I think it might be some kind of bug in gcc-4.4.x, relating to amd64 weirdness; the pari people have had such gcc-4 issues before. What's pre-release is the Ubuntu 9.10 OS, and pre-release OSs sometimes in fact use lookahead prerelease packages, but not always. The actual gcc-4.4.1 they send out should be the same one everybody else gets. Of course, you don't just have to worry about the compiler what with all of the various utils and libs and stuff that get used, and the whole assortment may in fact be 'pre-release'.
I don't have an earlier Ubuntu 9.04 to check out their gcc-4.4.1. And you're right about the pari mailing list: I _think_ I jumped through all the hoops to get my email sent, but then I never actually received it, so who knows how it all works. In any case, I'm updating 9.10 daily, and every day there are some libc6 / gcc / g++ updates, so I'll keep trying the pari build until they work. On Aug 6, 4:45 pm, William Stein <[email protected]> wrote: > 2009/8/6 J Elaych <[email protected]> > > > > > Hi all, > > > I've been working with the Sage / Python package which > > includes pari-2.3.4 and noticed that the build failed at > > the command > > > /usr/bin/gcc -c -O3 -Wall -fno-strict-aliasing -fomit-frame-pointer -I. > > -I../src/headers -fPIC -o base3.o ../src/basemath/base3.c > > > when using the latest gcc-4.4. When I switched back to > > gcc-4.3 pari built fine. Unfortunately because of the way > > the Sage/Python packages are released, we need to be > > able to use the latest gcc's. I'm using the Ubuntu prerelease > > of 9.10. > > > I have found that the problem is completely reproducible > > from the pari-2.3.4.tar.gz and pari-2.4.2.alpha.tar.gz > > tarballs. With 'make -j2 all' the compile extends to > > init.c, where presumably the second thread halts. Using > > 'top' shows cc1 taking up 99% cpu and most of the > > memory, for hours until its interrupted. > > > I didn't find any gcc-4.4 related issues when I searched the > > mailing list archives, so I want to bring this issue to your > > attention if you don't already know about the problem. > > I don't think this is a bug in GCC-4.4 or 4.4.1 released versions. I > regularly build Sage on the Skynet cluster which had GCC 4.4.0 for a long > time, and lately has had GCC 4.4.1, and I haven't had any specific compiler > problems with building PARI. I just checked and my build with GCC 4.4.1 of > Sage-4.1.1.rc1 worked fine. Isn't your ubuntu GCC a prerelease or > something? Prerelease compilers almost never work to build Sage. > > Note that skynet only has Fedora, Suse, and Redhat boxes on it, and the GCC > on their was built by their sysadmin. It's > > [wst...@eno ~]$ gcc -v > Using built-in specs. > Target: x86_64-unknown-linux-gnu > Configured with: /usr/local/gcc-4.4.1/src/gcc-4.4.1/configure > --enable-languages=c,c++,fortran,java --with-gnu-as > --with-gnu-as=/usr/local/binutils-2.19.1/x86_64-Linux-fc-gcc-4.3.3/bin/as > --with-gnu-ld > --with-ld=/usr/local/binutils-2.19.1/x86_64-Linux-fc-gcc-4.3.3/bin/ld > --with-gmp=/usr/local/mpir-1.2.1/x86_64-Linux-core2-gcc-4.4.0 > --with-mpfr=/usr/local/mpfr-2.4.1/x86_64-Linux-core2-mpir-1.2.1-gcc-4.4.0 > --prefix=/usr/local/gcc-4.4.1/x86_64-Linux-core2-fc > Thread model: posix > gcc version 4.4.1 (GCC) > > (I'm not cc'ing this to the PARI list, since it is too difficult for me to > figure out how to post to their mailing list.) > > William --~--~---------~--~----~------------~-------~--~----~ To post to this group, send an email to [email protected] To unsubscribe from this group, send an email to [email protected] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
