- files/patch-configureac * use a better way to detect the SDL header * (the old way did not work for me on several systems) - Makefile - files/patch-plugins::Makefile.in - files/patch-plugins::configure * add WITH_ESOUND knob and enable the plugin build - files/patch-plugins::sound_esd::Makefile - files/patch-plugins::sound_esd::sound_esd.h * make the esound plugin build- and loadable - files/patch-plugins::vidinput_bsd::vidinput_bsd.cxx - files/patch-plugins::vidinput_bsd::vidinput_bsd.h * fix the bsdvideo (bktr(4) based) plugin * convert to the correct API - files/patch-src::ptlib::unix::udll.cxx * add a small "error handler" to detect unresolved symbols in plugins - pkg-plist * add more plugins * use @dirrmtry for DATADIR, to don't complain during upgrades (net/openh323 installs its codecs in ${PREFIX}/lib/pwlib/codecs) - Makefile - files/patch-plugins::vidinput_avc::Makefile - files/patch-plugins::vidinput_dc::Makefile.in * add very experimental support for firewire(4) based video devices * as we do not have ports for all necessary libraries, provide information how to obtain the archives for them from our FTP server ! As the appropriate part is completely commented in the ports Makefile, it breaks nothing. It is mostly for testing purposes for brave people with firewire(4) based video devices. ! Fix: - apply the patch - add new files to CVS How-To-Repeat: <code/Input/activities to reproduce the problem (multiple lines)>
State Changed From-To: open->feedback Awaiting maintainers feedback
This patch looks great. Good to see someone working on bsdvideo. Note that patch ports/95400 also needs to go in. Be great if they could go together and only get one version bump. -steve
Approved.
State Changed From-To: feedback->open Maintainer approved.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, regarding to the latest follow-ups for PR/95424 which where kindly provided by Boris Samorodov, here is a more refined version for files/patch-configureac in devel/pwlib. The problem is actually, that the configure script adds a line "#define P_SDL 0" if one uses the WITHOUT_SDL knob or SDL is not available at all. This is completely wrong as this actually defines P_SDL with value "0" instead of undefining it. I double checked the configure.ac file and found no other point beside the P_SDL where this is done in this way. I think the upstream maintainers made an error there. Unfortunately, this error prevents the patched version (PR/95424) of net/ohpone and maybe other ports too from building correctly on systems where SDL is not installed at allor disabled during the build of devel/pwlib by using the WITHOUT_SDL knob. The procedure for fixing this PR should now be: 1. apply the initial patch and add new files to CVS 2. apply the following patch to files/patch-configureac Thanks a lot Joerg - -- The beginning is the most important part of the work. -Plato - --- patch-configureac.diff begins here --- - --- patch-configureac.orig Wed Apr 12 00:07:09 2006 +++ patch-configureac Wed Apr 12 00:06:14 2006 @@ -1,5 +1,5 @@ --- configure.ac.orig Wed Feb 22 12:53:28 2006 - -+++ configure.ac Wed Apr 12 00:06:55 2006 ++++ configure.ac Tue Apr 11 23:47:45 2006 @@ -136,7 +136,7 @@ x86 | i686 | i586 | i486 | i386 ) MACHTYPE=x86 ;; @@ -107,6 +107,15 @@ #include <netdb.h>], [getnameinfo(NULL, 0, NULL, 0, NULL, 0, 0);], HAS_IPV6=yes, HAS_IPV6=no) if test $HAS_IPV6 = no ; then +@@ -1485,7 +1486,7 @@ + HAS_SDL= + if test "$enable_sdl" = "no" ; then + AC_MSG_NOTICE(SDL disabled) +- AC_DEFINE(P_SDL, 0) ++# AC_DEFINE(P_SDL, 0) + else + dnl SDL checks + if test "${SDLLIB:-unset}" = "unset" ; then @@ -1501,9 +1502,9 @@ AC_CHECK_LIB(SDL, SDL_CreateYUVOverlay, HAS_SDL=1, HAS_SDL=, -L${X11LIBDIR} -lpthread -lX11 -lXext) fi - --- patch-configureac.diff ends here --- -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.3 (FreeBSD) iD8DBQFEPCnLSPOsGF+KA+MRAstMAJ9VVHVizRkrQaq16QBGUi+/Rm8sSwCffBez RwCbQZfq9qwok2Z061KBamQ= =i0LY -----END PGP SIGNATURE-----
Good catch. I concur completely. -steve
State Changed From-To: open->closed Committed, thanks!