Re: UPDATE: graphics/opencv
On Sat Jun 25, 2016 at 08:07:36PM +0200, Rafael Sadowski wrote: > Hi ports@, > > update OpenCV stable branch to 2.4.13. OpenCV 3 from WIP githup is not > ready for other ports in tree. > > core ports features: > - remove XXX tags in digikam-kde4 and kde-applications. > -- Not more necessary since the last opencv update > - use OpenCVs testsuite to run tests (looks much better) > - update to MODJAVA_VER 1.8 (burn in hell insecure JAVA) > naddy@ informed me that I forgot to update the distinfo in my last diff. Here is a new version, tested last night with -current ports tree. Thanks naddy@ Index: graphics/digikam-kde4/Makefile === RCS file: /cvs/ports/graphics/digikam-kde4/Makefile,v retrieving revision 1.21 diff -u -p -u -p -r1.21 Makefile --- graphics/digikam-kde4/Makefile 6 May 2016 21:39:03 - 1.21 +++ graphics/digikam-kde4/Makefile 11 Aug 2016 21:49:36 - @@ -178,11 +178,6 @@ TEST_IS_INTERACTIVE = X11 .include -# XXX should fix OpenCV CMake config on Java-enabled archs instead! -.if ${ARCH} == "amd64" -BUILD_DEPENDS +=graphics/opencv,-java -.endif - #.if ${ARCH} == "i386" CONFIGURE_ENV =CXXFLAGS="${CXXFLAGS} -DOPENBSD_LIBKDCRAW_NO_CRASH=0" #.else Index: graphics/opencv/Makefile === RCS file: /cvs/ports/graphics/opencv/Makefile,v retrieving revision 1.37 diff -u -p -u -p -r1.37 Makefile --- graphics/opencv/Makefile15 Apr 2016 08:53:49 - 1.37 +++ graphics/opencv/Makefile11 Aug 2016 21:49:36 - @@ -1,46 +1,45 @@ # $OpenBSD: Makefile,v 1.37 2016/04/15 08:53:49 shadchin Exp $ ONLY_FOR_ARCHS = ${GCC4_ARCHS} -ONLY_FOR_ARCHS-java = amd64 i386 +ONLY_FOR_ARCHS-java = ${ONLY_FOR_ARCHS} CATEGORIES = graphics devel COMMENT-docs = OpenCV documentation and examples COMMENT-main = library of programming functions for real time cv COMMENT-java = Java bindings for OpenCV -V =2.4.12.3 +V =2.4.13 DISTNAME = opencv-$V PKGNAME-main = ${PKGNAME} PKGNAME-docs = opencv-docs-$V PKGNAME-java = opencv-java-$V -REVISION-docs =0 HOMEPAGE = http://www.opencv.org/ MAINTAINER = Rafael Sadowski-SHARED_LIBS = opencv_calib3d 1.1 #1.2 -SHARED_LIBS += opencv_contrib 1.1 #1.2 -SHARED_LIBS += opencv_core 1.1 #1.2 -SHARED_LIBS += opencv_features2d 1.1 #1.2 -SHARED_LIBS += opencv_flann1.1 #1.2 -SHARED_LIBS += opencv_gpu 1.1 #1.2 -SHARED_LIBS += opencv_highgui 1.1 #1.2 -SHARED_LIBS += opencv_imgproc 1.1 #1.2 -SHARED_LIBS += opencv_legacy 1.1 #1.2 -SHARED_LIBS += opencv_ml 1.1 #1.2 -SHARED_LIBS += opencv_nonfree 0.1 #0.2 -SHARED_LIBS += opencv_objdetect1.1 #1.2 -SHARED_LIBS += opencv_photo0.1 #0.2 -SHARED_LIBS += opencv_stitching0.1 #0.2 -SHARED_LIBS += opencv_superres 0.1 #0.2 -SHARED_LIBS += opencv_ts 0.1 #0.2 -SHARED_LIBS += opencv_video1.1 #1.2 -SHARED_LIBS += opencv_videostab0.1 #0.2 -SHARED_LIBS += opencv_ocl 0.1 #2.5 +SHARED_LIBS = opencv_calib3d 1.2 #1.3 +SHARED_LIBS += opencv_contrib 1.2 #1.3 +SHARED_LIBS += opencv_core 1.2 #1.3 +SHARED_LIBS += opencv_features2d 1.2 #1.3 +SHARED_LIBS += opencv_flann1.2 #1.3 +SHARED_LIBS += opencv_gpu 1.2 #1.3 +SHARED_LIBS += opencv_highgui 1.2 #1.3 +SHARED_LIBS += opencv_imgproc 1.2 #1.3 +SHARED_LIBS += opencv_legacy 1.2 #1.3 +SHARED_LIBS += opencv_ml 1.2 #1.3 +SHARED_LIBS += opencv_nonfree 0.2 #0.3 +SHARED_LIBS += opencv_objdetect1.2 #1.3 +SHARED_LIBS += opencv_photo0.2 #0.3 +SHARED_LIBS += opencv_stitching0.2 #0.3 +SHARED_LIBS += opencv_superres 0.2 #0.3 +SHARED_LIBS += opencv_ts 0.2 #0.3 +SHARED_LIBS += opencv_video1.2 #1.3 +SHARED_LIBS += opencv_videostab0.2 #0.3 +SHARED_LIBS += opencv_ocl 0.2 #2.6 GH_ACCOUNT = Itseez GH_PROJECT = opencv -GH_COMMIT =
NEW: devel/py-codestyle 2.0.0 (was pep8)
Hi, ok to import? This package used to be called pep8 but was renamed to pycodestyle to reduce confusion. -- Alexandr Shadchin py-codestyle.tgz Description: application/tar-gz
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: shadc...@cvs.openbsd.org2016/08/11 21:02:55 Modified files: devel/py-parsing: Makefile distinfo Log message: Update to py-parsing 2.1.7
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: j...@cvs.openbsd.org2016/08/11 20:44:36 Modified files: x11: Makefile x11/dzen2 : Makefile x11/dzen2/patches: patch-config_mk Log message: add an xft flavor to dzen2 ok maintainer
Re: The W^X situation: we need wxneeded
> > On Mon, Aug 08, 2016 at 09:33:50AM -0600, Theo de Raadt wrote: > >> > On 2016-08-07, Carlin Binghamwrote: > >> > > >> > > What should be done for ports like (just for example) calibre? > >> > > It uses python, and uses PyQtWebkit to pull in Qt5Webkit which maps WX > >> > > for its jit. Python itself doesn't need wxneeded but for calibre (and > >> > > possibly other ports) to work it will. > >> > > >> > I don't know how to handle this. The model of tying wxneeded > >> > permissions to an executable is ill suited for this. > >> > > >> > Maybe have both python and python-wx and use the latter for things > >> > like calibre? > >> > Do we have Perl module that violates W^X in the ports tree? > >> > > >> > Ideas? > >> > >> Just mark python, and move on. > >> > >> In a few months we'll be able to circle back and decide. > > This does the trick and allows me to run something using > > PyQtWebkit/Qt5Webkit: > > It would be nice if we could create wxneeded executables *after* link > time. It would then be easy to tweak ports that use > libs like PyQtWebkit/Qt5Webkit to point at eg > /usr/local/bin/python2.7-wxneeded. I don't like it. The purpose of wxneeded is to get us over this hump where the ecosystem has some software which requires W|X behaviour. By doing so, it will curate a list of software which has the bad practice. If all software which needs the option has the flag, there's your list. Eventually we can get request that all the software on this list gets fixed -- or choose with our feet to use software which gets it right -- and then we can delete the wxneeded flag. One by one. I am completely serious. I was serious when we made the stack-protector mandatory, and got the world to accept that. I was serious when we made ASLR mandatory, and got the world to accept that. I was serious lots of other times when we made the safe code path mandatory, and slowly dragged software on the wrong track towards the right path. Yes, it will take 1-2 decades, like the others. And that doesn't bother me. Check back in 20 years and tell me if I was right or wrong! In the meantime, let's not abuse a system-dependent temporary measure.
