[UTC] Support to test annotated IR #165419

Camsyn · 2025-10-28T15:35:56Z

Some analysis/transformation, e.g., predicate info/ mem ssa, insert instruction annotations as comments, referring to #165249 (comment).

This PR makes UTC support checking these instruction annotations with an extra UTC option -check-inst-comments.

E.g.,
Before:

; CHECK:         [[Z_0:%.*]] = bitcast i1 [[Z]] to i1

After:

; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[Z]] = and i1 [[XZ]], [[YZ]] Edge: [label [[TMP0:%.*]],label %nope], RenamedOp: [[Z]] }
; CHECK-NEXT:    [[Z_0:%.*]] = bitcast i1 [[Z]] to i1

This PR also regenerates all UTC-generated tests for PredicateInfo; No MemSSA test is updated, as there are no UTC-generated tests designated for print<memoryssa>.

llvmbot · 2025-10-28T15:36:39Z

@llvm/pr-subscribers-llvm-transforms

@llvm/pr-subscribers-testing-tools

Author: Kunqiu Chen (Camsyn)

Changes

Some analysis/transformation, e.g., predicate info/ mem ssa, insert annotations as comments, referring to #165249 (comment).

This PR makes UTC support checking these annotations with an extra UTC option -check-comments.

E.g.,
Before:

; CHECK:         [[Z_0:%.*]] = bitcast i1 [[Z]] to i1

After:

; CHECK-NEXT:  ; Has predicate info
; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[Z]] = and i1 [[XZ]], [[YZ]] Edge: [label [[TMP0:%.*]],label %nope], RenamedOp: [[Z]] }
; CHECK-NEXT:    [[Z_0:%.*]] = bitcast i1 [[Z]] to i1

This PR also regenerates all UTC-generated tests for PredicateInfo; No MemSSA test is updated, as there are no UTC-generated tests designated for print<memoryssa>.

Patch is 76.69 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/165419.diff

9 Files Affected:

(modified) llvm/test/Transforms/Util/PredicateInfo/branch-on-same-cond.ll (+22-15)
(modified) llvm/test/Transforms/Util/PredicateInfo/condprop.ll (+82-28)
(modified) llvm/test/Transforms/Util/PredicateInfo/diamond.ll (+14-6)
(modified) llvm/test/Transforms/Util/PredicateInfo/edge.ll (+39-21)
(modified) llvm/test/Transforms/Util/PredicateInfo/pr33456.ll (+12-7)
(modified) llvm/test/Transforms/Util/PredicateInfo/pr33457.ll (+17-10)
(modified) llvm/test/Transforms/Util/PredicateInfo/testandor.ll (+320-108)
(modified) llvm/utils/UpdateTestChecks/common.py (+19-5)
(modified) llvm/utils/update_test_checks.py (+8)

diff --git a/llvm/test/Transforms/Util/PredicateInfo/branch-on-same-cond.ll b/llvm/test/Transforms/Util/PredicateInfo/branch-on-same-cond.ll
index 0be13ee76bece..3d4cf57d734a3 100644
--- a/llvm/test/Transforms/Util/PredicateInfo/branch-on-same-cond.ll
+++ b/llvm/test/Transforms/Util/PredicateInfo/branch-on-same-cond.ll
@@ -1,4 +1,4 @@
-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-comments
 ; RUN: opt -S -passes=print-predicateinfo < %s 2>&1 >/dev/null | FileCheck %s
 
 ; FIXME:  RenamedOp should be %cmp or %x in all cases here,
@@ -9,24 +9,31 @@ define i32 @test(i32 %x) {
 ; CHECK-NEXT:    br label [[BB1:%.*]]
 ; CHECK:       bb1:
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp eq i32 [[X:%.*]], 0
-; CHECK: RenamedOp: [[CMP]]
-; CHECK:         [[CMP_0:%.*]] = bitcast i1 [[CMP]] to i1
-; CHECK: RenamedOp: [[X]]
-; CHECK:         [[X_0:%.*]] = bitcast i32 [[X]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[X]], 0 Edge: [label [[BB1]],label %bb2], RenamedOp: [[CMP]] }
+; CHECK-NEXT:    [[CMP_0:%.*]] = bitcast i1 [[CMP]] to i1
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[X]], 0 Edge: [label [[BB1]],label %bb2], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_0:%.*]] = bitcast i32 [[X]] to i32
 ; CHECK-NEXT:    br i1 [[CMP]], label [[BB2:%.*]], label [[EXIT1:%.*]]
 ; CHECK:       bb2:
-; CHECK: RenamedOp: [[CMP_0]]
-; CHECK:         [[CMP_0_1:%.*]] = bitcast i1 [[CMP_0]] to i1
-; CHECK: RenamedOp: [[X]]
-; CHECK:         [[X_0_1:%.*]] = bitcast i32 [[X_0]] to i32
-; CHECK: RenamedOp: [[X_0]]
-; CHECK:         [[X_0_4:%.*]] = bitcast i32 [[X_0]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[X]], 0 Edge: [label [[BB2]],label %bb3], RenamedOp: [[CMP_0]] }
+; CHECK-NEXT:    [[CMP_0_1:%.*]] = bitcast i1 [[CMP_0]] to i1
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[X]], 0 Edge: [label [[BB2]],label %bb3], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_0_1:%.*]] = bitcast i32 [[X_0]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP]] = icmp eq i32 [[X]], 0 Edge: [label [[BB2]],label %exit2], RenamedOp: [[X_0]] }
+; CHECK-NEXT:    [[X_0_4:%.*]] = bitcast i32 [[X_0]] to i32
 ; CHECK-NEXT:    br i1 [[CMP_0]], label [[BB3:%.*]], label [[EXIT2:%.*]]
 ; CHECK:       bb3:
-; CHECK: RenamedOp: [[X]]
-; CHECK:         [[X_0_1_2:%.*]] = bitcast i32 [[X_0_1]] to i32
-; CHECK: RenamedOp: [[X_0_1]]
-; CHECK:         [[X_0_1_3:%.*]] = bitcast i32 [[X_0_1]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[X]], 0 Edge: [label [[BB3]],label %exit3], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_0_1_2:%.*]] = bitcast i32 [[X_0_1]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP]] = icmp eq i32 [[X]], 0 Edge: [label [[BB3]],label %exit4], RenamedOp: [[X_0_1]] }
+; CHECK-NEXT:    [[X_0_1_3:%.*]] = bitcast i32 [[X_0_1]] to i32
 ; CHECK-NEXT:    br i1 [[CMP_0_1]], label [[EXIT3:%.*]], label [[EXIT4:%.*]]
 ; CHECK:       exit1:
 ; CHECK-NEXT:    ret i32 0
diff --git a/llvm/test/Transforms/Util/PredicateInfo/condprop.ll b/llvm/test/Transforms/Util/PredicateInfo/condprop.ll
index 256d0d908ec1e..2720e8b2c06e2 100644
--- a/llvm/test/Transforms/Util/PredicateInfo/condprop.ll
+++ b/llvm/test/Transforms/Util/PredicateInfo/condprop.ll
@@ -1,4 +1,4 @@
-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-comments
 ; RUN: opt -passes=print-predicateinfo -disable-output < %s 2>&1 | FileCheck %s
 
