Re: [cp-patches] Make qt-peers build and run out of the box

[Dalibor Topic]

Mark Wielaard wrote:
> Hi,
>
> After a short discussion on irc people seemed to assume that the .pc
> $includedir given by pkg-config was in error. So I changed the patch a
> little bit and included a simple autoconf check to make sure we always
> get the include dirs in QT_CFLAGS right (in case the $includedir in
> the .pc file gets "fixed" in a later release).

That's a very good idea!

> 2005-08-23  Mark Wielaard  <address@hidden>
>
>        * configure.ac: Check for QtGui >= 4.0.1. Check for and
>        add extra include dirs ($includedir/Qt) to QT_CFLAGS.

fine

>        * native/jni/qt-peer/.cvsignore: Add slotcallbacks.moc.h.

fine

>        * native/jni/qt-peer/Makefile.am (QT_CXXFLAGS): Removed.
>        (libqtpeer_la_CXXFLAGS): Removed.
>        (libqtpeer_la_LDFLAGS): Removed.
>        (AM_LDFLAGS): Set to @CLASSPATH_MODULE@ @address@hidden
>        (AM_CXXFLAGS): Set to @address@hidden

fine.

I tried it and works like a charm in my setup as well, builds nice with 
qt4.0.1, fails otherwise, and the gtk build still works fine, too ;)

Thank you very much for improving the patch so quickly. Good work!

>        * native/jni/qt-peer/.cvsignore: Add slotcallbacks.moc.h.

while that's a cool hack, there is no need to list it twice in the 
ChangeLog ;)

> With this no "cleanup" of the #include statements in the .cpp files is
> necessary anymore.

Yeah, thanks for not messing with the include files.

> It seems to work fine for everything I throw at it.

Works fine, here, too.

> OK to commit?

Fine from my side. It is a really nice improvement over the previous 
state of affairs, thanks for doing the hard work, and responding so 
readily to peer review. You rock ;)

cheers,
dalibor topic