Re: GMP on OSX errors

[Kevin Ryde]

glenbo <address@hidden> writes:
>
> ok 'm4 ../config.m4' returns eighteen empty lines then the caron
> character.

Hmm.  Ok, we've got something fishy with m4 (or with the macros).  How
about the following

divert(-1)
discarded
divert

or instead the following

m4wrap(foo)
define(foo)

or even the following

m4wrap(foo)
define(foo,`ifelse(1,0,dummy)')

Do you know what version of m4 you've got?  You might get more joy
from GNU m4, if it supports macos.


> 'ranlib .libs/libgmp.a' returns:
> 
> ranlib: archive member: .libs/libgmp.a(diveby3.o) offset in archive not a
> multiple of sizeof(unsigned long) (must be since member is an object file)

Now it's just getting weird.

You can see the "ar" command libtool runs.  (Delete libgmp.la and do
"make" again to force it to re-run.)  It should be a straightforward,
if rather long "ar" invocation.  Are there maybe some magic options
needed when making a library out of object files or something?

I know very little about macos, if you couldn't tell.  You might have
to hunt around for where experts on things like "ar" on your system
hang out.  But we'd very much like to know what gmp needs to do to
work though.