 @a = external global i32		; <ptr> [#uses=7]
@@ -98,11 +98,21 @@ define void @test3(i32 %x, i32 %y) {
 ; CHECK-NEXT:    [[XZ:%.*]] = icmp eq i32 [[X:%.*]], 0
 ; CHECK-NEXT:    [[YZ:%.*]] = icmp eq i32 [[Y:%.*]], 0
 ; CHECK-NEXT:    [[Z:%.*]] = and i1 [[XZ]], [[YZ]]
-; CHECK:         [[Z_0:%.*]] = bitcast i1 [[Z]] to i1
-; CHECK:         [[XZ_0:%.*]] = bitcast i1 [[XZ]] to i1
-; CHECK:         [[X_0:%.*]] = bitcast i32 [[X]] to i32
-; CHECK:         [[YZ_0:%.*]] = bitcast i1 [[YZ]] to i1
-; CHECK:         [[Y_0:%.*]] = bitcast i32 [[Y]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[Z]] = and i1 [[XZ]], [[YZ]] Edge: [label [[TMP0:%.*]],label %nope], RenamedOp: [[Z]] }
+; CHECK-NEXT:    [[Z_0:%.*]] = bitcast i1 [[Z]] to i1
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[XZ]] = icmp eq i32 [[X]], 0 Edge: [label [[TMP0]],label %both_zero], RenamedOp: [[XZ]] }
+; CHECK-NEXT:    [[XZ_0:%.*]] = bitcast i1 [[XZ]] to i1
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[XZ]] = icmp eq i32 [[X]], 0 Edge: [label [[TMP0]],label %both_zero], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_0:%.*]] = bitcast i32 [[X]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[YZ]] = icmp eq i32 [[Y]], 0 Edge: [label [[TMP0]],label %both_zero], RenamedOp: [[YZ]] }
+; CHECK-NEXT:    [[YZ_0:%.*]] = bitcast i1 [[YZ]] to i1
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[YZ]] = icmp eq i32 [[Y]], 0 Edge: [label [[TMP0]],label %both_zero], RenamedOp: [[Y]] }
+; CHECK-NEXT:    [[Y_0:%.*]] = bitcast i32 [[Y]] to i32
 ; CHECK-NEXT:    br i1 [[Z]], label [[BOTH_ZERO:%.*]], label [[NOPE:%.*]]
 ; CHECK:       both_zero:
 ; CHECK-NEXT:    call void @foo(i1 [[XZ_0]])
@@ -133,7 +143,9 @@ define void @test4(i1 %b, i32 %x) {
 ; CHECK-LABEL: @test4(
 ; CHECK-NEXT:    br i1 [[B:%.*]], label [[SW:%.*]], label [[CASE3:%.*]]
 ; CHECK:       sw:
-; CHECK:         [[X_0:%.*]] = bitcast i32 [[X:%.*]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; switch predicate info { CaseValue: i32 1 Edge: [label [[SW]],label %case1], RenamedOp: [[X:%.*]] }
+; CHECK-NEXT:    [[X_0:%.*]] = bitcast i32 [[X]] to i32
 ; CHECK-NEXT:    switch i32 [[X]], label [[DEFAULT:%.*]] [
 ; CHECK-NEXT:      i32 0, label [[CASE0:%.*]]
 ; CHECK-NEXT:      i32 1, label [[CASE1:%.*]]
@@ -180,10 +192,18 @@ case3:
 define i1 @test5(i32 %x, i32 %y) {
 ; CHECK-LABEL: @test5(
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp eq i32 [[X:%.*]], [[Y:%.*]]
-; CHECK:         [[X_0:%.*]] = bitcast i32 [[X]] to i32
-; CHECK:         [[X_1:%.*]] = bitcast i32 [[X]] to i32
-; CHECK:         [[Y_0:%.*]] = bitcast i32 [[Y]] to i32
-; CHECK:         [[Y_1:%.*]] = bitcast i32 [[Y]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[X]], [[Y]] Edge: [label [[TMP0:%.*]],label %same], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_0:%.*]] = bitcast i32 [[X]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP]] = icmp eq i32 [[X]], [[Y]] Edge: [label [[TMP0]],label %different], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_1:%.*]] = bitcast i32 [[X]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[X]], [[Y]] Edge: [label [[TMP0]],label %same], RenamedOp: [[Y]] }
+; CHECK-NEXT:    [[Y_0:%.*]] = bitcast i32 [[Y]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP]] = icmp eq i32 [[X]], [[Y]] Edge: [label [[TMP0]],label %different], RenamedOp: [[Y]] }
+; CHECK-NEXT:    [[Y_1:%.*]] = bitcast i32 [[Y]] to i32
 ; CHECK-NEXT:    br i1 [[CMP]], label [[SAME:%.*]], label [[DIFFERENT:%.*]]
 ; CHECK:       same:
 ; CHECK-NEXT:    [[CMP2:%.*]] = icmp ne i32 [[X_0]], [[Y_0]]
@@ -253,10 +273,18 @@ different:
 define i1 @test7(i32 %x, i32 %y) {
 ; CHECK-LABEL: @test7(
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp sgt i32 [[X:%.*]], [[Y:%.*]]
-; CHECK:         [[X_0:%.*]] = bitcast i32 [[X]] to i32
-; CHECK:         [[X_1:%.*]] = bitcast i32 [[X]] to i32
-; CHECK:         [[Y_0:%.*]] = bitcast i32 [[Y]] to i32
-; CHECK:         [[Y_1:%.*]] = bitcast i32 [[Y]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp sgt i32 [[X]], [[Y]] Edge: [label [[TMP0:%.*]],label %same], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_0:%.*]] = bitcast i32 [[X]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP]] = icmp sgt i32 [[X]], [[Y]] Edge: [label [[TMP0]],label %different], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_1:%.*]] = bitcast i32 [[X]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp sgt i32 [[X]], [[Y]] Edge: [label [[TMP0]],label %same], RenamedOp: [[Y]] }
