[Mlir-commits] [mlir] [mlir][python] Fixup the constantTypes of pdl.TypesOp (PR #75812)

[Jiefeng Wang]

jiefwo wrote:

After this commit [[MLIR][python bindings] Add some AttrBuilder and port _exts to use them.](https://reviews.llvm.org/D149287), this python script produces mlir as follows. It's `%0 = types : []` rather than `%0 = types` which may result in PDL interpreter failure.
The checks `# CHECK:     %0 = types` in `llvm-project/mlir/test/python/dialects/pdl_ops.py` only match the result partially, not the whole line, so this test still passes.
```python
from mlir.ir import *
from mlir.dialects.pdl import *

def constructAndPrintInModule(f):
    print("\nTEST:", f.__name__)
    with Context(), Location.unknown():
        module = Module.create()
        with InsertionPoint(module.body):
            f()
        print(module)
    return f

@constructAndPrintInModule
def test_operation_results():
    valueRange = RangeType.get(ValueType.get())
    pattern = PatternOp(1, "operation_results")
    with InsertionPoint(pattern.body):
        types = TypesOp()
        inputOp = OperationOp(types=[types])
        results = ResultsOp(valueRange, inputOp)
        root = OperationOp(args=[results])
        RewriteOp(root, name="rewriter")
```

```mlir
TEST: test_operation_results
module {
  pdl.pattern @operation_results : benefit(1) {
    %0 = types : []
    %1 = operation  -> (%0 : !pdl.range<type>)
    %2 = results of %1
    %3 = operation(%2 : !pdl.range<value>)
    rewrite %3 with "rewriter"
  }
}
```

https://github.com/llvm/llvm-project/pull/75812