vitalybuka wrote: > > It happens later, in LLVM backend, it needs to be fixed. > > From [#65972 > (comment)](https://github.com/llvm/llvm-project/pull/65972#issuecomment-1971855638) > > Is this something you have planned to fix? If not would replacing the .size() > counter with perhaps a seeded random uint8 be acceptable? > > My use case prevents me from shipping the minimal runtime alongside the > instrumentation so my goal was to create an improvement (definitely > imperfect!) to the debugability of a production deployment of BoundsSan. This > PR as is would revert that behavior entirely.
I don't plan to do anything about it. My point that it does not work even on trivial example as in description. Unless you/someone else is willing to work on real fix, this behavior is not worse of preserving. https://github.com/llvm/llvm-project/pull/83470 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