+; CHECK-NEXT:    [[Y_0:%.*]] = bitcast i32 [[Y]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP]] = icmp sgt i32 [[X]], [[Y]] Edge: [label [[TMP0]],label %different], RenamedOp: [[Y]] }
+; CHECK-NEXT:    [[Y_1:%.*]] = bitcast i32 [[Y]] to i32
 ; CHECK-NEXT:    br i1 [[CMP]], label [[SAME:%.*]], label [[DIFFERENT:%.*]]
 ; CHECK:       same:
 ; CHECK-NEXT:    [[CMP2:%.*]] = icmp sle i32 [[X_0]], [[Y_0]]
@@ -280,10 +308,18 @@ different:
 define i1 @test7_fp(float %x, float %y) {
 ; CHECK-LABEL: @test7_fp(
 ; CHECK-NEXT:    [[CMP:%.*]] = fcmp ogt float [[X:%.*]], [[Y:%.*]]
-; CHECK:         [[X_0:%.*]] = bitcast float [[X]] to float
-; CHECK:         [[X_1:%.*]] = bitcast float [[X]] to float
-; CHECK:         [[Y_0:%.*]] = bitcast float [[Y]] to float
-; CHECK:         [[Y_1:%.*]] = bitcast float [[Y]] to float
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = fcmp ogt float [[X]], [[Y]] Edge: [label [[TMP0:%.*]],label %same], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_0:%.*]] = bitcast float [[X]] to float
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP]] = fcmp ogt float [[X]], [[Y]] Edge: [label [[TMP0]],label %different], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_1:%.*]] = bitcast float [[X]] to float
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = fcmp ogt float [[X]], [[Y]] Edge: [label [[TMP0]],label %same], RenamedOp: [[Y]] }
+; CHECK-NEXT:    [[Y_0:%.*]] = bitcast float [[Y]] to float
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP]] = fcmp ogt float [[X]], [[Y]] Edge: [label [[TMP0]],label %different], RenamedOp: [[Y]] }
+; CHECK-NEXT:    [[Y_1:%.*]] = bitcast float [[Y]] to float
 ; CHECK-NEXT:    br i1 [[CMP]], label [[SAME:%.*]], label [[DIFFERENT:%.*]]
 ; CHECK:       same:
 ; CHECK-NEXT:    [[CMP2:%.*]] = fcmp ule float [[X_0]], [[Y_0]]
@@ -353,8 +389,12 @@ different:
 define i32 @test9(i32 %i, i32 %j) {
 ; CHECK-LABEL: @test9(
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp eq i32 [[I:%.*]], [[J:%.*]]
-; CHECK:         [[I_0:%.*]] = bitcast i32 [[I]] to i32
-; CHECK:         [[J_0:%.*]] = bitcast i32 [[J]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[I]], [[J]] Edge: [label [[TMP0:%.*]],label %cond_true], RenamedOp: [[I]] }
+; CHECK-NEXT:    [[I_0:%.*]] = bitcast i32 [[I]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[I]], [[J]] Edge: [label [[TMP0]],label %cond_true], RenamedOp: [[J]] }
+; CHECK-NEXT:    [[J_0:%.*]] = bitcast i32 [[J]] to i32
 ; CHECK-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[RET:%.*]]
 ; CHECK:       cond_true:
 ; CHECK-NEXT:    [[DIFF:%.*]] = sub i32 [[I_0]], [[J_0]]
@@ -376,8 +416,12 @@ ret:
 define i32 @test10(i32 %j, i32 %i) {
 ; CHECK-LABEL: @test10(
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp eq i32 [[I:%.*]], [[J:%.*]]
-; CHECK:         [[I_0:%.*]] = bitcast i32 [[I]] to i32
-; CHECK:         [[J_0:%.*]] = bitcast i32 [[J]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[I]], [[J]] Edge: [label [[TMP0:%.*]],label %cond_true], RenamedOp: [[I]] }
+; CHECK-NEXT:    [[I_0:%.*]] = bitcast i32 [[I]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[I]], [[J]] Edge: [label [[TMP0]],label %cond_true], RenamedOp: [[J]] }
+; CHECK-NEXT:    [[J_0:%.*]] = bitcast i32 [[J]] to i32
 ; CHECK-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[RET:%.*]]
 ; CHECK:       cond_true:
 ; CHECK-NEXT:    [[DIFF:%.*]] = sub i32 [[I_0]], [[J_0]]
@@ -403,14 +447,20 @@ define i32 @test11(i32 %x) {
 ; CHECK-NEXT:    [[V0:%.*]] = call i32 @yogibar()
 ; CHECK-NEXT:    [[V1:%.*]] = call i32 @yogibar()
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp eq i32 [[V0]], [[V1]]
-; CHECK:         [[V0_0:%.*]] = bitcast i32 [[V0]] to i32
-; CHECK:         [[V1_0:%.*]] = bitcast i32 [[V1]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP]] = icmp eq i32 [[V0]], [[V1]] Edge: [label [[TMP0:%.*]],label %next], RenamedOp: [[V0]] }
+; CHECK-NEXT:    [[V0_0:%.*]] = bitcast i32 [[V0]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[V0]], [[V1]] Edge: [label [[TMP0]],label %cond_true], RenamedOp: [[V1]] }
+; CHECK-NEXT:    [[V1_0:%.*]] = bitcast i32 [[V1]] to i32
 ; CHECK-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[NEXT:%.*]]
 ; CHECK:       cond_true:
 ; CHECK-NEXT:    ret i32 [[V1_0]]
 ; CHECK:       next:
 ; CHECK-NEXT:    [[CMP2:%.*]] = icmp eq i32 [[X:%.*]], [[V0_0]]