UPDATE+FIX: PyPy
Hi Ports, Armin, This diff makes PyPy build on OpenBSD once more, and in doing so updates to the latest version. I only started out to mark PyPy wxneeded, but it led to an odd situation with PyPy: * The bootstrap tarball no longer works, since the contained binary is not marked wxneeded, so we have to make a new bootstrap using CPython. * However, the build system itself (under CPython) is not W^X compliant either. The diff makes the PyPy build work under CPython without requiring CPython itself to be marked wxneeded (although I see another thread where people are discussing marking CPython anyway). It also marks the PyPy binary wxneeded. Since we have to have a new bootstrap, I figured I may aswell update PyPy to the latest version too. Armin, this is a minimal distribution patch that I am going to use for now until we can figure out this rmmap refactoring. It simply duplicates alloc_hinted thus avoiding the NonConstant problem. I'm still working on the correct refactoring separately. People, note that this does not make the PyPy JIT W^X compatible. Although I hope we can be there eventually, apparently this will not be a trivial task (it's going to be hard to know all the places where we need to flip the page protection flags, and when we get it wrong and get a crash, it's going to be hard to know where the crash came from). Anyway, this is the first step. OK anyone? Bootstrap tarball is here (please could you host this Stuart?): http://theunixzoo.co.uk/random/pypy-bootstrap-amd64-5.3.1.tar.xz Index: Makefile === RCS file: /cvs/ports/lang/pypy/Makefile,v retrieving revision 1.24 diff -u -p -r1.24 Makefile --- Makefile13 Jun 2016 08:44:17 - 1.24 +++ Makefile11 Aug 2016 20:19:10 - @@ -18,7 +18,7 @@ COMMENT = fast implementation of the Py # You can use the no_bootstrap FLAVOR to make a new bootstrap. This uses # CPython to build PyPy instead. Note this is slower. -V =5.3.0 +V =5.3.1 BOOTSTRAP_V = ${V} DISTNAME = pypy2-v${V}-src PKGNAME = pypy-${V} @@ -95,7 +95,8 @@ do-build: pypy --jit loop_longevity=300 \ ../../rpython/bin/rpython --source --opt=jit .endif - cd ${WRKDIR}/usession/testing_1 && ${SETENV} ${MAKE_ENV} ${MAKE_PROGRAM} + cd ${WRKDIR}/usession/testing_1 && ${SETENV} ${MAKE_ENV} \ + ${MAKE_PROGRAM} LDFLAGS_LINK="-Wl,-z,wxneeded -pthread" PYPY_INST_LIB_DIR = ${PREFIX}/pypy/lib PYPY_LD_LIBRARY_PATH = LD_LIBRARY_PATH=${PYPY_INST_LIB_DIR} Index: distinfo === RCS file: /cvs/ports/lang/pypy/distinfo,v retrieving revision 1.16 diff -u -p -r1.16 distinfo --- distinfo13 Jun 2016 08:44:17 - 1.16 +++ distinfo11 Aug 2016 20:19:10 - @@ -1,4 +1,4 @@ -SHA256 (pypy/pypy-bootstrap-amd64-5.3.0.tar.xz) = edy2uGOHyc3D9NndGeSZ86szEvpiJmFSQdAC/wtLBpg= -SHA256 (pypy/pypy2-v5.3.0-src.tar.bz2) = QULrj0A4ELyIpJEXkrtaUC4VLflYBuM+aQUMgozRYNU= -SIZE (pypy/pypy-bootstrap-amd64-5.3.0.tar.xz) = 21976516 -SIZE (pypy/pypy2-v5.3.0-src.tar.bz2) = 17361429 +SHA256 (pypy/pypy-bootstrap-amd64-5.3.1.tar.xz) = J8lRGufV48uBbJu302zRSAf3RdXxjEokruugShyOJ0U= +SHA256 (pypy/pypy2-v5.3.1-src.tar.bz2) = MaUrq1hKvzoPDe/Rv5opEx2rCN9DiF5+7d/H3Jtxg24= +SIZE (pypy/pypy-bootstrap-amd64-5.3.1.tar.xz) = 22269372 +SIZE (pypy/pypy2-v5.3.1-src.tar.bz2) = 17361760 Index: patches/patch-rpython_jit_backend_x86_detect_feature_py === RCS file: patches/patch-rpython_jit_backend_x86_detect_feature_py diff -N patches/patch-rpython_jit_backend_x86_detect_feature_py --- /dev/null 1 Jan 1970 00:00:00 - +++ patches/patch-rpython_jit_backend_x86_detect_feature_py 11 Aug 2016 20:19:10 - @@ -0,0 +1,30 @@ +$OpenBSD$ + +Make the CPython bootstrap W^X compatible. + +--- rpython/jit/backend/x86/detect_feature.py.orig Tue Jun 14 08:46:04 2016 rpython/jit/backend/x86/detect_feature.py Thu Aug 11 08:49:22 2016 +@@ -1,17 +1,20 @@ + import sys + import struct + from rpython.rtyper.lltypesystem import lltype, rffi +-from rpython.rlib.rmmap import alloc, free ++from rpython.rlib.rmmap import alloc, free, set_pages_executable + ++CPU_INFO_SZ = 4096 ++ + def cpu_info(instr): +-data = alloc(4096) ++data = alloc(CPU_INFO_SZ, no_exec=True) + pos = 0 + for c in instr: + data[pos] = c + pos += 1 + fnptr = rffi.cast(lltype.Ptr(lltype.FuncType([], lltype.Signed)), data) ++set_pages_executable(data, CPU_INFO_SZ) + code = fnptr() +-free(data, 4096) ++free(data, CPU_INFO_SZ) + return code + + def detect_sse2(): Index: patches/patch-rpython_rlib_rmmap_py === RCS file: patches/patch-rpython_rlib_rmmap_py diff -N patches/patch-rpython_rlib_rmmap_py ---
Re: UPDATE: games/eduke32
