CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2012/12/11 04:01:31 Modified files: converters/mpack: Makefile distinfo converters/mpack/patches: patch-decode_c patch-encode_c patch-magic_c patch-unixos_c patch-unixpk_c Added files: converters/mpack/patches: patch-uudecode_c Log message: - fix unpacking of multiple parts (from Gentoo, possibly via FreeBSD). - recognise PNG files and set appropriate mime type (from FreeBSD). - remove botched mkstemp() conversion which broke mpack(1); a proper mkstemp conversion is intrusive. - use some additional in the temporary filename. - let this build with CLang scan-build.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: dco...@cvs.openbsd.org 2012/12/11 04:51:39 Modified files: devel/cmake: Makefile Log message: Add DPB_PROPERTIES=parallel Test/OK sthen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: dco...@cvs.openbsd.org 2012/12/11 04:59:31 Modified files: devel/libevent2: Makefile distinfo Log message: Bugfix update to libevent-2.0.21 From David Hill
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: lan...@cvs.openbsd.org 2012/12/11 06:31:38 ports/mail/geary/patches Update of /cvs/ports/mail/geary/patches In directory cvs.openbsd.org:/tmp/cvs-serv24219/patches Log Message: Directory /cvs/ports/mail/geary/patches added to the repository
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: lan...@cvs.openbsd.org 2012/12/11 06:33:06 Modified files: mail/geary : Makefile distinfo mail/geary/pkg : PLIST Added files: mail/geary/patches: patch-cmake_FindIntltool_cmake patch-src_client_geary-application_vala Log message: Update to geary 0.2.2. Based on a diff from Juan Francisco Cantero, added gettext MODULE and run depends on gnome-keyring.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: m...@cvs.openbsd.org2012/12/11 07:27:29 Modified files: devel/tig : Makefile distinfo Log message: Update to tig 1.1 ok tobiasu@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2012/12/11 07:46:41 Modified files: telephony/asterisk: Makefile distinfo telephony/asterisk/patches: patch-configure_ac patch-pbx_pbx_spool_c Log message: bugfix update to Asterisk 10.11.0 - while there, revise pbx_spool.c kevent timeout fix; rather than clamping the timestamp, in the particular problem situation we hit the loop (where dirlist is empty), pass in NULL rather than INT_MAX-timenow similar to what's done in the inotify case.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2012/12/11 08:30:56 Modified files: net/net-snmp : Makefile Added files: net/net-snmp/patches: patch-agent_mibgroup_default_modules_h Log message: disable notification-log-mib for now; causes an 8KB/trap leak in snmptrapd
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2012/12/11 08:33:55 Modified files: net/net-snmp : Tag: OPENBSD_5_2 Makefile Added files: net/net-snmp/patches: Tag: OPENBSD_5_2 patch-agent_mibgroup_default_modules_h Log message: MFC: disable notification-log-mib for now; causes an 8KB/trap leak in snmptrapd
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2012/12/11 10:11:07 Modified files: x11/gnome/at-spi2-core: Makefile distinfo x11/gnome/at-spi2-core/pkg: PLIST Log message: Bugfix update to at-spi2-core-2.6.3.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2012/12/11 10:26:51 Modified files: sysutils/ruby-puppet: Makefile sysutils/ruby-puppet/patches: patch-lib_puppet_defaults_rb Log message: Fix default modulepath to follow upstream's behavior. ok robert@ (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2012/12/11 10:54:13 Modified files: x11/gnome/grilo-plugins: Makefile distinfo x11/gnome/grilo-plugins/pkg: PLIST Log message: - update to grilo-plugins-0.2.5
