This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: libg++-2_8_2-980731 failed make check on egcs-19980824 on rs
- To: rodneybrown at pmsc dot com
- Subject: Re: libg++-2_8_2-980731 failed make check on egcs-19980824 on rs
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Wed, 02 Sep 1998 11:13:05 -0600
- cc: egcs at cygnus dot com
- Reply-To: law at cygnus dot com
In message <9809029047.AA904711431@cc.pmsc.com>you write:
>
> libg++-2_8_2-980731 failed to complete the make check using
> egcs-19980824 on rs6000-ibm-aix4.2.0.0
>
> The make check failed in libg++/etc/lf with
> egcs-19980824.obj/gcc/include/curses.h:130: \
> abstract declarator used as declaration
> when this was fixed the link failed with
> collect2: Library libtermcap not found
> Does libg++ assume GNU termcap?
> The failure in curses.h is caused by a 'typedef int bool;' which isn't
> guarded by __cplusplus as the successive 'typedef char bool;' and which
> is used in the intervening function prototypes. Kludging around this gave t he
> link failure with termcap.
> Should the curses.h problem be fixed by fixincludes?
I believe curses.h needs to be fixed by fixincludes.
I believe "bool" is a reserved keyword for C++ and C (ISOC9X). It
should not be used by header files.
Many systems had this problem. If you look at regress.html you'll see
AIX (and others) as passing libg++ after fixing the curses.h botch.
jeff