[mlir] Change generate-test-checks.py to `%[[V:.*]]` style

[silvas]

Summary

This style seems to be the preferred style lately as it plays better
with op syntaxes that include [ and ].

Input:

module {
  func @f(%arg0 : i32) -> i32 {
    %0 = addi %arg0, %arg0 : i32
    %1 = muli %arg0, %0 : i32
    return %1 : i32
  }
}

New output:

// NOTE: Assertions have been autogenerated by utils/generate-test-checks.py
// CHECK:       module {

// CHECK-LABEL:   func @f(
// CHECK-SAME:            %[[VAL_0:.*]] : i32) -> i32 {
// CHECK:           %[[VAL_1:.*]] = addi %[[VAL_0]], %[[VAL_0]] : i32
// CHECK:           %[[VAL_2:.*]] = muli %[[VAL_0]], %[[VAL_1]] : i32
// CHECK:           return %[[VAL_2]] : i32
// CHECK:         }
// CHECK:       }

Old output:

// NOTE: Assertions have been autogenerated by utils/generate-test-checks.py
// CHECK:       module {

// CHECK-LABEL:   func @f(
// CHECK-SAME:            [[VAL_0:%.*]] : i32) -> i32 {
// CHECK:           [[VAL_1:%.*]] = addi [[VAL_0]], [[VAL_0]] : i32
// CHECK:           [[VAL_2:%.*]] = muli [[VAL_0]], [[VAL_1]] : i32
// CHECK:           return [[VAL_2]] : i32
// CHECK:         }
// CHECK:       }

[ftynse]

I prefer this style because it makes clear what's value and what's something else when also pattern-matching named attributes for example.

[nicolasvasilache]

On my end, the biggest benefit I see is when captured values are used to index.
FileCheck does not like [[[ and so one must resort to heavy escaping sequences.
[%[[ works much better.

mlir/utils/generate-test-checks.py

for chunk in line_chunks:
     variable = None
     for scope in variable_namer.scopes:
       variable = scope.get(ssa_name)
       if variable is not None:
         break
 
     # If one exists, then output the existing name.
     if variable is not None:
-      output_line += '[[' + variable + ']]'
+      output_line += '%[[' + variable + ']]'
     else:
       # Otherwise, generate a new variable.
       variable = variable_namer.generate_name(ssa_name)
-      output_line += '[[' + variable + ':%.*]]'
+      output_line += '%[[' + variable + ':.*]]'
 
     # Append the non named group.
     output_line += chunk[len(ssa_name):]
 
   return output_line + '\n'
 
 
 # Pre-process a line of input to remove any character sequences that will be