Re: [Xsb-development] XSB specialization is broken

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 454-5900

No, I am compiling with the default.
My compilers are versions 2.96 and 3.04.
As I said earlier, it seems like a gcc bug possibly on PIII machines.
This is because when I compile with the opt level O1 then everything works.

	--michael  

> Hi Michael,
> 
> I tried it on a newly created tree (from CVS as of today) with the
> following versions of gcc, and got the same (good) result. Are you
> compiling with some options different than the default? 
> 
> lfmobile:~$ gcc -v
> Reading specs from /usr/lib/gcc-lib/i386-linux/2.95.4/specs
> gcc version 2.95.4 20011002 (Debian prerelease)
> lfmobile:~$ gcc-3.0 -v
> Reading specs from /usr/lib/gcc-lib/i386-linux/3.0.4/specs
> Configured with: ../src/configure -v --enable-languages=c,c++,java,f77,objc --prefix=/usr --infodir=/share/info --mandir=/share/man --enable-shared --with-gn
> u-as --with-gnu-ld --with-system-zlib --enable-long-long --enable-nls --without-included-gettext --disable-checking --enable-threads=posix --enable-java-gc=b
> oehm --with-cpp-install-dir=bin --enable-objc-gc i386-linux
> Thread model: posix
> gcc version 3.0.4
> lfmobile:~$ gcc-3.1 -v
> Reading specs from /usr/lib/gcc-lib/i386-linux/3.1.1/specs
> Configured with: /mnt/data/gcc-3.1/gcc-3.1-3.1.1ds2/src/configure -v --enable-languages=c,c++,java,f77,proto,objc,ada --prefix=/usr --mandir=/usr/share/man -
> -infodir=/usr/share/info --with-gxx-include-dir=/usr/include/g++-v3-3.1 --enable-shared --with-system-zlib --enable-long-long --enable-nls --without-included
> -gettext --enable-clocale=gnu --enable-__cxa_atexit --enable-threads=posix --enable-java-gc=boehm --enable-objc-gc i386-linux
> Thread model: posix
> gcc version 3.1.1 20020703 (Debian prerelease)
> 
> -Luis
> 
> ki...@cs... (Michael Kifer) writes:
> 
> > I don't know what is going on. I wiped out everything (make distclean),
> > configured with gcc-3.04, compiled, then tried
> > 
> > [user]
> > ?- X=23.234567, number_codes(X,L),atom_codes(Y,L),writeln([X,L,Y]).
> > EOF
> > 
> > Then I wiped out everything again and configured with gcc-2.96, recompiled,
> > then did the same as above. In both cases it gave a wrong result.
> > 
> > Then I wiped out all the xwam files in syslib, cmplib, and elsewhere,
> > updated from cvs, and ran the test again. Same result.
> > 
> > Aha!
> > I then recompiled with --optimization-level=O1, and it works!  I dunno
> > what the calprit is. It is not the compiler, since I tried 2.96 as well.
> > 
> > Maybe libc? Mine is 2.2.4.
> > Or could be wrong optimization that gcc does for P3 somewhere.
> > 
> > 
> > 	--michael  
> > 
> > 
> > > Michael wrote: 
> > > 
> > > > And to back it up, I recompiled XSB with gcc-2.96, did superdevel,
> > > wiped out all the xwam files in the testsuite, and then ran it again.
> > > The testsuite has failed.
> > > 
> > > I suspect that you were using old, good .xwam files in the testsuite, which
> > > is why it didn't fail for you. 
> > > 
> > > ------------------
> > > 
> > > Ok, 
> > > 
> > > 
> > > -- I updated my system pretty completely, (there are a few diffs from
> > > the common system, but I looked at them all, and took out any of my
> > > changes that had a conceivable chance of accounting for the error)
> > > 
> > > -- I remade xsb with superdevel
> > > 
> > > -- I removed all .xwam files from the testsuite.  The testsuite passed
> > > when I ran it.
> > > 
> > > -- Just to make sure there weren't any timing problems in the make
> > > process, I repeated the make and testsuite.  It ran again.
> > > 
> > > Also, 
> > > 
> > > 
> > > | ?- [user].
> > > [Compiling user]
> > > ?- X=23.234567,number_codes(X,L),writeln(L).
> > > [user compiled, cpu time used: 0.0900 seconds]
> > > [user loaded]
> > > [50,46,51,50,51,52,53,54,101,43,48,49]
> > > 
> > > yes
> > > | ?- X=23.234567,number_codes(X,L),writeln(L).
> > > [50,46,51,50,51,52,53,54,101,43,48,49]
> > > 
> > > X = 23.2346
> > > L = [50,46,51,50,51,52,53,54,101,43,48,49]
> > > 
> > > 
> > > yes
> > > 
> > > This seems contradictory to Michael's description.  I'm guessing that
> > > somewhere Michael has a library file that is different than one I use,
> > > and this is accounting for the different results (?)
> > > 
> > > 
> > > 
> > > 
> > 
> > 
> > 
> > 
> > -------------------------------------------------------
> > This sf.net email is sponsored by:ThinkGeek
> > Gadgets, caffeine, t-shirts, fun stuff.
> > http://thinkgeek.com/sf
> > _______________________________________________
> > Xsb-development mailing list
> > Xsb...@li...
> > https://lists.sourceforge.net/lists/listinfo/xsb-development
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Xsb-development mailing list
> Xsb...@li...
> https://lists.sourceforge.net/lists/listinfo/xsb-development
> 