getter(MemberVariable v, MemberFunction f, Class c) {
       v2 != v
     ) and
     f.getNumberOfParameters() = 0 and
-    sameBaseType(f.getType(), v.getType())
-}
-
-/**
- * Holds if `t1` and `t2` are the same type up to typedefs, specifiers,
- * and removing a single layer of pointers or references (but not arrays).
- * Equates, for example, `const int*` with `int`, but not `int**` with `int`
- * or `int[]` with `int`.
- */
-predicate sameBaseType(Type t1, Type t2) {
-  exists (Type base1, Type base2 |
-    base1 = t1.getUnspecifiedType() and
-    base2 = t2.getUnspecifiedType().getUnspecifiedType() and
-    (
-      base1 = base2 or
-      base1.(PointerType).getBaseType() = base2 or
-      base2.(PointerType).getBaseType() = base1 or
-      base1.(ReferenceType).getBaseType() = base2 or
-      base2.(ReferenceType).getBaseType() = base1
-    )
-  )
+    f.getType().stripType() = v.getType().stripType()
 }