Re: GMP on OSX errors

[glenbo]

the bad character, called a caron I found out, is at the end of the file,
and it is not in add_n.asm originally. when add_n.asm becomes tmp-add_n.s
the caron is added to the end of the file.

I wasn't quite sure what you were suggesting I try below, but I edited
add_n.asm with your siggestions below, and still had the problems. Adding:
        dnl
(with a tab in front of it) at the end of the add_n.asm file would make
the caron 1 tab length further in.

I started playing with the tmp-add_n.s file and was able to get it to
compile actually. I created tmp-add_n.2 removed the bad char, typed make
in one window and 'cp tmp-add_n2 tmp-add_n.s' in another. That worked, but
then it hoarfed on the next asm file. So, I did it for all 7 .asm files in
mpn/. Iknow that's a total kludge, but after the asm files were out of the
way it finished compiling with no errors. I was happy until I ran 'make
check' which didn't work:

Making check in mpn
Making check in tests
make[2]: Nothing to be done for `check'.
make[2]: Nothing to be done for `check-am'.
Making check in mpz
Making check in tests
make  t-mul t-tdiv t-tdiv_ui t-fdiv t-fdiv_ui t-gcd dive t-sqrtrem convert
io logic bit t-powm t-powm_ui t-pow_ui t-2exp reuse t-root t-jac t-bin
t-misc
cc -DHAVE_CONFIG_H -I. -I. -I../.. -I../..    -g -O2 -mpowerpc -c t-mul.c
/bin/sh ../../libtool --mode=link cc  -g -O2 -mpowerpc   -o t-mul  t-mul.o
../../libgmp.la 
mkdir .libs
cc -g -O2 -mpowerpc -o t-mul t-mul.o  ../../.libs/libgmp.a
/usr/bin/ld: archive: ../../.libs/libgmp.a has no table of contents, add
one with ranlib(1) (can't load from it)
make[3]: *** [t-mul] Error 1
make[2]: *** [check-am] Error 2
make[1]: *** [check-recursive] Error 1
make: *** [check-recursive] Error 1

So, I don't know what adds the caron, or if my trick is an honest way to
get a solid compile, but getting make check to work is where it breaks
now.

--glen

01100111 01101100 01100101
01101110 01100010 01101111


On 21 Dec 2000, Kevin Ryde wrote:

> glenbo <address@hidden> writes:
> >
> > well after running it with your suggestions below it changes the error
> > slightly
> >     tmp-add_n.s:64:Junk character -1 ().
> > 
> > hrm. Pine will not let me print the character that is problematic.
> > interesting.  The offending character is an upside down caret, like
> > SHIFT-6, but pointing down, and really small. I'm not even sure how to
> > type it. It's possible that that was the error all along and that it never
> > let me paste it into an email...
> >
> > ...
> >     addze   r3,r3           
> >     blr
> > 
> >     ^
> > 
> > I replaced the offending character with a caret.
> 
> Very mysterious.  Cc'ed to a reporter of success with Darwin+GMP, for
> any insights.
> 
> This is at the end of file I guess.  The add_n.asm doesn't contain a
> bad character to start with does it?  The distributed tarball is
> clean, naturally, but perhaps something got introduced.
> 
> If not, then does it happen if you go to the mpn directory can put the
> following through m4
> 
>         include(`../config.m4')
> 
> or
> 
>         include(`../config.m4')
>         EPILOGUE(foo)
> 
> or, with a tab before the dnl
> 
>         include(`../config.m4')
>                 dnl
> 
> I have a suspicion that with "define(<SIZE>, <dnl>)" in config.m4 that
> EPILOGUE will leave the last line of the file with a tab and no
> newline, if that might upset m4 or whatever.
>