-; CHECK:         [[V0_0_1:%.*]] = bitcast i32 [[V0_0]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP2]] = icmp eq i32 [[X]], [[V0_0]] Edge: [label [[NEXT]],label %cond_true2], RenamedOp: [[V0_0]] }
+; CHECK-NEXT:    [[V0_0_1:%.*]] = bitcast i32 [[V0_0]] to i32
 ; CHECK-NEXT:    br i1 [[CMP2]], label [[COND_TRUE2:%.*]], label [[NEXT2:%.*]]
 ; CHECK:       cond_true2:
 ; CHECK-NEXT:    ret i32 [[V0_0_1]]
@@ -439,8 +489,12 @@ next2:
 define i32 @test12(i32 %x) {
 ; CHECK-LABEL: @test12(
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp eq i32 [[X:%.*]], 0
-; CHECK:         [[X_0:%.*]] = bitcast i32 [[X]] to i32
-; CHECK:         [[X_1:%.*]] = bitcast i32 [[X]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[X]], 0 Edge: [label [[TMP0:%.*]],label %cond_true], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_0:%.*]] = bitcast i32 [[X]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP]] = icmp eq i32 [[X]], 0 Edge: [label [[TMP0]],label %cond_false], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_1:%.*]] = bitcast i32 [[X]] to i32
 ; CHECK-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
 ; CHECK:       cond_true:
 ; CHECK-NEXT:    br label [[RET:%.*]]
diff --git a/llvm/test/Transforms/Util/PredicateInfo/diamond.ll b/llvm/test/Transforms/Util/PredicateInfo/diamond.ll
index ac2c9a1026e76..6900a2a710c72 100644
--- a/llvm/test/Transforms/Util/PredicateInfo/diamond.ll
+++ b/llvm/test/Transforms/Util/PredicateInfo/diamond.ll
@@ -1,16 +1,20 @@
-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt -passes=print-predicateinfo < %s 2>&1 | FileCheck %s
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-comments
+; RUN: opt -passes=print-predicateinfo -disable-output < %s 2>&1 | FileCheck %s
 define i1 @f(i32 %x, i1 %y) {
 ; CHECK-LABEL: @f(
 ; CHECK-NEXT:    br i1 [[Y:%.*]], label [[BB0:%.*]], label [[BB1:%.*]]
 ; CHECK:       bb0:
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp sge i32 [[X:%.*]], 0
-; CHECK:         [[X_0:%.*]] = bitcast i32 [[X]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp sge i32 [[X]], 0 Edge: [label [[BB0]],label %bb2], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_0:%.*]] = bitcast i32 [[X]] to i32
 ; CHECK-NEXT:    br i1 [[CMP]], label [[BB2:%.*]], label [[BB3:%.*]]
 ; CHECK:       bb1:
 ; CHECK-NEXT:    [[X2:%.*]] = add nuw nsw i32 [[X]], 1
 ; CHECK-NEXT:    [[CMP2:%.*]] = icmp sge i32 [[X2]], 2
-; CHECK:         [[X2_0:%.*]] = bitcast i32 [[X2]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP2]] = icmp sge i32 [[X2]], 2 Edge: [label [[BB1]],label %bb2], RenamedOp: [[X2]] }
+; CHECK-NEXT:    [[X2_0:%.*]] = bitcast i32 [[X2]] to i32
 ; CHECK-NEXT:    br i1 [[CMP2]], label [[BB2]], label [[BB3]]
 ; CHECK:       bb2:
 ; CHECK-NEXT:    [[X3:%.*]] = phi i32 [ [[X_0]], [[BB0]] ], [ [[X2_0]], [[BB1]] ]
@@ -38,12 +42,16 @@ define i1 @g(i32 %x, i1 %y) {
 ; CHECK-NEXT:    br i1 [[Y:%.*]], label [[BB0:%.*]], label [[BB1:%.*]]
 ; CHECK:       bb0:
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp sge i32 [[X:%.*]], 0
-; CHECK:         [[X_0:%.*]] = bitcast i32 [[X]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP]] = icmp sge i32 [[X]], 0 Edge: [label [[BB0]],label %bb2], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_0:%.*]] = bitcast i32 [[X]] to i32
 ; CHECK-NEXT:    br i1 [[CMP]], label [[BB3:%.*]], label [[BB2:%.*]]
 ; CHECK:       bb1:
 ; CHECK-NEXT:    [[X2:%.*]] = add nuw nsw i32 [[X]], 1
 ; CHECK-NEXT:    [[CMP2:%.*]] = icmp sge i32 [[X2]], 2
-; CHECK:         [[X2_0:%.*]] = bitcast i32 [[X2]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP2]] = icmp sge i32 [[X2]], 2 Edge: [label [[BB1]],label %bb2], RenamedOp: [[X2]] }
+; CHECK-NEXT:    [[X2_0:%.*]] = bitcast i32 [[X2]] to i32
 ; CHECK-NEXT:    br i1 [[CMP2]], label [[BB3]], label [[BB2]]
 ; CHECK:       bb2:
 ; CHECK-NEXT:    [[X3:%.*]] = phi i32 [ [[X_0]], [[BB0]] ], [ [[X2_0]], [[BB1]] ]
diff --git a/llvm/test/Transforms/Util/PredicateInfo/edge.ll b/llvm/test/Transforms/Util/PredicateInfo/edge.ll
index ef757f323921a..86acca2245e3e 100644
--- a/llvm/test/Transforms/Util/PredicateInfo/edge.ll
+++ b/llvm/test/Transforms/Util/PredicateInfo/edge.ll
@@ -1,16 +1,18 @@
-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt -passes=print-predicateinfo < %s 2>&1 | FileCheck %s
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-comments
+; RUN: opt -passes=print-predicateinfo -disable-output < %s 2>&1 | FileCheck %s
 
