bug-ncurses
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: ncurses 5.7 compile failure when crosscompiling to arm


From: Thomas Dickey
Subject: Re: ncurses 5.7 compile failure when crosscompiling to arm
Date: Sat, 13 Mar 2010 14:37:45 -0500
User-agent: Mutt/1.5.20 (2009-06-14)

On Tue, Mar 09, 2010 at 01:18:01PM +0100, Bernhard Rosenkraenzer wrote:
> On Tue, 09 Mar 2010 12:56:04 +0100, Bernhard Rosenkraenzer
> <address@hidden> wrote:
> > Hi,
> > ncurses 5.7 compiles and works fine on x86_64, but when I try to
> > crosscompile it to armv6j-unknown-linux-gnueabi, I get (while compiling
> a
> > tool that is run on the x86_64 build machine, not even while building
> the
> > arm lib):
> > 
> > gcc -o make_hash -DHAVE_CONFIG_H -I../ncurses -I. -I./../include
> > -I../include   ./tinfo/make_hash.c
> > In file included from ./tinfo/make_hash.c:40:0:
> > ../ncurses/curses.priv.h:559:2: error: expected specifier-qualifier-list
> > before ‘cchar_t’
> 
> Comparing what happens when I build it for x86_64 with what happens when
> building it for arm, the difference that matters is that the native build
> passes -D_GNU_SOURCE while the crosscompiler build doesn't. If I force
> -D_GNU_SOURCE into BUILD_CCFLAGS, the problem goes away.

yes, that's been reported a couple of times: the configure script is checking
for whatever _XXX_SOURCE, but only for the target.  It doesn't make configure
checks for the build-compiler.  Fixing it in the configure script is "doable",
but only with a lot of duplication.

On the other hand, I see that I can suppress the definition as used by
make_hash, as done for the termlib special case, i.e., #define NCURSES_CH_T to
void.

I'll add that to today's patch...

-- 
Thomas E. Dickey <address@hidden>
http://invisible-island.net
ftp://invisible-island.net

Attachment: signature.asc
Description: Digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]