On Wed, Aug 10, 2016 at 11:30:19PM -0700, Ryan Freeman wrote: > Hey, > > Trying this one again, updates eduke32 to a much newer release, moving to > SDL2 and seems to run very well still on current/amd64. If it helps at all, my testing seems to indicate there is no need for wxneeded. > > Original post(s): http://marc.info/?l=openbsd-ports=146784932621112=2 > One independent test: http://marc.info/?l=openbsd-ports=146786520124196=2 > > Copying the details for ease: > > > Here is a patch against current ports cvs triggered by Daniel's update. > > WANTLIB and LIB_DEPENDS have been regenerated. the do-install paths have > > been corrected, I removed the dos2unix stuff as I don't think it is that > > big a deal in the end, and it matches upstream. > > > > The datafile path substitutions have been reduced to one file (yay), and > > perl has been switched out in lieu of sed. > > > > Unfortunately the Makefile.common has some mean tricks to really break our > > ability to feed CC/CXX, so I did some hack'n'slash patching of it. I am > > not sure that this is best, I have employed their recommended > > PACKAGE_REPOSITORY=1 setting to 'not completely gut all their > > optimizations'.. > > but I'd love it if anyone can confirm its okay :-) looks to just use -O2. > > > > Turning on aggressive debugging things like DEBUGANYWAY kinda broke mouse > > input making it really hard to play, so avoiding that path. > > > > Lastly I think we need to use ports gcc now. > > > > works good here on current amd64 > > Anyone willing to ok/commit? > > Cheers! > -ryan > > > ? eduke32.diff > Index: Makefile > === > RCS file: /cvs/ports/games/eduke32/Makefile,v > retrieving revision 1.7 > diff -u -p -r1.7 Makefile > --- Makefile 22 Oct 2014 13:55:07 - 1.7 > +++ Makefile 19 Jun 2016 06:43:09 - > @@ -1,12 +1,11 @@ > # $OpenBSD: Makefile,v 1.7 2014/10/22 13:55:07 jasper Exp $ > > COMMENT =Enhanced Duke Nukem 3D engine > -RDATE = 20120301 > -RTAG = 2394 > +RDATE = 20160610 > +RTAG = 5775 > DISTNAME = eduke32_src_${RDATE}-${RTAG} > PKGNAME =eduke32-2.0.0.${RTAG} > -REVISION = 2 > -EXTRACT_SUFX = .tar.bz2 > +EXTRACT_SUFX = .tar.xz > CATEGORIES = games x11 > > HOMEPAGE = http://www.eduke32.com/ > @@ -25,12 +24,15 @@ BUILD_DEPENDS = archivers/unzip \ > BUILD_DEPENDS += devel/nasm > .endif > > -WANTLIB =SDL c m pthread stdc++ SDL_mixer>=3 vorbisfile vpx > +WANTLIB += c m ogg png pthread vorbis vorbisfile vpx z > +WANTLIB += SDL2 SDL2_mixer FLAC execinfo > > LIB_DEPENDS =audio/libvorbis \ > - multimedia/libvpx \ > - devel/sdl \ > - devel/sdl-mixer > + audio/flac \ > + devel/sdl2 \ > + devel/sdl2-mixer \ > + devel/libexecinfo \ > + multimedia/libvpx > > # dlopen()'d; shows as Extra in check-lib-depends but it is used > LIB_DEPENDS += x11/gtk+2 > @@ -38,31 +40,23 @@ WANTLIB +=gtk-x11-2.0 > > RUN_DEPENDS =games/duke3ddata > > -MASTER_SITES = > http://dukeworld.duke4.net/eduke32/synthesis/old/${RDATE}-${RTAG}/ \ > - http://dukeworld.duke4.net/eduke32/synthesis/${RDATE}-${RTAG}/ > +MASTER_SITES = > http://dukeworld.duke4.net/eduke32/synthesis/${RDATE}-${RTAG}/ > + > +MODULES =gcc4 > +MODGCC4_LANGS = c c++ > +MODGCC4_ARCHS = * > > # make sure we can see compilation output, respect CC/CXX, don't strip! > -MAKE_FLAGS = PRETTY_OUTPUT=0 CC="${CC}" CXX="${CXX}" DEBUGANYWAY=1 > +MAKE_ENV = CC="${CC}" CXX="${CXX}" PRETTY_OUTPUT=0 PACKAGE_REPOSITORY=1 > USE_GMAKE = Yes > NO_TEST =Yes > > -WRKDIST =${WRKDIR}/eduke32_${RDATE}-${RTAG}/ > +WRKDIST =${WRKDIR}/eduke32_${RDATE}-${RTAG} > > post-extract: > # check for data files in correct location > - @perl -pi -e "s,/usr/share/games/eduke32,${TRUEPREFIX}/share/duke3d," \ > - ${WRKSRC}/source/astub.c > - @perl -pi -e "s,/usr/share/games/eduke32,${TRUEPREFIX}/share/duke3d," \ > - ${WRKSRC}/source/game.c > - @perl -pi -e "s,./eduke32,eduke32," ${WRKSRC}/build/include/editor.h > - # make plaintext files less annoying to read... > - @perl -MExtUtils::Command -e dos2unix \ > - ${WRKSRC}/tiles.cfg \ > - ${WRKSRC}/names.h \ > - ${WRKSRC}/GNU.TXT \ > - ${WRKSRC}/buildlic.txt \ > - ${WRKSRC}/samples/*.sample \ > - ${WRKSRC}/samples/*.txt > + @sed -i "s,/usr/share/games/eduke32,${TRUEPREFIX}/share/duke3d," \ > + ${WRKSRC}/source/common.c > > do-install: > ${INSTALL_PROGRAM} ${WRKBUILD}/eduke32 ${PREFIX}/bin > @@ -71,17 +65,25 @@ do-install: > ${INSTALL_DATA_DIR} ${PREFIX}/share/doc/eduke32 > ${INSTALL_DATA_DIR} ${PREFIX}/share/examples/eduke32 > # mapster32 help files > -
Re: The W^X situation: we need wxneeded
Jasper Lievisse Adriaansewrites: > On Mon, Aug 08, 2016 at 09:33:50AM -0600, Theo de Raadt wrote: >> > On 2016-08-07, Carlin Bingham wrote: >> > >> > > What should be done for ports like (just for example) calibre? >> > > It uses python, and uses PyQtWebkit to pull in Qt5Webkit which maps WX >> > > for its jit. Python itself doesn't need wxneeded but for calibre (and >> > > possibly other ports) to work it will. >> > >> > I don't know how to handle this. The model of tying wxneeded >> > permissions to an executable is ill suited for this. >> > >> > Maybe have both python and python-wx and use the latter for things >> > like calibre? >> > Do we have Perl module that violates W^X in the ports tree? >> > >> > Ideas? >> >> Just mark python, and move on. >> >> In a few months we'll be able to circle back and decide. > This does the trick and allows me to run something using PyQtWebkit/Qt5Webkit: It would be nice if we could create wxneeded executables *after* link time. It would then be easy to tweak ports that use libs like PyQtWebkit/Qt5Webkit to point at eg /usr/local/bin/python2.7-wxneeded. In the meantime, I think this is the way to go. ok jca@ -- jca | PGP: 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
