Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 10 Apr 2010 16:20:40 +0200
From:      Thierry Thomas <thierry@FreeBSD.org>
To:        Andrea Venturoli <ml@netfence.it>
Cc:        pav@freebsd.org, freebsd-ports@freebsd.org
Subject:   Re: Problem with devel/libtool22? [was Re: OpenCASCADE marked broken]
Message-ID:  <20100410142040.GA64311@graf.pompo.net>
In-Reply-To: <4BBAD161.7060706@netfence.it>
References:  <4BA93494.9080601@netfence.it> <20100405222608.GD37471@graf.pompo.net> <4BBAD161.7060706@netfence.it>

next in thread | previous in thread | raw e-mail | index | archive | help
Le Mar  6 avr 10 à  8:14:57 +0200, Andrea Venturoli <ml@netfence.it>
 écrivait :

> Are you sure OpenCASCADE requires f77?
> I looked into the source and didn't find any file like "*.f" or "*.for", 
> although of course they might have other extension or be generated 
> during the build process.
> Also I can confirm it compiles with the base system gcc, although I 
> don't remember if I had gcc44 installed at the time, so it might also be 
> it picked f77 from gcc44 anyway.

You're right: it builds succesfully in a tinderbox without USE_FORTRAN,
which is just searched by configure.
Should be some remaining of an older version...

> The whole point of the above is that, due to gcc44's libstdc++ recently 
> becoming incompatible with older versions, if you mix different g++ 
> versions in the same project, you'll get "Undefined symbol" errors when 
> starting your executable.
> 
> In my case, I'm linking (amongst others) against Boost and OpenCASCADE, 
> so I need them to use the same compiler (and use that myself).
> 
> 
> 
> Again, this is not a problem with OpenCASCADE only, but potentially 
> affects any C++ library which is build with gcc44.

True: that's why I think that libtool should'nt force $CC and $LD.

Best regards,
-- 
Th. Thomas.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20100410142040.GA64311>