[Mlir-commits] [mlir] [mlir][transform] Overhaul `RegionBranchOpInterface` implementations. (PR #111408)

[Matthias Springer]

Ingo =?utf-8?q?Müller?= <ingomueller at google.com>
Message-ID:
In-Reply-To: <llvm.org/llvm/llvm-project/pull/111408 at github.com>


================
@@ -104,16 +104,8 @@ transform::AlternativesOp::getEntrySuccessorOperands(RegionBranchPoint point) {
 
 void transform::AlternativesOp::getSuccessorRegions(
     RegionBranchPoint point, SmallVectorImpl<RegionSuccessor> &regions) {
-  for (Region &alternative : llvm::drop_begin(
----------------
matthias-springer wrote:

Yes.

I just took a look at `scf.forall`:

```
void ForallOp::getSuccessorRegions(RegionBranchPoint point,
                                   SmallVectorImpl<RegionSuccessor> &regions) {
  // Both the operation itself and the region may be branching into the body or
  // back into the operation itself. It is possible for loop not to enter the
  // body.
  regions.push_back(RegionSuccessor(&getRegion()));
  regions.push_back(RegionSuccessor());
}
```

Maybe the entire modeling of operands could be optional.


https://github.com/llvm/llvm-project/pull/111408