NEW devel/p5-Test-MockRandom
Hi, I need devel/p5-Test-MockRandom as a test dependency. Comment: replace random number generation with non-random Description: This perhaps ridiculous-seeming module was created to test routines that manipulate random numbers by providing a known output from rand. Given a list of seeds with srand, it will return each in turn. After seeded random numbers are exhausted, it will always return 0. Seed numbers must be of a form that meets the expected output from rand as called with no arguments -- i.e. they must be between 0 (inclusive) and 1 (exclusive). In order to facilitate generating and testing a nearly-one number, this module exports the function oneish, which returns a number just fractionally less than one. ok to import? bluhm p5-Test-MockRandom.tgz Description: application/tar-gz
Re: NEW: sysutils/bitrot
Michael McConville wrote: > Ray Lai wrote: > > Detects bit rotten files on the hard drive to save your precious > > photo and music collection from slow decay. > > > > Don't let your bits rot, ffs. > > Attached is an iteration with a few fixes: > > * update to 0.8.0 > * specify that it's released under the MIT license > * tweak $COMMENT and DESCR > > Upstream seems to have stopped making releases on GitHub - I just > emailed them about that. For now, I'll stick to the commit approach. Ray doesn't have time to work on this, so I adopted it. Attached is a new version that fetches from PyPI, since upstream is exclusively using that for new releases. I've never worked with the PyPI make infrastructure before, so that's worth checking. Thanks for your time, Mike bitrot.tar.gz Description: application/tar-gz
Re: The W^X situation: we need wxneeded
On Mon, Aug 08, 2016 at 09:33:50AM -0600, Theo de Raadt wrote: > > On 2016-08-07, Carlin Binghamwrote: > > > > > What should be done for ports like (just for example) calibre? > > > It uses python, and uses PyQtWebkit to pull in Qt5Webkit which maps WX > > > for its jit. Python itself doesn't need wxneeded but for calibre (and > > > possibly other ports) to work it will. > > > > I don't know how to handle this. The model of tying wxneeded > > permissions to an executable is ill suited for this. > > > > Maybe have both python and python-wx and use the latter for things > > like calibre? > > Do we have Perl module that violates W^X in the ports tree? > > > > Ideas? > > Just mark python, and move on. > > In a few months we'll be able to circle back and decide. This does the trick and allows me to run something using PyQtWebkit/Qt5Webkit: Index: Makefile.inc === RCS file: /cvs/ports/lang/python/Makefile.inc,v retrieving revision 1.115 diff -u -p -r1.115 Makefile.inc --- Makefile.inc11 May 2016 22:43:49 - 1.115 +++ Makefile.inc11 Aug 2016 15:24:15 - @@ -123,7 +123,7 @@ CONFIGURE_ENV += OPT='${CFLAGS} -fPIC' C LOCALBASE=${LOCALBASE} X11BASE=${X11BASE} MAKE_ENV +=LOCALBASE=${LOCALBASE} X11BASE=${X11BASE} MAKE_FLAGS += LD_LIBRARY_PATH=${WRKSRC} PATH="${WRKDIST}:${PORTPATH}" -MAKE_FLAGS += LDFLAGS='-L${WRKSRC}' +MAKE_FLAGS += LDFLAGS='-L${WRKSRC} -Wl,-z,wxneeded' FAKE_FLAGS += RANLIB=: ALL_TARGET = all ./Lib/plat-openbsd6 Index: 2.7/Makefile === RCS file: /cvs/ports/lang/python/2.7/Makefile,v retrieving revision 1.46 diff -u -p -r1.46 Makefile --- 2.7/Makefile3 Jul 2016 18:16:00 - 1.46 +++ 2.7/Makefile11 Aug 2016 15:24:15 - @@ -7,6 +7,7 @@ VERSION = 2.7 PATCHLEVEL = .12 +REVISION = 0 SHARED_LIBS = python2.7 0.0 VERSION_SPEC = >=2.7,<2.8 Index: 3.4/Makefile === RCS file: /cvs/ports/lang/python/3.4/Makefile,v retrieving revision 1.12 diff -u -p -r1.12 Makefile --- 3.4/Makefile3 Jul 2016 18:16:00 - 1.12 +++ 3.4/Makefile11 Aug 2016 15:24:15 - @@ -7,6 +7,7 @@ VERSION = 3.4 PATCHLEVEL = .5 +REVISION = 0 SHARED_LIBS = python3.4m 1.0 VERSION_SPEC = >=3.4,<3.5 Index: 3.5/Makefile === RCS file: /cvs/ports/lang/python/3.5/Makefile,v retrieving revision 1.9 diff -u -p -r1.9 Makefile --- 3.5/Makefile3 Jul 2016 18:16:00 - 1.9 +++ 3.5/Makefile11 Aug 2016 15:24:15 - @@ -7,6 +7,7 @@ VERSION = 3.5 PATCHLEVEL = .2 +REVISION = 0 SHARED_LIBS = python3.5m 0.0 VERSION_SPEC = >=3.5,<3.6 -- jasper
Re: Dropping lang/libv8
On Thu, Aug 11, 2016 at 08:09:29AM -0600, Aaron Bieber wrote: > Hola, > > I am dropping MAINTAINER for lang/libv8, haven't really touched it > since I started building lang/node with builtin v8. > > If anyone wants to take it over, I have attached an updated version > that robert@ did some work on. I can also continue to host the dist > files if need be! > > This attached version is probably well out of date by now. When picking > a version to update to, one needs to take into account the versions of > v8 that each dep build against. Last I checked the-ruby-racer was > holding things back. > > Also attached is a diff to clean the Makefile up a bit and add > "-Wl,-z,wxneeded" woops. attached now! > > Cheers, > Aaron > Index: Makefile === RCS file: /cvs/ports/lang/libv8/Makefile,v retrieving revision 1.7 diff -u -p -r1.7 Makefile --- Makefile18 Mar 2016 22:14:47 - 1.7 +++ Makefile11 Aug 2016 14:02:41 - @@ -5,13 +5,13 @@ ONLY_FOR_ARCHS= amd64 i386 COMMENT= V8 JavaScript for clients and servers V8_VERSION=3.12.19 -REVISION= 1 +REVISION= 2 EPOCH= 0 -V8_ARCH= ${MACHINE_ARCH:S/amd64/x64/:S/i386/ia32/} -V8_RELEASE= release +V8_ARCH= ${MACHINE_ARCH:S/amd64/x64/:S/i386/ia32/} +V8_RELEASE=release DISTNAME= v8-${V8_VERSION} -PKGNAME= lib${DISTNAME} +PKGNAME= lib${DISTNAME} CATEGORIES=lang @@ -19,37 +19,36 @@ SHARED_LIBS=v80.0 HOMEPAGE= https://bugs.chromium.org/p/v8/ -MAINTAINER=Aaron Bieber- # BSD PERMIT_PACKAGE_CDROM= Yes MASTER_SITES= http://deftly.net/ -BUILD_DEPENDS= devel/gyp +BUILD_DEPENDS= devel/gyp -LIB_DEPENDS= devel/libexecinfo +LIB_DEPENDS= devel/libexecinfo WANTLIB += c execinfo m pthread ssl stdc++ USE_GMAKE= Yes -CONFIGURE_STYLE= python.port.mk -MODULES= lang/python +CONFIGURE_STYLE= python.port.mk +MODULES= lang/python SUBST_VARS+= WRKDIR SUBST_VARS+= V8_ARCH -MAKE_FLAGS+= library=shared ARCHES=${V8_ARCH} MODES=${V8_RELEASE} \ -DEFAULT_ARCHES=${V8_ARCH} +MAKE_FLAGS+= library=shared ARCHES=${V8_ARCH} MODES=${V8_RELEASE} \ + DEFAULT_ARCHES=${V8_ARCH} MAKE_ENV+= V=1 CXX=c++ CCFLAGS+="${CFLAGS}" \ - CXXFLAGS="${CXXFLAGS} -I${WRKSRC}/src -I${LOCALBASE}/include" \ - PATH=${WRKSRC}/bin:${PATH} + CXXFLAGS="${CXXFLAGS} -I${WRKSRC}/src -I${LOCALBASE}/include" \ + LDFLAGS="-L${LOCALBASE}/lib -L${X11BASE}/lib -Wl,-z,wxneeded" \ + PATH=${WRKSRC}/bin:${PATH} NO_CCACHE= Yes -TEST_TARGET= check +TEST_TARGET= check pre-configure: @# Bad practice, but prevents a whole stack of patches.
Dropping lang/libv8
Hola, I am dropping MAINTAINER for lang/libv8, haven't really touched it since I started building lang/node with builtin v8. If anyone wants to take it over, I have attached an updated version that robert@ did some work on. I can also continue to host the dist files if need be! This attached version is probably well out of date by now. When picking a version to update to, one needs to take into account the versions of v8 that each dep build against. Last I checked the-ruby-racer was holding things back. Also attached is a diff to clean the Makefile up a bit and add "-Wl,-z,wxneeded" Cheers, Aaron libv8-4.9.385.28.tar.gz Description: application/tar-gz
fluxter on Fluxbox: Cannot connect to window manager
Hello, launching fluxter with Fluxbox: ┌──[just22@poseidon]-[0]-[✓]-[tmp] └─› fluxter Local fluxbox config file: "/home/just22/.fluxbox/fluxter.bb" Default fluxbox config file: "/usr/local/share/fluxbox/fluxter.bb" ┌──[just22@poseidon]-[0]-[✓]-[tmp] └─› fluxter -n Local fluxter config file: "/home/just22/.fluxbox/fluxter.nobb" Global fluxter config file: "/usr/local/share/fluxbox/fluxter.nobb" Cannot connect to window manager Anybody had more luck with this port? Or it has been proven working with other window managers? If not, probably it may be removed (the project is currently unmaintained). All the best -- Alessandro DE LAURENZIS [mailto:jus...@atlantide.t28.net] LinkedIn: http://it.linkedin.com/in/delaurenzis
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: j...@cvs.openbsd.org2016/08/11 06:04:17 Modified files: net/knot : Makefile distinfo Log message: Update to knot-1.6.8 from Pierre Emeriaud
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 05:41:42 Modified files: sysutils/ruby-puppet/4: Makefile distinfo Log message: update to puppet-4.5.3
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: j...@cvs.openbsd.org2016/08/11 05:34:13 Modified files: lang/smlnj : Makefile lang/smlnj/patches: patch-base_runtime_objs_mk_ppc-openbsd patch-base_runtime_objs_mk_x86-openbsd Log message: wxneeded
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 05:29:36 Modified files: games/solarus/zsdx: Makefile distinfo games/solarus/zsdx/patches: patch-CMakeLists_txt Log message: update to zsdx-1.11.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 05:29:15 Modified files: games/solarus/zsxd: Makefile distinfo games/solarus/zsxd/patches: patch-CMakeLists_txt Log message: update to zsxd-1.11.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 05:29:10 Modified files: games/solarus/roth: Makefile distinfo games/solarus/roth/patches: patch-CMakeLists_txt Log message: update to zelda_roth_se-1.1.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 05:28:42 Modified files: games/solarus/solarus: Makefile distinfo games/solarus/solarus/pkg: PLIST Log message: update to solarus-1.5.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 05:01:45 Modified files: sysutils/rofi : Makefile distinfo Log message: update to rofi-1.1.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 04:59:08 Modified files: www/kibana : Makefile distinfo www/kibana/pkg : PLIST Log message: - update to kibana-4.5.4 from Pavel Korovin (MAINTAINER)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 04:57:30 Modified files: textproc/elasticsearch: Makefile distinfo Log message: - update to elasticsearch-2.3.5 from Pavel Korovin
UPDATE: devel/py-pexpect 4.2.0
