Hi Dana On Thu, May 23, 2013 at 10:57:35PM -0700, Dana Jacobsen wrote: > I just found this report of build failures on the Debian mailing list -- I > had not seen these before. I managed to reproduce it on a Power7 machine, > and found the fix. > > I see that the issue came up when I switched the ranged moebius return > value from an IV* (long *) to char*. On every other machine, char is > being treated as signed. On these platforms it is unsigned (which they > technically may do). I switched to explicit signed char and that seems to > have fixed it. I'll get a release out next week.
I know you just said above there will be a release fixing this in the coming week. To keep the diff a bit smaller at least I have uploaded the 0.28 package to unstable (including the patch), so can be used as additional step to verify it builds now again[1]. [1] https://buildd.debian.org/status/package.php?p=libmath-prime-util-perl Regards, Salvatore -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
