On 15 June 2010 01:12, Bill Hart <goodwillh...@googlemail.com> wrote: > Just a note: the standard gcc with Debian Lenny is gcc 4.3.2. This is > so buggy it fails to compile MPIR correctly. And yes, we've checked > that it is a gcc bug and it has been reported. > > It is fixed in a later gcc. But that means all Debian Lenny users will > not be able to get correct results until they upgrade their gcc. >4.3.2 > I think this is unprecedented, in that whilst standard system > libraries distributed with a distro have had bugs before, this is the > first gcc used as standard in a distro whose bugs are severe. It's a > general rule that x.y.0 is buggy, but it is unusual for an x.y.2 to be > so bad. > > Bill.
If need be, we could stop people using gcc 4.3.2 quite easily, just as we do with gcc's older than 4.0.1. Again, we could let them set an environment variable if they insist on doing so. Does this produce a broken version of MPIR or does it stop MPIR from building? If it does compile, perhaps you can give us an example of what generates an error. If that could be used as a doc test in Sage, we would have a test for the problem. Dave -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
