shafik added inline comments.
================
Comment at: clang/lib/AST/Interp/ByteCodeExprGen.cpp:259
+
+ if (Op != BO_Add && Op != BO_Sub)
+ return false;
----------------
or neither left or right operand is a pointer type
================
Comment at: clang/lib/AST/Interp/ByteCodeExprGen.cpp:271
+
+ if (LHS->getType()->isPointerType() && RHS->getType()->isPointerType()) {
+ assert(E->getType()->isIntegerType());
----------------
If it is two pointers then it must be a subtraction.
================
Comment at: clang/lib/AST/Interp/ByteCodeExprGen.cpp:280
+ PrimType OffsetType;
+ if (LHS->getType()->isIntegerType() && RHS->getType()->isPointerType()) {
+ if (!visit(RHS) || !visit(LHS))
----------------
================
Comment at: clang/lib/AST/Interp/ByteCodeExprGen.cpp:284
+ OffsetType = *LT;
+ } else {
+ if (!visit(LHS) || !visit(RHS))
----------------
================
Comment at: clang/lib/AST/Interp/ByteCodeExprGen.cpp:288
+ OffsetType = *RT;
+ }
+
----------------
================
Comment at: clang/lib/AST/Interp/ByteCodeExprGen.cpp:241
+ // Pointer arithmethic special case. This is supported for one of
+ // LHS and RHS being a pointer type and the other being an integer type.
+ if (BO->getType()->isPointerType()) {
----------------
tbaeder wrote:
> shafik wrote:
> > I am not sure if this is the right place to handle this but there are a
> > bunch of other cases.
> >
> > - `nullptr` can have `0` added or subtracted
> > - You can only do addition/subtraction from a pointer if the result in
> > within bounds or one after the end
> > - You can subtract two pointers if they point to the same object.
> >
> > godbolt: https://godbolt.org/z/5YTY93z8M
> I will probably move this special case out of the function and into its own.
> Thanks for the tests, I think this should all already work except for adding
> two pointers. I'll update the diff once I checked that out.
You will also want to cover this case as well:
https://eel.is/c++draft/expr.add#6
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D135858/new/
https://reviews.llvm.org/D135858
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