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gonz...@cvs.openbsd.org 2012/12/11 11:15:13 Modified files: net/balance: Makefile distinfo net/balance/patches: patch-balance_h net/balance/pkg: PLIST Removed files: net/balance/patches: patch-balance_c Log message: Update for balance to 3.54 * A bug with hash_fold() regarding incoming IPv4 and IPv6 source addresses has been fixed. The hash group mechanism is now working as expected. * Buffering has been disabled for interactive shell IO. A new assign command has been added to allow changing of the host:port assignment of a channel (only if disabled). A locking bug has been fixed. * A new option -6 has been added to force IPv6 bind. * Problems with setting IPV6_V6ONLY socket option are now handled more nicely with a syslog warning message. * Balance now compiles also on systems where IPV6_V6ONLY is undefined (like some Solaris systems). * IPv6 support on the listening side has been added. MAXCHANNELS in balance.h has been increased to 64. Ok aja@ brad@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2012/12/11 11:22:54 Modified files: www/squid : Makefile www/squid/patches: patch-configure_ac Log message: mark BROKEN-hppa and remove my broken attempt at fixing the broken autoconf test for atomic ops.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: lan...@cvs.openbsd.org 2012/12/11 13:47:45 Modified files: security/gnupg/patches: patch-cipher_Makefile_in patch-g10_Makefile_in patch-keyserver_Makefile_in patch-tools_Makefile_in patch-util_Makefile_in Log message: Regen patches
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: na...@cvs.openbsd.org 2012/12/11 14:09:18 Modified files: misc/xnc : Makefile misc/xnc/patches: patch-configure_in Log message: use openpty()
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: lan...@cvs.openbsd.org 2012/12/11 14:12:36 Modified files: security/gnupg : Makefile Log message: Fix on hppa, broken since the PIE switch. mpi/hppa1.1/udiv-qrnnd.S is not PIE-safe, let's use mpi/hppa/udiv-qrnnd.S instead. with kettenis@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2012/12/11 15:12:12 Modified files: lang/lua : lua.port.mk Log message: allow MODLUA_VERSION to be overridden; ok jasper@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: rpoin...@cvs.openbsd.org2012/12/11 23:07:14 Modified files: devel/mercurial: Makefile distinfo devel/mercurial/pkg: PLIST Log message: Update mercurial to 2.4. ok landry@, thanks to Samir Saada to have tested on loongson.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: rpoin...@cvs.openbsd.org2012/12/11 23:07:53 Modified files: devel/tortoisehg: Makefile distinfo devel/tortoisehg/pkg: PLIST Log message: update tortoisehg to 2.6. ok landry@.
Re: FW: Re: [ipxe-devel] [OpenBSD] 'internal compiler error: Segmentation' or 'Error: bad expression used with @GOTOFF' fault]
On Mon, Dec 10, 2012 at 05:56:56PM -0500, Jiri B wrote: Hi, I tried to make IPXE running on OpenBSD and got this reply from IPXE guy asking for binutils version - huh, how to check it? ld -v, as -v... you know what binutils is, right ?
Re: FW: Re: [ipxe-devel] [OpenBSD] 'internal compiler error: Segmentation' or 'Error: bad expression used with @GOTOFF' fault]
On 2012/12/10 17:56, Jiri B wrote: Hi, I tried to make IPXE running on OpenBSD and got this reply from IPXE guy asking for binutils version - huh, how to check it? My OpenBSD install was some last November snapshot for amd64. 2.15 (2004 vintage), yes it's old, and yes there are reasons why it hasn't been updated. ;) [BUILD] bin/tls.o net/tls.c: In function 'tls_new_server_hello': net/tls.c:1668: internal compiler error: Segmentation fault Please submit a full bug report, with preprocessed source if appropriate. See URL:http://gcc.gnu.org/bugs.html for instructions. gmake: *** [bin/tls.o] Error 1 This sounds like a gcc-4.2.1 bug. Using a newer version is the way to go. Or identifying the fix and see if similar can be done to gcc-4.2.1. Or have you tried clang?