 define i32 @f1(i32 %x) {
 ; CHECK-LABEL: @f1(
 ; CHECK-NEXT:  bb0:
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp eq i32 [[X:%.*]], 0
-; CHECK:         [[X_0:%.*]] = bitcast i32 [[X]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 1 Comparison: [[CMP]] = icmp eq i32 [[X]], 0 Edge: [label [[BB0:%.*]],label %bb2], RenamedOp: [[X]] }
+; CHECK-NEXT:    [[X_0:%.*]] = bitcast i32 [[X]] to i32
 ; CHECK-NEXT:    br i1 [[CMP]], label [[BB2:%.*]], label [[BB1:%.*]]
 ; CHECK:       bb1:
 ; CHECK-NEXT:    br label [[BB2]]
 ; CHECK:       bb2:
-; CHECK-NEXT:    [[COND:%.*]] = phi i32 [ [[X_0]], [[BB0:%.*]] ], [ 0, [[BB1]] ]
+; CHECK-NEXT:    [[COND:%.*]] = phi i32 [ [[X_0]], [[BB0]] ], [ 0, [[BB1]] ]
 ; CHECK-NEXT:    [[FOO:%.*]] = add i32 [[COND]], [[X]]
 ; CHECK-NEXT:    ret i32 [[FOO]]
 ;
@@ -29,12 +31,14 @@ define i32 @f2(i32 %x) {
 ; CHECK-LABEL: @f2(
 ; CHECK-NEXT:  bb0:
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp ne i32 [[X:%.*]], 0
-; CHECK:         [[X_0:%.*]] = bitcast i32 [[X]] to i32
+; CHECK-NEXT:  ; Has predicate info
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[CMP]] = icmp ne i32 [[X]], 0 Edge: [label...
[truncated]

