Re: [PATCHSET] [4/4] Fix various issue in TUI

[Eli Zaretskii <eliz@gnu.org>]

> Date: Mon, 05 Jan 2015 19:54:42 +0000
> From: Pedro Alves <palves@redhat.com>
> 
> On 12/31/2014 05:56 PM, Eli Zaretskii wrote:
> > Well, one patch is Windows-specific after all.  This patch makes sure
> > windows-termcap is not compiled when GDB is linked against ncurses,
> 
> ...
> 
> > and also makes the file a no-op should it compile in that
> > configuration.  
> 
> With the configure.ac change, how can that happen?

It shouldn't.

> > --- gdb/configure.ac~0	2014-10-29 21:45:50 +0200
> > +++ gdb/configure.ac	2014-12-30 07:42:27 +0200
> > @@ -627,9 +627,10 @@
> >      ac_cv_search_tgetent="none required"
> >      ;;
> >    *mingw32*)
> > -    ac_cv_search_tgetent="none required"
> > -    CONFIG_OBS="$CONFIG_OBS windows-termcap.o"
> > -    ;;
> > +    if test x"$prefer_curses" = xyes; then
> > +      ac_cv_search_tgetent="none required"
> > +      CONFIG_OBS="$CONFIG_OBS windows-termcap.o"
> 
> I'm confused on the predicate here.  How can we tell
> from $prefer_curses that the curses library doesn't include
> termcap?  AFAICS, if the TUI is enabled, it'll always be
> "yes".  But isn't that the case where you _don't_
> want windows-termcap.o?  And shouldn't this be checking
> $curses_found?

Sorry, I sent the wrong patch here.  It should be !=, of course.

> > --- gdb/windows-termcap.c~0	2014-06-11 18:34:41 +0300
> > +++ gdb/windows-termcap.c	2014-12-29 15:42:44 +0200
> > @@ -19,6 +19,11 @@
> >     You should have received a copy of the GNU General Public License
> >     along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
> >  
> > +
> > +#include "config.h"
> 
> All files in gdb/ should start with "defs.h" instead.

OK.

Thanks.