Re: [update] devel/libevent2
On Mon, Dec 10, 2012 at 8:42 PM, David Hill dh...@mindcry.org wrote: update libevent2 to 2.0.21. Tested with transmission and cyphertite on i386. Please test and commit. Tested and builds/works ok. I'll commit this later today, if nobody beats me. Ciao, David
Re: boost 1.52.0 status
On Mon, Dec 10, 2012 at 10:56 PM, Stuart Henderson s...@spacehopper.org wrote: On 2012/12/10 13:21, Brad Smith wrote: On Mon, Dec 10, 2012 at 08:04:41AM -0600, Amit Kulkarni wrote: what about having DPB_PROPERTIES=parallel ? Not sure about that. Someone else who is familiar with this feature should comment as to whether it should be added or not. I've done at least half a dozen builds with boost 1.42.0 marked that way without any problems, it is on the critical path for libreoffice in a bulk package build and takes a fair amount of time to build, I am OK with this. (Likewise for devel/cmake). If ok, I will add it to cmake. Ciao, David
Re: boost 1.52.0 status
what about having DPB_PROPERTIES=parallel ? Not sure about that. Someone else who is familiar with this feature should comment as to whether it should be added or not. I've done at least half a dozen builds with boost 1.42.0 marked that way without any problems, it is on the critical path for libreoffice in a bulk package build and takes a fair amount of time to build, I am OK with this. (Likewise for devel/cmake). If ok, I will add it to cmake. please, it takes time for a new cmake to build, and it unlocks all of kde4. likewise a DPB_PARALLEL is needed for kdelibs4.
Re: redland-bindings
the perl bindings are working well for me. On 13/11/2012, at 1:31 AM, Stuart Henderson s...@spacehopper.org wrote: Here's a redland-bindings port, which requires the recently commited updates to RDF libs. Based on a port from dlg, I've tweaked the multi packages stuff. Usual problem with the Perl bindings, they need libpthread preloading, there is a MESSAGE file to explain this for now. redland-bindings.tgz
UPDATE: Adsuck-2.5.0
Hi, Update for Adsuck to 2.5.0: * Automatically watch for changes to resolv.conf and reread it when that happens. * Refresh all the hosts files. * Start using libevent2 in favor of libevent Ok? Comments? Cheers. -- Sending from my VCR... Index: Makefile === RCS file: /cvs/ports/net/adsuck/Makefile,v retrieving revision 1.23 diff -u -p -r1.23 Makefile --- Makefile27 Sep 2012 18:20:41 - 1.23 +++ Makefile11 Dec 2012 17:13:13 - @@ -2,7 +2,7 @@ COMMENT= DNS relay for ad blocking -DISTNAME= adsuck-2.4.3 +DISTNAME= adsuck-2.5.0 CATEGORIES=net HOMEPAGE= http://opensource.conformal.com/wiki/Adsuck @@ -17,9 +17,10 @@ PERMIT_DISTFILES_FTP=Yes MASTER_SITES= http://opensource.conformal.com/snapshots/adsuck/ -WANTLIB += c event ldns +WANTLIB += c event_core event_extra ldns -LIB_DEPENDS= net/ldns/libldns=1.6.5 +LIB_DEPENDS= devel/libevent2 \ + net/ldns/libldns=1.6.5 MAKE_FLAGS=BINDIR=${TRUEPREFIX}/sbin \ MANDIR=${TRUEPREFIX}/man/man Index: distinfo === RCS file: /cvs/ports/net/adsuck/distinfo,v retrieving revision 1.13 diff -u -p -r1.13 distinfo --- distinfo27 Sep 2012 18:20:41 - 1.13 +++ distinfo11 Dec 2012 17:13:13 - @@ -1,2 +1,2 @@ -SHA256 (adsuck-2.4.3.tgz) = Sm0xYeyvx/54cu4+eeYgv0IGJmu4FG0ovapWpqZEGWg= -SIZE (adsuck-2.4.3.tgz) = 10492872 +SHA256 (adsuck-2.5.0.tgz) = EmAuySScbHMD+QF7yQdHKK9jxFq+wVuYPkemZNcCQmM= +SIZE (adsuck-2.5.0.tgz) = 10443584
Re: boost 1.52.0 status
Brad Smith b...@comstyle.com wrote: what about having DPB_PROPERTIES=parallel ? Not sure about that. Someone else who is familiar with this feature should comment as to whether it should be added or not. You run a build with MAKE_JOBS=1 and one with MAKE_JOBS=2 (or better 4 if you have the hardware) and compare the build times. If it sorta, kinda scales, we enable DPB_PROPERTIES=parallel. The last time I checked boost it scarled poorly enough that I thought enabling the parallel build wasn't worth it, but I don't remember the figures. -- Christian naddy Weisgerber na...@mips.inka.de
Re: boost 1.52.0 status
what about having DPB_PROPERTIES=parallel ? Not sure about that. Someone else who is familiar with this feature should comment as to whether it should be added or not. You run a build with MAKE_JOBS=1 and one with MAKE_JOBS=2 (or better 4 if you have the hardware) and compare the build times. If it sorta, kinda scales, we enable DPB_PROPERTIES=parallel. The last time I checked boost it scarled poorly enough that I thought enabling the parallel build wasn't worth it, but I don't remember the figures. i will try running with values of 1, 2 and 4 with a rm -rf /usr/ports/pobj/.ccache/ and report.