arichardson · 2025-10-28T16:15:37Z

llvm/test/Transforms/Util/PredicateInfo/pr33457.ll

 ; CHECK-NEXT:    store i32 6, ptr @e, align 4
 ; CHECK-NEXT:    br label [[TMP1:%.*]]
-; CHECK:         [[TMP2:%.*]] = load i32, ptr @d, align 4
+; CHECK:       1:


These should probably be variable captures. Any idea why it's not working?

Maybe the script regex for labels expected a leading alphabetical character?

Blocks are only variables in newer UTC versions.

Actually, it works. The change here is adding a check for block labels.
From

; CHECK: [[TMP2:%.*]] = load i32, ptr @d, align 4

To

; CHECK: 1: ; CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr @d, align 4

Maybe the script regex for labels expected a leading alphabetical character?

In fact, this change was not introduced by this PR, and I don't know why the label check here is CHECK: 1:.

arichardson

Should this be the default behaviour (with a --version bump)?

Being able to use CHECK-NEXT more instead of the more fragile CHECK is IMO a good thing.

arichardson · 2025-10-28T16:16:04Z

llvm/test/Transforms/Util/PredicateInfo/pr33457.ll

 ; CHECK-NEXT:    store i32 6, ptr @e, align 4
 ; CHECK-NEXT:    br label [[TMP1:%.*]]
-; CHECK:         [[TMP2:%.*]] = load i32, ptr @d, align 4
+; CHECK:       1:


Maybe the script regex for labels expected a leading alphabetical character?

llvm/test/Transforms/Util/PredicateInfo/branch-on-same-cond.ll

Camsyn · 2025-10-28T16:58:51Z

Hi, @arichardson

Should this be the default behaviour (with a --version bump)?

This change is the default behaviour, but only enabled when an extra UTC option --check-comments specified.
And I don't think it's necessary to control it with versions.

Being able to use CHECK-NEXT more instead of the more fragile CHECK is IMO a good thing.

This change is not introduced by this PR. And I think using CHECK instead of CHECK-NEXT is somehow reasonable, since there is an empty line between the two lines required checking:

; CHECK-NEXT:    br label [[TMP1:%.*]]
; CHECK:         [[TMP2:%.*]] = load i32, ptr @d, align 4
....
  br label %1

; <label>:1:                                      ; preds = %17, %0
  %2 = load i32, ptr @d, align 4

arichardson

We also have a --check-attributes option to check another subset of comments. Could you add a test showing whether the new --check-comments flag also checks the function attributes comment?

arichardson · 2025-10-28T17:06:57Z

llvm/utils/UpdateTestChecks/common.py

+            scrubbed_line = (
+                # Ignore any comments, since the check lines will too.
+                SCRUB_IR_COMMENT_RE.sub(r"", line)
+                if ignore_all_comments
+                # Ignore comments of predecessors only.
+                else SCRUB_PRED_COMMENT_RE.sub(r"", line)
+            )


Suggested change

scrubbed_line = (

# Ignore any comments, since the check lines will too.

SCRUB_IR_COMMENT_RE.sub(r"", line)

if ignore_all_comments

# Ignore comments of predecessors only.

else SCRUB_PRED_COMMENT_RE.sub(r"", line)

)

if ignore_all_comments:

# Ignore any comments, since the check lines will too.

scrubbed_line = SCRUB_IR_COMMENT_RE.sub(r"", line)

else:

# Ignore comments of predecessors only.

scrubbed_line = SCRUB_PRED_COMMENT_RE.sub(r"", line)

Same number of lines and IMO slightly easier to read than the ternary.

arichardson · 2025-10-28T17:10:05Z

Hi, @arichardson

Should this be the default behaviour (with a --version bump)?

This change is the default behaviour, but only enabled when an extra UTC option --check-comments specified. And I don't think it's necessary to control it with versions.

Being able to use CHECK-NEXT more instead of the more fragile CHECK is IMO a good thing.

This change is not introduced by this PR. And I think using CHECK instead of CHECK-NEXT is somehow reasonable, since there is an empty line between the two lines required checking:
; CHECK-NEXT:    br label [[TMP1:%.*]]
; CHECK:         [[TMP2:%.*]] = load i32, ptr @d, align 4
....
  br label %1

; <label>:1:                                      ; preds = %17, %0
  %2 = load i32, ptr @d, align 4

We could add a CHECK-EMPTY: and a CHECK-NEXT for the label comment, but that should probably be a separate change.

Camsyn · 2025-10-28T17:22:25Z

We also have a --check-attributes option to check another subset of comments. Could you add a test showing whether the new --check-comments flag also checks the function attributes comment?

It does not.
--check-comments only checks IR comments, orthogonal to --check-attributes, which checks the function comments.

Maybe we should rename this option to --check-ir-comments?

Camsyn · 2025-10-28T17:24:49Z

We could add a CHECK-EMPTY: and a CHECK-NEXT for the label comment, but that should probably be a separate change.

I will take a look.

github-actions · 2025-10-28T17:25:51Z

✅ With the latest revision this PR passed the Python code formatter.

arichardson · 2025-10-28T17:58:53Z

We also have a --check-attributes option to check another subset of comments. Could you add a test showing whether the new --check-comments flag also checks the function attributes comment?

It does not. --check-comments only checks IR comments, orthogonal to --check-attributes, which checks the function comments.

Maybe we should rename this option to --check-ir-comments?

Technically I guess the whole file is IR but I can't think of a better name that isn't super long (--check-comments-in-function). So this sounds good to me.

Camsyn · 2025-10-28T18:17:22Z

Technically I guess the whole file is IR but I can't think of a better name that isn't super long (--check-comments-in-function). So this sounds good to me.

My mistake. It should be comments attached to IR instructions.

I'm wondering if we need to check all comments through this option, i.e., global/function/BB/instruction comments (print-predicate-info/print<memssa> insert instruction comments; but I don't know if there is any other pass inserting other comments required checking...).
If not, maybe --check-inst-comments is a good name, and I will validate that it only works on IR instructions.

nikic

LGTM after a rebase.

arichardson · 2025-10-29T11:00:55Z

llvm/test/Transforms/Util/PredicateInfo/branch-on-same-cond.ll

@@ -1,4 +1,4 @@
-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-comments


Suggested change

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-comments

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-inst-comments

And the same below

nikic · 2025-10-29T11:03:28Z

llvm/utils/update_test_checks.py

        help="Check global entries (global variables, metadata, attribute sets, ...) for functions",
    )
+    parser.add_argument(
+        "--check-inst_comments",


Suggested change

"--check-inst_comments",

"--check-inst-comments",

arichardson

LGTM with updated tests

See #165419 (comment) for details. The extra annotation `"; Has predicate info"` does not provide any extra information and might poison the UTC-generated checks introduced by #165419.

