On 22 Mar 2015, at 23:11, Garrett Cooper <yaneurabeya_at_gmail.com> wrote: > > On Mar 22, 2015, at 15:09, Dimitry Andric <dim_at_FreeBSD.org> wrote: > >> On 22 Mar 2015, at 23:04, Garrett Cooper <yaneurabeya_at_gmail.com> wrote: >>> >>> On Mar 22, 2015, at 15:01, Craig Rodrigues <rodrigc_at_FreeBSD.org> wrote: >>> >>> ... >>> >>>> OK, converting expr.y to use unsigned integers would require a bit of work. >>>> >>>> Can you commit your patch to the Makefile? It fixes the problem for now. >>> >>> +1 >>> >>> I’d still like to know why clang 3.5 doesn’t have this behavior though — there might be other potential issues lurking around that need to be solved (either here, in ports, or both). >> >> Because this version optimizes better around undefined behavior. There >> are most likely many issues lurking around, and most certainly in ports. >> >> I would recommend using UBSan to tackle this kind of thing. > > I hope this got a ports tinderbox run first… It is called an "exp-run" in ports land, but that only really tests whether ports *build* successfully. For most ports, there is no good way of testing them, certainly not if they don't have their own testing facilities. That said, I have built quite a number of ports with 3.6.0, even before the import, and saw zero problems until now. Then again, I'm just one user, and that is a very low sample size. ;) I don't expect a great many problems with integer wrapping though. You have to realize that quite a lot of open source projects already went through this phase, when gcc started optimizing integer wrapping a few years ago. We are just slowly catching up. > Adding UBSan to tinderbox runs for toolchain upgrades [in the future] might be a good idea. Maybe, but please note that both ASan and UBSan are still beta-ish under FreeBSD, there will most likely be bumps along the road. Any bug reports in that area are welcome. -Dimitry
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:56 UTC