Parallel cmake (was: Re: boost 1.52.0 status)
David Coppa dco...@gmail.com wrote: If ok, I will add [DPB_PROPERTIES=parallel] to cmake. Which does absolutely nothing. MAKE_JOBS has no effect on cmake. The build is strictly sequential. You can parallelize the bootstrap phase with CONFIGURE_ARGS += --parallel=${MAKE_JOBS} but the main build remains sequential so the overall gain is negligible. I haven't been able to figure out how to get cmake to build things in parallel, but then I'm largely unfamiliar with cmake. -- Christian naddy Weisgerber na...@mips.inka.de
UPDATE: net/nmap
Update to latest version, a lua.port.mk diff is needed to build it because it needs at least lua 5.2. Comments ? Hints ? Ok ? Cheers Giovanni Index: lua.port.mk === RCS file: /cvs/ports/lang/lua/lua.port.mk,v retrieving revision 1.23 diff -u -p -r1.23 lua.port.mk --- lua.port.mk 31 Oct 2012 13:39:32 - 1.23 +++ lua.port.mk 11 Dec 2012 18:30:59 - @@ -8,7 +8,7 @@ CATEGORIES+= lang/lua # # Define the default version and use that if MODLUA_VERSION is not set. -MODLUA_DEFAULT_VERSION= 5.1 +MODLUA_DEFAULT_VERSION?= 5.1 # If a port already has flavors, append our lua flavors to it, unless it requests a specific # version of lua. Otherwise set the FLAVORS list to just the lua flavors. Index: Makefile === RCS file: /cvs/ports/net/nmap/Makefile,v retrieving revision 1.104 diff -u -p -r1.104 Makefile --- Makefile 16 Nov 2012 17:05:13 - 1.104 +++ Makefile 11 Dec 2012 18:30:05 - @@ -3,14 +3,11 @@ COMMENT-main= scan ports and fingerprint stack of network hosts COMMENT-zenmap= graphical frontend for nmap -MODPY_EGG_VERSION= 6.01 +MODPY_EGG_VERSION= 6.25 DISTNAME= nmap-${MODPY_EGG_VERSION} PKGNAME-main= ${DISTNAME} PKGNAME-zenmap= nmap-zenmap-${MODPY_EGG_VERSION} -REVISION-main= 1 -REVISION-zenmap=1 - CATEGORIES= net security MASTER_SITES= http://nmap.org/dist/ EXTRACT_SUFX= .tgz @@ -28,6 +25,8 @@ PERMIT_DISTFILES_FTP= Yes MULTI_PACKAGES= -main -zenmap USE_GMAKE= yes +MODLUA_SA= Yes +MODLUA_DEFAULT_VERSION=5.2 MODULES= lang/python \ lang/lua Index: distinfo === RCS file: /cvs/ports/net/nmap/distinfo,v retrieving revision 1.26 diff -u -p -r1.26 distinfo --- distinfo 16 Jul 2012 07:39:31 - 1.26 +++ distinfo 11 Dec 2012 18:30:05 - @@ -1,2 +1,2 @@ -SHA256 (nmap-6.01.tgz) = 5p++vOu32VCcqRM8ObL5GZGkRzjRhEyvOF8fP2b3OXk= -SIZE (nmap-6.01.tgz) = 23204848 +SHA256 (nmap-6.25.tgz) = QsVP1cxfBeXfkJdX1iMHECqP8d+vXCYmsctfLs0Anlc= +SIZE (nmap-6.25.tgz) = 9374716 Index: patches/patch-FPEngine_cc === RCS file: /cvs/ports/net/nmap/patches/patch-FPEngine_cc,v retrieving revision 1.1 diff -u -p -r1.1 patch-FPEngine_cc --- patches/patch-FPEngine_cc 16 Jul 2012 07:39:31 - 1.1 +++ patches/patch-FPEngine_cc 11 Dec 2012 18:30:05 - @@ -1,7 +1,7 @@ $OpenBSD: patch-FPEngine_cc,v 1.1 2012/07/16 07:39:31 giovanni Exp $ FPEngine.cc.orig Fri May 25 11:47:44 2012 -+++ FPEngine.cc Fri May 25 11:49:23 2012 -@@ -479,7 +479,7 @@ void FPNetworkControl::response_reception_handler(nsoc +--- FPEngine.cc.orig Wed Oct 31 18:00:55 2012 FPEngine.cc Fri Nov 30 12:41:47 2012 +@@ -477,7 +477,7 @@ void FPNetworkControl::response_reception_handler(nsoc enum nse_type type = nse_type(nse); const u8 *rcvd_pkt = NULL;/* Points to the captured packet */ size_t rcvd_pkt_len = 0; /* Lenght of the captured packet */ Index: patches/patch-Makefile_in === RCS file: /cvs/ports/net/nmap/patches/patch-Makefile_in,v retrieving revision 1.13 diff -u -p -r1.13 patch-Makefile_in --- patches/patch-Makefile_in 16 Jul 2012 07:39:31 - 1.13 +++ patches/patch-Makefile_in 11 Dec 2012 18:30:05 - @@ -1,6 +1,6 @@ $OpenBSD: patch-Makefile_in,v 1.13 2012/07/16 07:39:31 giovanni Exp $ Makefile.in.orig Tue Apr 3 02:21:44 2012 -+++ Makefile.in Wed May 30 10:48:53 2012 +--- Makefile.in.orig Thu Sep 13 02:17:42 2012 Makefile.in Fri Nov 30 12:41:47 2012 @@ -45,7 +45,7 @@ DEFS += -D_FORTIFY_SOURCE=2 # Should only be enabled during debugging and not in any real release. # DEFS += -DMTRACE=1 @@ -10,7 +10,7 @@ $OpenBSD: patch-Makefile_in,v 1.13 2012/ export CFLAGS = $(CXXFLAGS) # CFLAGS = $(DEFS) $(INCLS) STATIC = -@@ -284,8 +284,8 @@ $(NPINGDIR)/nping.h: nmap.h +@@ -285,8 +285,8 @@ $(NPINGDIR)/nping.h: nmap.h mv -f $@.tmp $@ # Update the version number used by Zenmap. Index: patches/patch-configure === RCS file: /cvs/ports/net/nmap/patches/patch-configure,v retrieving revision 1.15 diff -u -p -r1.15 patch-configure --- patches/patch-configure 16 Jul 2012 07:39:31 - 1.15 +++ patches/patch-configure 11 Dec 2012 18:30:05 - @@ -1,6 +1,6 @@ $OpenBSD: patch-configure,v 1.15 2012/07/16 07:39:31 giovanni Exp $ configure.orig Sat Apr 7 10:16:08 2012 -+++ configure Wed Jul 11 15:05:35 2012 +--- configure.orig Fri Nov 16 01:54:16 2012 configure Fri Nov 30 12:41:47 2012 @@ -6311,6 +6311,7 @@ else extern char pcap_version[]; int main() { @@ -8,13 +8,13 @@ $OpenBSD: patch-configure,v 1.15 2012/07 + exit(0); /* fool it, use our libpcap */ sscanf(pcap_version,%d.%d.%d, major, minor1, minor2); if (major 0) - exit(0); + return 0; @@
Re: UPDATE: net/nmap
On Tue, Dec 11, 2012 at 7:39 PM, Giovanni Bechis giova...@bigio.snb.it wrote: Update to latest version, a lua.port.mk diff is needed to build it because it needs at least lua 5.2. Comments ? Hints ? Ok ? Cheers Giovanni imho this is not right. Just set MODLUA_VERSION=5.2 into net/nmap/Makefile ciao, David
Re: UPDATE: net/nmap
On 12/11/12 19:49, David Coppa wrote: On Tue, Dec 11, 2012 at 7:39 PM, Giovanni Bechis giova...@bigio.snb.it wrote: Update to latest version, a lua.port.mk diff is needed to build it because it needs at least lua 5.2. Comments ? Hints ? Ok ? Cheers Giovanni imho this is not right. Just set MODLUA_VERSION=5.2 into net/nmap/Makefile This does not work, lua.port.mk still thinks that lua-5.1 is needed. $ make patch === nmap-6.25 depends on: lua-=5.1,5.2 - lua-5.1.5p1 === Verifying specs: c crypto m pcap ssl stdc++ pcre pthread lua5.1 === found c.66.0 crypto.21.0 m.7.1 pcap.7.0 ssl.19.0 stdc++.55.0 pcre.3.0 pthread.17.0 lua5.1.5.1 Cheers Giovanni