Hi, This diff updates py-pexpect to the latest release. Tested on amd64. Comments ? OK ? -- Alexandr Shadchin Index: Makefile === RCS file: /cvs/ports/devel/py-pexpect/Makefile,v retrieving revision 1.24 diff -u -p -r1.24 Makefile --- Makefile19 Jun 2016 12:17:06 - 1.24 +++ Makefile11 Aug 2016 10:21:24 - @@ -2,7 +2,7 @@ COMMENT= pure Python Expect-like module -MODPY_EGG_VERSION= 4.1.0 +MODPY_EGG_VERSION= 4.2.0 DISTNAME= pexpect-${MODPY_EGG_VERSION} PKGNAME= py-${DISTNAME} CATEGORIES=devel Index: distinfo === RCS file: /cvs/ports/devel/py-pexpect/distinfo,v retrieving revision 1.9 diff -u -p -r1.9 distinfo --- distinfo19 Jun 2016 12:17:06 - 1.9 +++ distinfo11 Aug 2016 10:21:24 - @@ -1,2 +1,2 @@ -SHA256 (pexpect-4.1.0.tar.gz) = w4HGDxmHNVtl348Ion9CiDGRTIqBCRvRd4rDNvovJ+c= -SIZE (pexpect-4.1.0.tar.gz) = 140294 +SHA256 (pexpect-4.2.0.tar.gz) = v2gWuMyNMBpJnnrfM4gos5vHVI62Tb7U3UEO2T2V+FM= +SIZE (pexpect-4.2.0.tar.gz) = 147948 Index: pkg/PLIST === RCS file: /cvs/ports/devel/py-pexpect/pkg/PLIST,v retrieving revision 1.9 diff -u -p -r1.9 PLIST --- pkg/PLIST 19 Jun 2016 12:17:06 - 1.9 +++ pkg/PLIST 11 Aug 2016 10:21:24 - @@ -118,5 +118,6 @@ share/examples/${MODPY_PY_PREFIX}pexpect share/examples/${MODPY_PY_PREFIX}pexpect/script.py share/examples/${MODPY_PY_PREFIX}pexpect/ssh_tunnel.py share/examples/${MODPY_PY_PREFIX}pexpect/table_test.html +share/examples/${MODPY_PY_PREFIX}pexpect/telnet.py share/examples/${MODPY_PY_PREFIX}pexpect/topip.py share/examples/${MODPY_PY_PREFIX}pexpect/uptime.py
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: shadc...@cvs.openbsd.org2016/08/11 04:20:36 Modified files: devel/py-parsing: Makefile distinfo devel/py-parsing/pkg: PLIST Log message: Update to py-parsing 2.1.6
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: shadc...@cvs.openbsd.org2016/08/11 03:50:05 Modified files: devel/py-test-cov: Makefile distinfo Log message: Minor update to py-test-cov 2.3.1
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 03:12:22 Modified files: graphics/birdfont: Makefile distinfo Added files: graphics/birdfont/patches: patch-configure patch-dodo_py Log message: - update to birdfont-2.16.5
Re: Security update textproc/wkhtmltopdf-0.12.3.2
On Thu, Aug 11, 2016 at 03:48:59PM +0800, Tinker wrote: > On 2016-08-11 15:38, Tinker wrote: > >On 2016-07-28 22:54, Frank Groeneveld wrote: > >>On Thu, Jul 28, 2016 at 12:54:52PM +0100, Stuart Henderson wrote: > >>>On 2016/07/28 01:02, Tinker wrote: > On 2016-07-27 16:36, Frank Groeneveld wrote: > > Attached patch updates wkhtmltopdf to the latest release. Some > > important > > security fixes (OpenSSL updates) were added .. > > Can't it use the system's LibreSSL version?? > > >>> > >>>I'm pretty sure it is, otherwise we wouldn't have needed the > >>>OPENSSL_NO_SSL3 > >>>patches. > >> > >>Yes, it doesn't use the bundled OpenSSl version, sorry for the > >>confusion. It seems [..] > > (Wait, by "does not use the bundled OpenSSL version", you mean that the > Wkhtmltopdf OpenBSD port not uses the *OS-*bundled *LibreSSL* library > right?) > With bundled I mean the library that comes in the wkhtmltopdf "bundle". As you can see we need patches to let the build know that the base LibreSSL doesn't support SSLv3, which leads me and Stuart to think that it uses the LibreSSL from base. I see Stuart has committed this update already. Thanks Stuart! Frank
Re: Haskell wxneeded
Hi, On 2016-08-10 22:04, Matthias Kilian wrote: > But I want to move forward with ghc (because ghc-8 works much better > on OpenBSD as Karel Gardas wrote some time ago). Not necessarily > right now, but at least during this release cycle (i.e. before > OpenBSD 6.1). Would it be possible to have multiple versions of GHC like the way it is done with Erlang or would that be too much work? In case that having multiple versions is not viable, I'll switch to Stackage nightly till GHC 8 becomes LTS. Anyway, thanks for your work on this port. Kind regards, Martijn Rijkeboer
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: bent...@cvs.openbsd.org 2016/08/11 02:18:02 Modified files: fonts : Makefile Log message: +gentium
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: bent...@cvs.openbsd.org 2016/08/11 02:17:30 Log message: Import gentium-5.000. Gentium is a typeface family designed to enable the diverse ethnic groups around the world who use the Latin, Cyrillic and Greek scripts to produce readable, high-quality publications. It supports a wide range of Latin- and Cyrillic-based alphabets. ok benoit@ czarkoff@ Status: Vendor Tag: bentley Release Tags: bentley_20160811 N ports/fonts/gentium/Makefile N ports/fonts/gentium/distinfo N ports/fonts/gentium/pkg/DESCR N ports/fonts/gentium/pkg/PLIST No conflicts created by this import
Re: [NEW] graphics/py-pdfrw
On Wed, 10 Aug 2016 17:17:23 -0400, George Rosamondwrote: > New port for consideration, py-pdfrw. Tested a bit with python2.7. Just a quick look, the PLIST is wrong because of the python3 flavor, see lang/python in PORT-MODULES(5). Any reason to not to enable test? Minor nits: can you vertically align the Makefile? Cheers, Daniel
