On Sun, 2026-07-12 at 01:28 +0800, Jin Ma wrote: > Hi Jeff, > > Thank you for the review. > > > The riscv.md part is fine. The tests aren't. > > > > The problem with tests of this style is you're going to be incredibly > > sensitive to register allocation because you're scanning for specific > > registers in the output. > > You are right. The hardcoded a0 makes these tests fragile. > > > Or you could just avoid the test in this case. I'm a big proponent of > > aggressively adding tests to the testsuite, but in this case I could > > make the argument that the cost/benefit analysis on a test for this > > problem isn't great and that our time is better spent elsewhere. > > Agreed. The existing tests (zero-extend-rshift*.c, pr111501.c) > already cover the functional correctness of the splitter, so a > dedicated register-allocation test adds little value here. > > I will drop the two test files and send v2 with only the riscv.md > change.
Just FYI you may use check-function-bodies with a back reference for reg name in such a test case, if you really need one in the future. -- Xi Ruoyao <[email protected]>
