Re: wxterminal compilation

[<tim...@en...>]

Hi James,


James R. Van Zandt wrote:
> Here, wxterminal compilation fails as follows:
>
> Script started on Fri Jul 14 21:49:29 2006
> vanzandt:/usr/local/src/gnuplot/cvs/gnuplot$ make
>
> (...)
> if g++ -DHAVE_CONFIG_H -I. -I. -I..  -I../term -I../term -DBINDIR=3D\"/=
usr/local/bin\" -DX11_DRIVER_DIR=3D\"/usr/local/libexec/gnuplot/4.1\" -DG=
NUPLOT_PS_DIR=3D\"/usr/local/share/gnuplot/4.1/PostScript\" -DCONTACT=3D\=
"gnu...@li...\" -DHELPFILE=3D\"/usr/local/share/gnu=
plot/4.1/gnuplot.gih\" -DPREFIX=3D\"/usr/local\" -DVERSION_MAJOR=3D\"4.1\=
" -I/usr/X11R6/include -I/usr/include -I/usr/include/cairo   -I/usr/inclu=
de/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include    -g -O=
2 -I/usr/lib/wx/include/gtk2-unicode-release-2.6 -I/usr/include/wx-2.6 -D=
GTK_NO_CHECK_CASTS -D__WXGTK__ -D_FILE_OFFSET_BITS=3D64 -D_LARGE_FILES -D=
_LARGEFILE_SOURCE=3D1 -DNO_GCC_PRAGMA   -I/usr/include/cairo   -I/usr/inc=
lude/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include   -MT =
wxt_gui.o -MD -MP -MF ".deps/wxt_gui.Tpo" -c -o wxt_gui.o `test -f 'wxter=
minal/wxt_gui.cpp' || echo './'`wxterminal/wxt_gui.cpp; \
>  =20
My g++ command looks the same.

> 	then mv -f ".deps/wxt_gui.Tpo" ".deps/wxt_gui.Po"; else rm -f ".deps/w=
xt_gui.Tpo"; exit 1; fi
>
> In file included from /usr/lib/gcc/i486-linux-gnu/4.0.4/../../../../inc=
lude/c++/4.0.4/ext/new_allocator.h:37,
>
>      (...)
>  =20
> ../new:1: error: stray '@' in program
>  =20
Hmm, as you said below, these @ are in comments, they are used for=20
source documentation (doxygen or similar).

How can they appear in the compilation ???


> In file included from /usr/lib/gcc/i486-linux-gnu/4.0.4/../../../../inc=
lude/c++/4.0.4/bits/stl_construct.h:65,
>
>                  from /usr/lib/gcc/i486-linux-gnu/4.0.4/../../../../inc=
lude/c++/4.0.4/memo
> ...
> <~800 lines omitted>
> ...
>
> /usr/lib/gcc/i486-linux-gnu/4.0.4/../../../../include/c++/4.0.4/bits/st=
l_list.h:571: error: return type 'class std::allocator<gp_command>' is in=
complete
> /usr/lib/gcc/i486-linux-gnu/4.0.4/../../../../include/c++/4.0.4/bits/st=
l_list.h: In member function 'std::_List_node<_Tp>* std::_List_base<_Tp, =
_Alloc>::_M_get_node() [with _Tp =3D gp_event_t, _Alloc =3D std::allocato=
r<gp_event_t>]':
> /usr/lib/gcc/i486-linux-gnu/4.0.4/../../../../include/c++/4.0.4/bits/st=
l_list.h:438:   instantiated from 'std::_List_node<_Tp>* std::list<_Tp, _=
Alloc>::_M_create_node(const _Tp&) [with _Tp =3D gp_event_t, _Alloc =3D s=
td::allocator<gp_event_t>]'
> /usr/lib/gcc/i486-linux-gnu/4.0.4/../../../../include/c++/4.0.4/bits/st=
l_list.h:1152:   instantiated from 'void std::list<_Tp, _Alloc>::_M_inser=
t(std::_List_iterator<_Tp>, const _Tp&) [with _Tp =3D gp_event_t, _Alloc =
=3D std::allocator<gp_event_t>]'
> /usr/lib/gcc/i486-linux-gnu/4.0.4/../../../../include/c++/4.0.4/bits/st=
l_list.h:774:   instantiated from 'void std::list<_Tp, _Alloc>::push_back=
(const _Tp&) [with _Tp =3D gp_event_t, _Alloc =3D std::allocator<gp_event=
_t>]'
> wxterminal/wxt_gui.cpp:2661:   instantiated from here
> /usr/lib/gcc/i486-linux-gnu/4.0.4/../../../../include/c++/4.0.4/bits/st=
l_list.h:313: error: invalid use of undefined type 'class std::allocator<=
gp_event_t>'
> /usr/lib/gcc/i486-linux-gnu/4.0.4/../../../../include/c++/4.0.4/bits/st=
ringfwd.h:49: error: declaration of 'class std::allocator<gp_event_t>'
> /usr/lib/gcc/i486-linux-gnu/4.0.4/../../../../include/c++/4.0.4/bits/st=
l_list.h:313: error: '<declaration error>' is not a class or namespace
>  =20
Hmm (again).
I am far from being a C++ specialist, but I am using very basic features =

of the STL (standard template library), namely the template list<>,=20
which implements a linked list.
Looks like your g++ installation has problems.

> This is with a recent compiler:
>   $ gcc --version
>   gcc (GCC) 4.0.4 20060422 (prerelease) (Debian 4.0.3-2)
>   Copyright (C) 2006 Free Software Foundation, Inc.
>   This is free software; see the source for copying conditions.  There =
is NO
>   warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PU=
RPOSE.
>
> ...so maybe it's being more strict about the language.  However, "make
> CC=3Dgcc-3.3" fails the same way.
>  =20
I am using "stock" gcc-4.0.3, maybe Debian is patching it with something =

wrong... I don't really know...
> FWIW I think it's complaining about /usr/include/c++/4.0.4/new, but all=

> the "@" characters appear to be in comments:
>  =20
Yes, that's a bad sign to see failures there !
Are you using some sort of bleeding-edge repository ? I am sorry, but I=20
have no clue where it might come from. And google doesn't help here. I=20
will try to ask on irc...

Best regards,

Timoth=C3=A9e