[llvm] [SLP] Don't recognize rotated widened strided stores in analyzeRtStrideCandidate() (PR #204013)

[Mikhail Gudim via llvm-commits]

================
@@ -0,0 +1,477 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+
+; RUN: opt -mtriple=riscv64 -mattr=+m,+v,+unaligned-vector-mem -passes=slp-vectorizer -S < %s | FileCheck %s
+
+define void @constant_stride_widen_rotated1(ptr %pl, i64 %stride, ptr %ps) {
----------------
mgudim wrote:

maybe add pre-commit a test first in a separate pr?

https://github.com/llvm/llvm-project/pull/204013