Re: Security update textproc/wkhtmltopdf-0.12.3.2
On 2016-08-11 15:38, Tinker wrote: On 2016-07-28 22:54, Frank Groeneveld wrote: On Thu, Jul 28, 2016 at 12:54:52PM +0100, Stuart Henderson wrote: On 2016/07/28 01:02, Tinker wrote: > On 2016-07-27 16:36, Frank Groeneveld wrote: > > Attached patch updates wkhtmltopdf to the latest release. Some important > > security fixes (OpenSSL updates) were added .. > > Can't it use the system's LibreSSL version?? > I'm pretty sure it is, otherwise we wouldn't have needed the OPENSSL_NO_SSL3 patches. Yes, it doesn't use the bundled OpenSSl version, sorry for the confusion. It seems [..] (Wait, by "does not use the bundled OpenSSL version", you mean that the Wkhtmltopdf OpenBSD port not uses the *OS-*bundled *LibreSSL* library right?)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: bent...@cvs.openbsd.org 2016/08/11 01:42:27 Modified files: emulators/gambatte: Makefile Log message: Fix gambatte PKGNAME.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: bent...@cvs.openbsd.org 2016/08/11 01:43:20 Modified files: emulators/mupen64plus/ui-console: Makefile Added files: emulators/mupen64plus/ui-console/patches: patch-projects_unix_Makefile Log message: Mark mupen64plus wxneeded.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: bent...@cvs.openbsd.org 2016/08/11 01:42:52 Modified files: emulators/desmume: Makefile Log message: Mark desmume wxneeded.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 01:39:29 Modified files: x11/gnome/online-accounts: Makefile distinfo Log message: update to gnome-online-accounts-3.20.3
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 01:38:45 Modified files: databases/ruby-hiera3: Makefile distinfo Log message: update to hiera3-3.2.1
Re: Security update textproc/wkhtmltopdf-0.12.3.2
On 2016-07-28 22:54, Frank Groeneveld wrote: On Thu, Jul 28, 2016 at 12:54:52PM +0100, Stuart Henderson wrote: On 2016/07/28 01:02, Tinker wrote: > On 2016-07-27 16:36, Frank Groeneveld wrote: > > Attached patch updates wkhtmltopdf to the latest release. Some important > > security fixes (OpenSSL updates) were added .. > > Can't it use the system's LibreSSL version?? > I'm pretty sure it is, otherwise we wouldn't have needed the OPENSSL_NO_SSL3 patches. Yes, it doesn't use the bundled OpenSSl version, sorry for the confusion. It seems [..] I understand why Wkhtmltopdf builds its own QT: it needs cutom formatting patches and the like, that the ordinary QT not has. Wkhtmltopdf bundling its own OpenSSL sounds not only useless (bc Wkhtmltopdf does not rely on any custom functionality via patches, right?), but also dangerous? Afterall OpenSSL is one of the nastiest open-source libraries in widespread use. Could Wkhtmltopdf be made to use the OS-bundled LibreSSL instead, do you see any conceptual problems about doing that?
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 01:37:41 Modified files: x11/gnome/calculator: Makefile distinfo Log message: update to gnome-calculator-3.20.2
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/08/11 01:34:30 Modified files: devel/json-glib: Makefile distinfo devel/json-glib/pkg: PLIST Log message: update to json-glib-1.2.2
Re: UPDATE: security/lastpass-cli 1.0.0
On Tue 02/08/2016 11:55, Björn Ketelaars wrote: > This diff updates security/lastpass-cli to the latest release. Changelog: > https://github.com/lastpass/lastpass-cli/releases > > OK? Ping -- Björn Ketelaars GPG key: 0x4F0E5F21 diff --git security/lastpass-cli/Makefile security/lastpass-cli/Makefile index f56a3ea..b0e53bd 100644 --- security/lastpass-cli/Makefile +++ security/lastpass-cli/Makefile @@ -4,7 +4,7 @@ COMMENT = LastPass command line interface tool GH_ACCOUNT = lastpass GH_PROJECT = lastpass-cli -GH_TAGNAME = v0.9.0 +GH_TAGNAME = v1.0.0 CATEGORIES = security MAINTAINER = Bjorn Ketelaarsdiff --git security/lastpass-cli/distinfo security/lastpass-cli/distinfo index ecf9147..d80a8fa 100644 --- security/lastpass-cli/distinfo +++ security/lastpass-cli/distinfo @@ -1,2 +1,2 @@ -SHA256 (lastpass-cli-0.9.0.tar.gz) = 5zFPDf7shq3XwZwFPuNLt6F25GLnFyfSs0VIHS0TaAA= -SIZE (lastpass-cli-0.9.0.tar.gz) = 84433 +SHA256 (lastpass-cli-1.0.0.tar.gz) = QglsC9OXKw6cyc7zL78UHkewS54jh/s6vosQXhNftB4= +SIZE (lastpass-cli-1.0.0.tar.gz) = 92311 diff --git security/lastpass-cli/patches/patch-Makefile security/lastpass-cli/patches/patch-Makefile index 80dab4a..7732b17 100644 --- security/lastpass-cli/patches/patch-Makefile +++ security/lastpass-cli/patches/patch-Makefile @@ -1,7 +1,7 @@ $OpenBSD: patch-Makefile,v 1.1.1.1 2015/09/02 16:10:06 sthen Exp $ Makefile.orig Tue Aug 11 19:42:03 2015 -+++ Makefile Thu Aug 27 17:37:15 2015 -@@ -6,7 +6,7 @@ MANDIR ?= $(PREFIX)/share/man +--- Makefile.orig Wed Jul 20 14:46:49 2016 Makefile Tue Aug 2 11:45:58 2016 +@@ -7,7 +7,7 @@ COMPDIR ?= $(shell pkg-config --variable=completionsdi CFLAGS ?= -O3 -march=native -fomit-frame-pointer -pipe CFLAGS += -std=gnu99 -D_GNU_SOURCE diff --git security/lastpass-cli/patches/patch-upload-queue_c security/lastpass-cli/patches/patch-upload-queue_c index 447b8af..d2b9f24 100644 --- security/lastpass-cli/patches/patch-upload-queue_c +++ security/lastpass-cli/patches/patch-upload-queue_c @@ -2,9 +2,9 @@ $OpenBSD: patch-upload-queue_c,v 1.4 2016/04/17 09:16:43 landry Exp $ cope with 64-bit time_t on 32-bit arches upload-queue.c.origWed Mar 2 23:03:40 2016 -+++ upload-queue.c Thu Mar 17 06:08:05 2016 -@@ -128,7 +128,7 @@ static void upload_queue_write_entry(const char *entry +--- upload-queue.c.origWed Jul 20 14:46:49 2016 upload-queue.c Tue Aug 2 11:45:58 2016 +@@ -86,7 +86,7 @@ static void upload_queue_write_entry(const char *entry for (serial = 0; serial < ULONG_MAX; ++serial) { free(name);