…165434) See llvm/llvm-project#165419 (comment) for details. The extra annotation `"; Has predicate info"` does not provide any extra information and might poison the UTC-generated checks introduced by #165419.

nikic · 2025-10-29T15:18:30Z

llvm/test/Transforms/Util/PredicateInfo/condprop.ll

-; CHECK:         [[X_0:%.*]] = bitcast i32 [[X]] to i32
-; CHECK:         [[YZ_0:%.*]] = bitcast i1 [[YZ]] to i1
-; CHECK:         [[Y_0:%.*]] = bitcast i32 [[Y]] to i32
+; CHECK-NEXT:  ; branch predicate info { TrueEdge: 0 Comparison: [[Z]] = and i1 [[XZ]], [[YZ]] Edge: [label [[TMP0:%.*]],label %nope], RenamedOp: [[Z]] }


Here one of the labels gets a wildcard and the other doesn't. I wonder whether printing a space after the comma would give use consistent behavior?

I wonder whether printing a space after the comma would give use consistent behavior?

@nikic, Unfortunately, it does not work :(.

It is another reason to cause such inconsistency.

Maybe the wildcard is used to define [[TMP0]] when it first appears, and the following [[TMP0]]]s just reference the first definition, just like mermaid/dot grammar

I think these tests default to an old --version which doesn't use wildcards for labels. Could you try adding --version 6 to the start of the test?

Maybe the wildcard is used to define [[TMP0]] when it first appears, and the following [[TMP0]]]s just reference the first definition, just like mermaid/dot grammar

I may have misunderstood earlier. If @nikic thinks that the label %nope in ... Edge: [label [[TMP0 :%.*]], label % nope],... is mistakenly missing a wildcard, I think I have found the cause, not a version issue or "a missed space after the comma".

It's the mismatched variable regex r"(\s*)(?:%([\w$.-]+?))([,\s\}]|\Z)" that causes such an issue:

TWO approaches to fix:

Change PredicateInfo's printing

Change the regex (maybe the regex is wrong? Why does it expect some character after a variable?)

Full regex:

(\s*)((?:%([\w$.-]+?))|(?:#([0-9]+))|^(?:attributes #([0-9]+))|(?:@([0-9]+))|^(?:@([0-9]+))|^(?:@([a-zA-Z0-9_$\"\\.-]*[a-zA-Z_$\"\\.-][a-zA-Z0-9_$\"\\.-]*))|(?:!dbg (![0-9]+))|(?:!DIAssignID (![0-9]+))|(?:!prof (![0-9]+))|(?:!tbaa (![0-9]+))|(?:!tbaa.struct (![0-9]+))|(?:!range (![0-9]+))|(?:!llvm.loop (![0-9]+))|^(?:(![0-9]+))|(?:!llvm.access.group (![0-9]+))|(?:![a-z.]+ (![0-9]+))|(?:[, (](![0-9]+)))([,\s\}]|\Z)

I'm not sure why that specific character requirement exists, but it would probably be fine to add \] to the list?

597ac47 introduced such var regex suffix to help locate a var.

I think it should be safe to add ']' as another suffix, since phi i32 [%a, %then], [%b, %else] without space is also a legal LLVM IR.

@nikic, do you think the suffix change should be gated by a version? This suffix change might lead to a negligible impact on the current test, as such a pattern of ... %var] is rare.

I don't think we need a version for this.

See llvm#165419 (comment) for details. The extra annotation `"; Has predicate info"` does not provide any extra information and might poison the UTC-generated checks introduced by llvm#165419.

arichardson · 2025-10-30T18:37:05Z

llvm/test/Transforms/Util/PredicateInfo/pr33457.ll

@@ -1,5 +1,5 @@
-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt -passes=print-predicateinfo < %s 2>&1 | FileCheck %s
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-inst-comments


Suggested change

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-inst-comments

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-inst-comments --version 6

Would that add variable captures for the labels?

No, it wouldn't.
It's the mismatched variable regex r"(\s*)(?:%([\w$.-]+?))([,\s\}]|\Z)" that causes the missed capture of the labels in ... , %label].
See this exmpale:

* main: [SPIRV] Fix vector bitcast check in LegalizePointerCast (llvm#164997) [lldb][docs] Add troubleshooting section to scripting introduction [Sema] Fix parameter index checks on explicit object member functions (llvm#165586) To fix polymorphic pointer assignment in FORALL when LHS is unlimited polymorphic and RHS is intrinsic type target (llvm#164999) [CostModel][AArch64] Model cost of extract.last.active intrinsic (clastb) (llvm#165739) [MemProf] Select largest of matching contexts from profile (llvm#165338) [lldb][TypeSystem] Better support for _BitInt types (llvm#165689) [NVPTX] Move TMA G2S lowering to Tablegen (llvm#165710) [MLIR][NVVM] Extend NVVM mma ops to support fp64 (llvm#165380) [UTC] Support to test annotated IR (llvm#165419)

Previously, any blank lines in IR were ignored by UTC, leading to more fragile `CHECK`s being generated. This change lets UTC, 1) emit `CHECK-EMPTY` to check blank lines, and 2) generate more `CHECK-NEXT`s, landing the discussion #165419 (comment). Moreover, this change also aligns the behavior of IR check-gen to ASM check-gen, which has been emitting `CHECK-EMPTY` since a8a89c7.

