Re: ppc-darwin, failed compile

Posted by Luigi Ballabio-4 on Oct 26, 2001; 4:46am
URL: http://quantlib.414.s1.nabble.com/ppc-darwin-failed-compile-tp1784p1786.html

At 11:42 AM 10/26/01 +0200, Ferdinando Ametrano wrote:
>At 06:59 PM 10/25/01 +0200, you wrote:
> >I downloaded QuantLib-0.2.0 (Unix version) and attempted to compile for
> Mac OS X
> >(=BSD 4.4). It failed.
>
>Thank you Jonas.
>Would you please report if it fails also using the latest CVS snapshot?

Never mind the CVS snapshot, it would fail just the same.
Open the config.status that ./configure created. There will be a bunch of
lines in the middle of it which are fed to sed and look like the ones below:

s,@ACLOCAL@,${SHELL} ~/QuantLib-0.2.0/config
/home/users/l/lb/lballabio/QuantLib-0.2.0/config/missing --run aclocal,;t t

which for some reason are split over two lines. Is there any
autoconf/automake expert around here which knows why this happens and how
not to make it happen?

Edit the file and join the lines so that each sed command is on one line.
After that I've been able to run

localhost> ./config.status

which created the makefiles.

All is not well, though---I then ran make, and the files compiled, but the
final link failed with a whole bunch of multiple definition errors. From
the look of them, it seems to me that for some reason the compiler failed
to recognize inline declarations, so that the corresponding methods were
defined in each unit including the corresponding header. How this happens
beats me.
Is there any cc/darwin expert around here which knows why this happens and
how not to make it happen?

Sorry this doesn't help much,
                                 Luigi