On February 4, 2014 8:09:12 AM GMT+08:00, Austin English <austinengl...@gmail.com> wrote: > FWIW, I see the same failures on Gentoo amd64, at > 4e5f15c6851c69c4cc5da18209218eb918dbee77 (also tested against > 2eb844f8b5ea0ae159eb3fcec78aa50bd8c03f11, but the problem is there as > well
[...] > -ulltof: 18446744073709551616.000000 > +ulltof: 2334.000000 [...] > -ulltof: 18446744073709549568.000000 > +ulltof: 2334.000000 [...] > -ulltof: 18446744073709549282.000000 > +ulltof: 2334.000000 This is a test I just introduce when I fixed long long to float conversion for arm. It means the bug was already there for x86 and became exposed by this new test. I might give it a shot after I fix the problem of parameter passing of bit field (I only tried on arm but it seems to be present on all architectures). However my main machine is an arm machine without much space for a vm or chroot so it is more difficult to fix this x86 bug for me. Cheers, Thomas _______________________________________________ Tinycc-devel mailing list Tinycc-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/tinycc-devel