dialects: (Transform) Add new selecting operations #3015
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #3015 +/- ##
==========================================
- Coverage 89.84% 89.81% -0.04%
==========================================
Files 411 411
Lines 51641 51892 +251
Branches 8027 8071 +44
==========================================
+ Hits 46399 46608 +209
- Misses 3961 3986 +25
- Partials 1281 1298 +17 ☔ View full report in Codecov by Sentry. |
| "transform.named_sequence"() <{arg_attrs = [{transform.readonly}], function_type = (!transform.any_op) -> (), sym_name = "foo"}> ({ | ||
| ^bb0(%arg0: !transform.any_op): | ||
| "transform.yield"() : () -> () | ||
| }) : () -> () |
There was a problem hiding this comment.
Please use custom syntax in tests when possible, it makes them more readable
There was a problem hiding this comment.
I don't really see where custom syntax can be used in this case. are there any particular operations you're referring to?
There was a problem hiding this comment.
I don't have time for a super thorough review, but this looks very nice! @jorendumoulin ?
tests/filecheck/mlir-conversion/with-mlir/dialects/transform/transform_types.mlir
Outdated
Show resolved
Hide resolved
|
This is overall looking good, there are just small changes remaining. Could I just ask you to please make smaller pull requests in the future? A file that needs to be expanded explicitly in the GitHub UI is usually a signal to split up the changes into smaller diffs. For example, for this we could easily have merged the new attributes first, then the new operations. |
In this PR, 13 new core operations get added to the transform dialect. All of them useful to find and select the correct operation in a transformation. They are:
GetConsumerOfResult, GetDefiningOp, GetParentOp, GetProducerOfOperand, GetResultOp, GetTypeOp, IncludeOp, MatchOperationEmptyOp, MatchOperationNameOp, MatchParamCmpIOp, MergeHandlesOp, ParamConstantOp, SplitHandleOp,
All operations are tested using filecheck and pytest