Re: [configure bus while building] flex beta version 2.5.15

[Tim Van Holder]

On Thu, 2002-08-29 at 16:40, Bruce Lilly wrote:
> Akim Demaille wrote:
> 
> > | # Be careful to initialize this variable, since it used to be cached.
> > | # Otherwise an old cache value of `no' led to `EXEEXT = no' in a Makefile.
> > | ac_cv_exeext=
> > | for ac_file in a_out.exe a.exe conftest.exe a.out conftest a.* conftest.* 
> > ; do
> > |    if test -x $ac_file ; then
> > |      case $ac_file in
> > |        *.$ac_ext | *.o | *.obj | *.xcoff | *.tds | *.d | *.pdb | *.xSYM ) 
> > ;;
> > |        a.out ) # We found the default executable, but exeext='' is most
> > |                # certainly right.
> > |                break;;
> > |        *.* ) ac_cv_exeext=`expr "$ac_file" : '[^.]*\(\..*\)'`
> > |              # FIXME: I believe we export ac_cv_exeext for Libtool --akim.
> > |              export ac_cv_exeext
> > |              break;;
> > |        * ) break;;
> > |      esac
> > |      break
> > |    fi
> > | done
> > | 
> > | There are similar tests throughout the configure script.
> > 
> > That's better, indeed.  Just -f instead of -x.
> 
> Using -x ensures that only executables are found; a further
> improvement (w.r.t. maintainability) would then be to remove
> all the .o .obj .xcoff .tds .d .pdb .xSYM cruft [of course,
> if configure didn't insist on using -g when compiling throw-
> away programs, much of that would have been unnecessary in
> the first place...].

But when cross-compiling, the executable bit might not be set (not
entirely sure though).
As for removing the .o ... test, some systems don't have the
'executable' attribute, so test has to determine executability from the
file data, and objects are executable code so may get flagged as
executable.  Also, attributes may be f*ed up if building on a mounted
filesystem (eg on a Samba NTFS share).  Keeping this specific case in
doesn't hurt and helps portability to DOSish systems.