Manav, You may also wish to consult the man or info pages for your particular flavor of gcc regarding the interpretation of -ansi. There may be more specific alternatives that check whatever flavor of ISO compliance is important to you.
Unfortunately, the mpi specification was written before int32/int64 made it into the C specification. Ben On Sun, Aug 06, 2006 at 08:51:28AM -0400, Jeff Squyres wrote: > Open MPI was not designed to handle all of those flags. We do have an > "--enable-debug" mode that adds a bunch of compiler flags, but not all of > the ones that you specified. > > In particular, however, -pedantic works for me (it's automatically added > with --enable-debug). Running configure on my macbook with 10.4.7 with > --enable-debug shows: > > checking for long long... yes > > and > > checking size of long long... 8 > > and > > checking alignment of long long... 4 > > If you need more information, please send the data listed here: > http://www.open-mpi.org/community/help/ > > > > On 8/5/06 6:35 PM, "Manav Bhatia" <ma...@u.washington.edu> wrote: > > > Hi, > > > > I am using OpenMPI 1.1 on Mac OS 10.4.7 with intel processors. I > > compiled and installed using the following options: > > > > /configure --prefix=/usr/local/openmpi --disable-mpi-f77 --disable- > > mpi-f90 > > > > Everything compiles and runs fine in optimized mode using mpic++, > > however, I get the following compilation error message when compiling > > in debug mode > > > > /usr/local/openmpi/include/mpi.h:147: error: ISO C++ does not support > > 'long long' > > > > I am using the following compiler options in optimized mode > > > > -DNDEBUG -fno-common -O2 -felide-constructors -funroll-loops -fstrict- > > aliasing -Wdisabled-optimization > > > > and the following in debug mode > > > > -DDEBUG -fno-common -O0 -felide-constructors -g -ansi -pedantic -W - > > Wall -Wunused -Wpointer-arith -Wimplicit -Wformat -Wparentheses - > > Woverloaded-virtual -D_GLIBCXX_DEBUG -D_GLIBCXX_DEBUG_PEDANTIC > > > > Kindly help me understand what is happening here, and how I could > > resolve this error. Please let me know if I could help you with more > > information about this. > > > > Thanks > > Manav > > _______________________________________________ > > users mailing list > > us...@open-mpi.org > > http://www.open-mpi.org/mailman/listinfo.cgi/users > > > -- > Jeff Squyres > Server Virtualization Business Unit > Cisco Systems > _______________________________________________ > users mailing list > us...@open-mpi.org > http://www.open-mpi.org/mailman/listinfo.cgi/users
