libstdc++ in 4.5.0 b3

[Nick.SIMON at syntegra.bt.co.uk]

OK, I wasn't sure how libstdc++ got picked up.  What I should say is that it
doesn't appear to be linked in, hence the undefined cout, endl etc..  Should
there be an explicit -L option for these libraries in the link step, or
should this be unnecessary?  Which of the libraries you list (which are all
present on my system) should be used?


-- Nick Simon 

> -----Original Message-----
> From: Ralf Corsepius [mailto:corsepiu at faw.uni-ulm.de]
> Sent: 18 May 2000 12:24
> To: Nick.SIMON at syntegra.bt.co.uk
> Cc: rtems-users at oarcorp.com
> Subject: Re: libstdc++ in 4.5.0 b3
> 
> 
> Nick.SIMON at syntegra.bt.co.uk wrote:
> > 
> > For some reason this isn't getting included in builds, with 
> the result that
> > cout, endl etc come up as undefined. As well as in my own 
> code, this occurs
> > in cdtest if it's set to use cout.
> I don't understand, why should libstdc++ be included into rtems
> builts?
> libstdc++ comes with gcc and is not part of rtems sources. Therefore
> it is getting build when building gcc.
> 
> > 
> > There is no libstdc++ under the install directory, 'though 
> there is in
> > /opt/rtems/lib...
> 
> # rpm -q -l -p powerpc-rtems-gcc-gcc2.95.2newlib1.8.2-5.i386.rpm |
> grep libstdc
> /opt/rtems/lib/gcc-lib/powerpc-rtems/2.95.2/ca/libstdc++.a
> /opt/rtems/lib/gcc-lib/powerpc-rtems/2.95.2/le/ca/libstdc++.a
> /opt/rtems/lib/gcc-lib/powerpc-rtems/2.95.2/le/libstdc++.a
> /opt/rtems/lib/gcc-lib/powerpc-rtems/2.95.2/libstdc++.a
> /opt/rtems/lib/gcc-lib/powerpc-rtems/2.95.2/lin/libstdc++.a
> /opt/rtems/lib/gcc-lib/powerpc-rtems/2.95.2/nof/ca/libstdc++.a
> /opt/rtems/lib/gcc-lib/powerpc-rtems/2.95.2/nof/le/ca/libstdc++.a
> /opt/rtems/lib/gcc-lib/powerpc-rtems/2.95.2/nof/le/libstdc++.a
> /opt/rtems/lib/gcc-lib/powerpc-rtems/2.95.2/nof/libstdc++.a
> /opt/rtems/powerpc-rtems/lib/ca/libstdc++.a.2.10.0
> /opt/rtems/powerpc-rtems/lib/le/ca/libstdc++.a.2.10.0
> /opt/rtems/powerpc-rtems/lib/le/libstdc++.a.2.10.0
> /opt/rtems/powerpc-rtems/lib/libstdc++.a.2.10.0
> /opt/rtems/powerpc-rtems/lib/lin/libstdc++.a.2.10.0
> /opt/rtems/powerpc-rtems/lib/nof/ca/libstdc++.a.2.10.0
> /opt/rtems/powerpc-rtems/lib/nof/le/ca/libstdc++.a.2.10.0
> /opt/rtems/powerpc-rtems/lib/nof/le/libstdc++.a.2.10.0
> /opt/rtems/powerpc-rtems/lib/nof/libstdc++.a.2.10.0
> 
> Please make sure that these files are present in your installation.
> 
> > 
> > The configuration setting is:
> > 
> > ../rtems-4.5.0-beta3/configure \
> >      --target=powerpc-rtems \
> >      --prefix=/usr/rtm-src/tools/install-rtems-4.5.0-beta3 \
> >      --disable-posix \
> >      --disable-itron \
> >      --enable-cxx \
> >      --enable-rtemsbsp="dhe860boot"
> > 
> 
> I guess your problem is originating from the make-exe rule in your
> make/custom/<BSP>.cfg.
> 
> There is a known (and still unfixed) problem with the general
> working principle of make-exe and c++, which for some strange reason
> only shows with some permutations of multilib flags for some
> targets.
> You might try to add -lstdc++ to the make-exe rule in your <BSP>.cfg
> (cf. make/custom/gensh2.cfg - the sh2 is known to suffer from this
> problem).
> 
> Ralf
> 
> -- 
> Ralf Corsepius 
> Forschungsinstitut fuer Anwendungsorientierte Wissensverarbeitung
> (FAW)
> Helmholtzstr. 16, 89081 Ulm, Germany     Tel: +49/731/501-8690
> mailto:corsepiu at faw.uni-ulm.de           FAX: +49/731/501-999  
> http://www.faw.uni-ulm.de
>