Previously, any blank lines in IR were ignored by UTC, leading to more fragile `CHECK`s being generated. This change lets UTC, 1) emit `CHECK-EMPTY` to check blank lines, and 2) generate more `CHECK-NEXT`s, landing the discussion llvm/llvm-project#165419 (comment). Moreover, this change also aligns the behavior of IR check-gen to ASM check-gen, which has been emitting `CHECK-EMPTY` since llvm/llvm-project@a8a89c7.

See llvm#165419 (comment) for details. The extra annotation `"; Has predicate info"` does not provide any extra information and might poison the UTC-generated checks introduced by llvm#165419.

Some analysis/transformation, e.g., predicate info/ mem ssa, insert instruction annotations as comments, referring to llvm#165249 (comment). This PR makes UTC support checking these instruction annotations with an extra UTC option `-check-inst-comments`. E.g., Before: ```LLVM ; CHECK: [[Z_0:%.*]] = bitcast i1 [[Z]] to i1 ``` After: ```LLVM ; CHECK-NEXT: ; branch predicate info { TrueEdge: 0 Comparison: [[Z]] = and i1 [[XZ]], [[YZ]] Edge: [label [[TMP0:%.*]],label %nope], RenamedOp: [[Z]] } ; CHECK-NEXT: [[Z_0:%.*]] = bitcast i1 [[Z]] to i1 ``` This PR also regenerates all UTC-generated tests for PredicateInfo; No MemSSA test is updated, as there are no UTC-generated tests designated for `print<memoryssa>`.

Previously, any blank lines in IR were ignored by UTC, leading to more fragile `CHECK`s being generated. This change lets UTC, 1) emit `CHECK-EMPTY` to check blank lines, and 2) generate more `CHECK-NEXT`s, landing the discussion llvm#165419 (comment). Moreover, this change also aligns the behavior of IR check-gen to ASM check-gen, which has been emitting `CHECK-EMPTY` since llvm@a8a89c7.

Camsyn requested review from MatzeB, dtcxzyw and nikic October 28, 2025 15:35

llvmbot added testing-tools llvm:transforms labels Oct 28, 2025

arichardson reviewed Oct 28, 2025

View reviewed changes

nikic reviewed Oct 28, 2025

View reviewed changes

llvm/test/Transforms/Util/PredicateInfo/branch-on-same-cond.ll Outdated Show resolved Hide resolved

Camsyn mentioned this pull request Oct 28, 2025

[PredicateInfo] Drop redundant PredicateInfo annotation #165434

Merged

arichardson reviewed Oct 28, 2025

View reviewed changes

nikic approved these changes Oct 29, 2025

View reviewed changes

arichardson reviewed Oct 29, 2025

View reviewed changes

nikic reviewed Oct 29, 2025

View reviewed changes

arichardson approved these changes Oct 29, 2025

View reviewed changes

Camsyn added 6 commits October 29, 2025 22:07

[UTC] Support to check generated comments

aec8313

Regenerate the related tests

1626db8

Refactor to enhance readability.

fa39e19

Make formatter happy

4631c34

Rename option to check-inst-comments

b241a99

Address review comments

d890f94

Regenerate tests after rebase

d9ebb7c

Camsyn force-pushed the UTC-on-comments branch from 5869c7c to d9ebb7c Compare October 29, 2025 14:46

nikic reviewed Oct 29, 2025

View reviewed changes

Camsyn added 2 commits October 30, 2025 19:44

Add '\]' to the var regex suffix

0f57f27

Regenerated the tests after adding '\]' as suffix

e293632

Camsyn requested review from arichardson and nikic October 30, 2025 12:13

Camsyn mentioned this pull request Oct 30, 2025

[UTC] CHECK-EMPTY instead of skipping blank lines #165718

Merged

arichardson approved these changes Oct 30, 2025

View reviewed changes

Camsyn merged commit 511c9c0 into llvm:main Oct 31, 2025
10 checks passed

		@@ -1,4 +1,4 @@
		; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
		; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-comments

[UTC] Support to test annotated IR #165419

[UTC] Support to test annotated IR #165419

Uh oh!

Conversation

Camsyn commented Oct 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

llvmbot commented Oct 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Camsyn Oct 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

arichardson left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Camsyn commented Oct 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

arichardson left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

arichardson commented Oct 28, 2025

Uh oh!

Camsyn commented Oct 28, 2025

Uh oh!

Camsyn commented Oct 28, 2025

Uh oh!

github-actions bot commented Oct 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

arichardson commented Oct 28, 2025

Uh oh!

Camsyn commented Oct 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

nikic left a comment

Choose a reason for hiding this comment

Uh oh!

arichardson Oct 29, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

arichardson left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Camsyn Oct 29, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Camsyn Oct 30, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Camsyn commented Oct 28, 2025 •

edited

Loading

llvmbot commented Oct 28, 2025 •

edited

Loading

Camsyn Oct 28, 2025 •

edited

Loading

Camsyn commented Oct 28, 2025 •

edited

Loading

github-actions bot commented Oct 28, 2025 •

edited

Loading

Camsyn commented Oct 28, 2025 •

edited

Loading

arichardson Oct 29, 2025 •

edited

Loading

Camsyn Oct 29, 2025 •

edited

Loading

Camsyn Oct 30, 2025 •

edited

Loading