On Tue, 2 Jun 2026, Jin Ma wrote: > No test case is added: without the return the first gather is a dead > store to the target that the generic path immediately overwrites, so it > is removed by later RTL dead-store elimination and the final assembly is > identical with and without the fix. No test can therefore observe the > change.
Not even with the RTL dead-store elimination pass explicitly disabled? Also I guess an appropriate `-fdump-...' option would let one see it too. Not that I'd insist on an arcane test case for every single obvious bug, but the claim in the commit description seems suspicious to me in this case. Maciej
