RE: [Queue-developers] solaris gcc cross-compile cure

[Sam L. <sam...@an...>]

OK, in line 238,309,350,379 etc change the spelling of the word "undefine"

Weird; I;m sure something is interpreting that text in a way it shouldn't.

Sam

> -----Original Message-----
> From: Sam Liddicott [mailto:sam...@an...]
> Sent: 11 April 2001 15:06
> To: que...@li...
> Subject: RE: [Queue-developers] solaris gcc cross-compile cure
> 
> 
> You are right (thanks) the cvs of autoconf solves ./configure issues
> (apart from --sharedstatedir being ignored)
> 
> EXCEPT: make insists on re-running autoconf which fails:
> 
> bash-2.03# make
> cd . && aclocal
> cd . && automake --gnu Makefile
> configure.in: 26: automake requires `AM_PROG_LEX', not `AC_PROG_LEX'
> configure.in: 27: `AC_DECL_YYTEXT' is covered by `AM_PROG_LEX'
> cd . && autoconf
> configure.in:25: warning: AC_PROG_LEX invoked multiple times
> configure.in:26: warning: AC_PROG_LEX invoked multiple times
> configure.in:27: warning: AC_PROG_LEX invoked multiple times
> configure.in:238: error: undefine: undefined:
> configure.in:238: the top level
> make: *** [configure] Error 1
> 
> And I can't find anything wrong on line 238 of configure.in
> 
> Sam
> 
> > -----Original Message-----
> > From: W. G. Krebs [mailto:wer...@ya...]
> > Sent: 10 April 2001 18:29
> > To: que...@li...
> > Subject: Re: [Queue-developers] solaris gcc cross-compile cure
> > 
> > 
> > AFAIK, the solution is to use a development (CVS) version of 
> > GNU Autoconf, one
> > reason that having ./configure and other autoconf stuff in 
> the GQ CVS
> > respository is a good idea IMHO.
> > 
> > Sam Liddicott wrote:
> > 
> > > The reason ./configure thinks solaris gcc is a cross 
> > compiler is because
> > > some of the test cases (like to detect endian-ness) fail 
> > either way; so it
> > > assumes the test could not run because it is a cross compiler.
> > >
> > > I looked at why the bigendian test failed to run either 
> > way, it seems to be
> > > because solaris does *NOT* define either:
> > > BYTE_ORDER
> > > or
> > > BIG_ENDIAN
> > > or
> > > LITTLE_ENDIAN
> > >
> > > anywhere!  (Well, maybe in 
> > /usr/include/arpa/nameser_compat.h (ultimatly
> > > included from resolv.h) but possibly that is not part of 
> > the standard
> > > install?)
> > >
> > > /usr/include/sys/isa_defs.h DOES define _LITTLE_ENDIAN, 
> > _BIG_ENDIAN but not
> > > BYTE_ORDER
> > > _BIG_ENDIAN and _LITTLE_ENDIAN are not defined to have 
> > values, but just to
> > > be "set"
> > >
> > > BYTE_ORDER is only defined in name_ser.compat
> > >
> > > So whats the best way to solve this for solaris?
> > >
> > > Sam
> > >
> > > _______________________________________________
> > > Queue-developers mailing list 
> Que...@li...
> > > To unsubscribe, subscribe, or set options:
> > > http://lists.sourceforge.net/lists/listinfo/queue-developers
> > 
> > 
> > _______________________________________________
> > Queue-developers mailing list Que...@li...
> > To unsubscribe, subscribe, or set options:
> > http://lists.sourceforge.net/lists/listinfo/queue-developers
> > 
> 
> _______________________________________________
> Queue-developers mailing list Que...@li...
> To unsubscribe, subscribe, or set options:
> http://lists.sourceforge.net/lists/listinfo/queue-developers
>