steakhal wrote: Ah, I think rushed ahead of myself. I applied the patch to clang-17, where it of course we didn't have any issues even with this broken `isTainted`. Now that I applied the patch to our clang-18 based branch, the file analyzes in 1:40, which is still far off from the baseline run 32 seconds. But 1:40 is arguably a lot better than 1.23 hours.
I perfed again, and it shows this:  This is much better, and showcases the next slowdown bug I'm hunting for the days since I reported this one :D The remaining 1 extra minute is lost during Z3 refutation. So, yea, this PR fixes the bug I reported in #89045, so we can close that once this is merged. And stay tuned for the next slowdown bug. I can already tell you that the bitwise shift checker introduces new constraints, after which we can now refute more bugs! - but also spends more time for finding a valid bug report in an EQClass , while previously we just picked the first one as we found those path constraints `sat`. But I'll come back to that once I have something concrete to share. https://github.com/llvm/llvm-project/pull/89606 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
