This is finally the last patch for `sink-addsub-of-const.ll`! :)

Again only vectors affected. Frustrating. Let me take a look into that..

# Details

Details

- Reviewers
RKSimon craig.topper spatel - Commits
- rL362145: [DAGCombine] (x - C) - y -> (x - y) - C fold. Try 3

rG76696654322b: [DAGCombine] (x - C) - y -> (x - y) - C fold

rG57aa36ff9164: [DAGCombine] (x - C) - y -> (x - y) - C fold. Try 3

rGdfc34f0211b7: [DAGCombine] (x - C) - y -> (x - y) - C fold. Try 2

rL361856: [DAGCombine] (x - C) - y -> (x - y) - C fold

rL361874: [DAGCombine] (x - C) - y -> (x - y) - C fold. Try 2

# Diff Detail

Diff Detail

- Repository
- rL LLVM

### Event Timeline

Comment Actions

It seems i was overly optimistic that this was the last missing fold of this kind (not counting stuck D62266).

More to follow..