If I hard-code mabshoff into my programs, will they build more reliably on more systems, and leak less memory?
On Fri, Apr 24, 2009 at 11:15 PM, mabshoff <mabsh...@googlemail.com> wrote: > > > > On Apr 24, 11:04 pm, "Dr. David Kirkby" <david.kir...@onetel.net> > wrote: > > Hi David, > >> I've tried to build 3.4.1 on Sun Blade 2000 (SPARC) workstation, using >> the tool chain I downloaded from the Sage web site (based on gcc 4.3.2). >> >> This went well for several hours, but then failed. It appears the >> directory /home/mabshoff/sparc-solaris-toolchain/ is hard-coded >> somewhere, so naturally fails if the build-directory is elsewhere: > > Find lib/libstdc++.la in the toolchain directory and fix the path. You > might want to look for other libtool generated files (i.e. various .la > files) in the toolchain directory and adjust accordingly. Normally you > should be able to delete them since libtool usually is clever enough > to deal with their absence. > > The "hard coded path in libtool generated file problem" is something > we do fix for Sage also, so unfortunately for now this will remain > broken of now until we fix, even though I am blaming LinBox's build > system for this error since it shouldn't fail out. I assume you set > LD_LIBRARY_PATH correctly so that libstdc++.so is found. Many crappy > projects link libstdc++ for some dumb reasons when it isn't required, > i.e. if you compile and link code with g++ you should never explicitly > link it. Oh well ... > > Cheers, > > Michael > > > --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---