On Thu, Oct 09, 2008 at 03:18:50PM +0200, Jean-Marc Lasgouttes wrote: > Enrico Forestieri <[EMAIL PROTECTED]> writes: > >> Applied then. And now, is the nox.diff patch working in cygwin? > > > > Yes, it works. > > > >> I guess so, but I prefer to be sure... In particular, is Q_CYGWIN_WIN > >> defined as needed by pkg-config? > > > > Yes, it is, as --without-x still sets with_x=no. > > This was the whole point of my question :) Normally, you should be able > to just point pkg-config to the right place (qt/x11 or qt/win32) and lyx > should do the rest by itself...
For configuring, it is like that. > What is the difference between Q_CYGWIN_WIN and > defined(Q_WS_WIN32) && defined(Q_OS_CYGWIN) > ? > > It seems to me that this extra define is not really needed... It is needed because the window system choice is made in qglobal.h and I see no other way to have Q_WS_WIN32 defined, rather than Q_WS_X11. In other words, Q_WS_WIN32 gets defined on cygwin only because Q_CYGWIN_WIN is defined. When building Qt, this is accomplished in qmake.conf through QMAKE_COMPILER_DEFINES += Q_CYGWIN_WIN So, you are right that there is no difference, but only because Q_CYGWIN_WIN is defined, otherwise Q_WS_WIN32 and Q_OS_CYGWIN would never be both defined at the same time. On Mac, the role of Q_CYGWIN_WIN is played by __USE_WS_X11__ and maybe I could have used that. But I don't see the point, as instead of defining Q_CYGWIN_WIN you would have to anyway define __USE_WS_X11__ and then instead of simply #if defined(Q_OS_UNIX) && !defined(Q_CYGWIN_WIN) you would have to use the ugly #if defined(Q_OS_UNIX) && !(defined(Q_WS_WIN32) && defined(Q_OS_CYGWIN)) -- Enrico
