[NFC] Fix builds on recent GCC with C++20 enabled

[alexbatashev]

Summary

The following pattern fails on recent GCC versions with -std=c++20 flag passed
and succeeds with -std=c++17. Such behavior is not observed on Clang 16.0.

c++
template <typename T>
struct Foo {
    Foo<T>(int a) {}
};

This patch removes template parameter from constructor in two occurences to
make the following command complete successfully:
bazel build -c fastbuild --cxxopt=-std=c++20 --host_cxxopt=-std=c++20 @llvm-project//llvm/...

[alexbatashev]

A brief note for reviewers. This appears to be a bug in either Clang or GCC. Briefly looking at recent C++ defect reports, I couldn't find anything like this. So, I lean towards a bug in GCC 12 and 13 (I didn't check other versions). But if someone knows what the rootcause is, please, let me know.
A link to Compiler Explorer with the reproducer to play around: https://godbolt.org/z/1fKqdfnrM

[jdoerfert]

llvm/lib/Transforms/IPO/AttributorAttributes.cpp

3475

This part LGTM.

[Hahnfeld]

Thanks for fixing it! I also discovered it today... I think this is according to the standard: https://eel.is/c++draft/diff.cpp17.class#2 I guess Clang should warn about this; and GCC's error message isn't helpful at all (since version 11 btw)

[alexbatashev]

In D154782#4511812, @Hahnfeld wrote:

Thanks for fixing it! I also discovered it today... I think this is according to the standard: https://eel.is/c++draft/diff.cpp17.class#2 I guess Clang should warn about this; and GCC's error message isn't helpful at all (since version 11 btw)

Totally agree. Did you already submit bugs to both projects?

[Hahnfeld]

In D154782#4511815, @alexbatashev wrote:

In D154782#4511812, @Hahnfeld wrote:

Thanks for fixing it! I also discovered it today... I think this is according to the standard: https://eel.is/c++draft/diff.cpp17.class#2 I guess Clang should warn about this; and GCC's error message isn't helpful at all (since version 11 btw)

Totally agree. Did you already submit bugs to both projects?

No, didn't get to it yet. I didn't realize at first that Clang just silently accepts it...

@tstellar also FYI. I know that there are no more point releases planned for 16.0, but the fix in llvm/lib/Transforms/IPO/AttributorAttributes.cpp also affects GCC + C++20 builds of release/16.x. If there ever is another release of 16.0, this one would be great to include :)

llvm/include/llvm/CodeGen/ByteProvider.h

 
 /// Represents known origin of an individual byte in combine pattern. The
 /// value of the byte is either constant zero, or comes from memory /
 /// some other productive instruction (e.g. arithmetic instructions).
 /// Bit manipulation instructions like shifts are not ByteProviders, rather
 /// are used to extract Bytes.
 template <typename ISelOp> class ByteProvider {
 private:
-  ByteProvider<ISelOp>(std::optional<ISelOp> Src, int64_t DestOffset,
-                       int64_t SrcOffset)
+  ByteProvider(std::optional<ISelOp> Src, int64_t DestOffset, int64_t SrcOffset)
       : Src(Src), DestOffset(DestOffset), SrcOffset(SrcOffset) {}
 
   // TODO -- use constraint in c++20
   // Does this type correspond with an operation in selection DAG
   template <typename T> class is_op {
   private:
     using yes = std::true_type;
     using no = std::false_type;

llvm/lib/Transforms/IPO/AttributorAttributes.cpp

 } // namespace llvm
 
 namespace {
 
 template <typename BaseTy, typename ToTy>
 struct CachedReachabilityAA : public BaseTy {
   using RQITy = ReachabilityQueryInfo<ToTy>;
 
-  CachedReachabilityAA<BaseTy, ToTy>(const IRPosition &IRP, Attributor &A)
-      : BaseTy(IRP, A) {}
+  CachedReachabilityAA(const IRPosition &IRP, Attributor &A) : BaseTy(IRP, A) {}
 
   /// See AbstractAttribute::isQueryAA.
   bool isQueryAA() const override { return true; }
 
   /// See AbstractAttribute::updateImpl(...).
   ChangeStatus updateImpl(Attributor &A) override {
     ChangeStatus Changed = ChangeStatus::UNCHANGED;
     InUpdate = true;