UPDATE: games/eduke32
Hey, Trying this one again, updates eduke32 to a much newer release, moving to SDL2 and seems to run very well still on current/amd64. Original post(s): http://marc.info/?l=openbsd-ports=146784932621112=2 One independent test: http://marc.info/?l=openbsd-ports=146786520124196=2 Copying the details for ease: > Here is a patch against current ports cvs triggered by Daniel's update. > WANTLIB and LIB_DEPENDS have been regenerated. the do-install paths have > been corrected, I removed the dos2unix stuff as I don't think it is that > big a deal in the end, and it matches upstream. > > The datafile path substitutions have been reduced to one file (yay), and > perl has been switched out in lieu of sed. > > Unfortunately the Makefile.common has some mean tricks to really break our > ability to feed CC/CXX, so I did some hack'n'slash patching of it. I am > not sure that this is best, I have employed their recommended > PACKAGE_REPOSITORY=1 setting to 'not completely gut all their optimizations'.. > but I'd love it if anyone can confirm its okay :-) looks to just use -O2. > > Turning on aggressive debugging things like DEBUGANYWAY kinda broke mouse > input making it really hard to play, so avoiding that path. > > Lastly I think we need to use ports gcc now. > > works good here on current amd64 Anyone willing to ok/commit? Cheers! -ryan ? eduke32.diff Index: Makefile === RCS file: /cvs/ports/games/eduke32/Makefile,v retrieving revision 1.7 diff -u -p -r1.7 Makefile --- Makefile22 Oct 2014 13:55:07 - 1.7 +++ Makefile19 Jun 2016 06:43:09 - @@ -1,12 +1,11 @@ # $OpenBSD: Makefile,v 1.7 2014/10/22 13:55:07 jasper Exp $ COMMENT = Enhanced Duke Nukem 3D engine -RDATE =20120301 -RTAG = 2394 +RDATE =20160610 +RTAG = 5775 DISTNAME = eduke32_src_${RDATE}-${RTAG} PKGNAME = eduke32-2.0.0.${RTAG} -REVISION = 2 -EXTRACT_SUFX = .tar.bz2 +EXTRACT_SUFX = .tar.xz CATEGORIES = games x11 HOMEPAGE = http://www.eduke32.com/ @@ -25,12 +24,15 @@ BUILD_DEPENDS = archivers/unzip \ BUILD_DEPENDS += devel/nasm .endif -WANTLIB = SDL c m pthread stdc++ SDL_mixer>=3 vorbisfile vpx +WANTLIB += c m ogg png pthread vorbis vorbisfile vpx z +WANTLIB += SDL2 SDL2_mixer FLAC execinfo LIB_DEPENDS = audio/libvorbis \ - multimedia/libvpx \ - devel/sdl \ - devel/sdl-mixer + audio/flac \ + devel/sdl2 \ + devel/sdl2-mixer \ + devel/libexecinfo \ + multimedia/libvpx # dlopen()'d; shows as Extra in check-lib-depends but it is used LIB_DEPENDS += x11/gtk+2 @@ -38,31 +40,23 @@ WANTLIB += gtk-x11-2.0 RUN_DEPENDS = games/duke3ddata -MASTER_SITES = http://dukeworld.duke4.net/eduke32/synthesis/old/${RDATE}-${RTAG}/ \ - http://dukeworld.duke4.net/eduke32/synthesis/${RDATE}-${RTAG}/ +MASTER_SITES = http://dukeworld.duke4.net/eduke32/synthesis/${RDATE}-${RTAG}/ + +MODULES = gcc4 +MODGCC4_LANGS =c c++ +MODGCC4_ARCHS =* # make sure we can see compilation output, respect CC/CXX, don't strip! -MAKE_FLAGS = PRETTY_OUTPUT=0 CC="${CC}" CXX="${CXX}" DEBUGANYWAY=1 +MAKE_ENV = CC="${CC}" CXX="${CXX}" PRETTY_OUTPUT=0 PACKAGE_REPOSITORY=1 USE_GMAKE =Yes NO_TEST = Yes -WRKDIST = ${WRKDIR}/eduke32_${RDATE}-${RTAG}/ +WRKDIST = ${WRKDIR}/eduke32_${RDATE}-${RTAG} post-extract: # check for data files in correct location - @perl -pi -e "s,/usr/share/games/eduke32,${TRUEPREFIX}/share/duke3d," \ - ${WRKSRC}/source/astub.c - @perl -pi -e "s,/usr/share/games/eduke32,${TRUEPREFIX}/share/duke3d," \ - ${WRKSRC}/source/game.c - @perl -pi -e "s,./eduke32,eduke32," ${WRKSRC}/build/include/editor.h - # make plaintext files less annoying to read... - @perl -MExtUtils::Command -e dos2unix \ - ${WRKSRC}/tiles.cfg \ - ${WRKSRC}/names.h \ - ${WRKSRC}/GNU.TXT \ - ${WRKSRC}/buildlic.txt \ - ${WRKSRC}/samples/*.sample \ - ${WRKSRC}/samples/*.txt + @sed -i "s,/usr/share/games/eduke32,${TRUEPREFIX}/share/duke3d," \ + ${WRKSRC}/source/common.c do-install: ${INSTALL_PROGRAM} ${WRKBUILD}/eduke32 ${PREFIX}/bin @@ -71,17 +65,25 @@ do-install: ${INSTALL_DATA_DIR} ${PREFIX}/share/doc/eduke32 ${INSTALL_DATA_DIR} ${PREFIX}/share/examples/eduke32 # mapster32 help files - ${INSTALL_DATA} ${WRKDIST}/m32help.hlp ${PREFIX}/share/duke3d/ - ${INSTALL_DATA} ${WRKDIST}/SEHELP.HLP ${PREFIX}/share/duke3d/ - ${INSTALL_DATA} ${WRKDIST}/STHELP.HLP ${PREFIX}/share/duke3d/ + ${INSTALL_DATA} ${WRKDIST}/package/sdk/m32help.hlp \ + ${PREFIX}/share/duke3d/ + ${INSTALL_DATA} ${WRKDIST}/package/sdk/SEHELP.HLP \ +