Re: Parallel cmake (was: Re: boost 1.52.0 status)
Which does absolutely nothing. MAKE_JOBS has no effect on cmake. The build is strictly sequential. You can parallelize the bootstrap phase with CONFIGURE_ARGS += --parallel=${MAKE_JOBS} but the main build remains sequential so the overall gain is negligible. I haven't been able to figure out how to get cmake to build things in parallel, but then I'm largely unfamiliar with cmake. in cmake based builds when you cd WRKOBJDIR and do a gmake -j4 it usually works in driving all cores to utilization. I tried with USE_GMAKE = Yes and it drives all cores so I think please add CONFIGURE_ARGS = --parallel=${MAKE_JOBS} and USE_GMAKE = Yes and then both configure and build will go parallel.
Re: new: net/tinc
Ping? Is there more testing needed or are there other issues preventing this port from being included in the ports tree? -- Gregor Best
Re: Parallel cmake (was: Re: boost 1.52.0 status)
On Tue, Dec 11, 2012 at 7:59 PM, Amit Kulkarni amitk...@gmail.com wrote: Which does absolutely nothing. MAKE_JOBS has no effect on cmake. The build is strictly sequential. You can parallelize the bootstrap phase with CONFIGURE_ARGS += --parallel=${MAKE_JOBS} but the main build remains sequential so the overall gain is negligible. I haven't been able to figure out how to get cmake to build things in parallel, but then I'm largely unfamiliar with cmake. in cmake based builds when you cd WRKOBJDIR and do a gmake -j4 it usually works in driving all cores to utilization. I tried with USE_GMAKE = Yes and it drives all cores so I think please add CONFIGURE_ARGS = --parallel=${MAKE_JOBS} and USE_GMAKE = Yes and then both configure and build will go parallel. Please just let me know what should I do, because I know nothing about DPB and related stuff (yet) ciao, david
Re: Parallel cmake (was: Re: boost 1.52.0 status)
Which does absolutely nothing. MAKE_JOBS has no effect on cmake. The build is strictly sequential. You can parallelize the bootstrap phase with CONFIGURE_ARGS += --parallel=${MAKE_JOBS} but the main build remains sequential so the overall gain is negligible. I haven't been able to figure out how to get cmake to build things in parallel, but then I'm largely unfamiliar with cmake. in cmake based builds when you cd WRKOBJDIR and do a gmake -j4 it usually works in driving all cores to utilization. I tried with USE_GMAKE = Yes and it drives all cores so I think please add CONFIGURE_ARGS = --parallel=${MAKE_JOBS} and USE_GMAKE = Yes and then both configure and build will go parallel. Please just let me know what should I do, because I know nothing about DPB and related stuff (yet) Ok, apply this diff ? cmake.diff Index: Makefile === RCS file: /cvs/ports/devel/cmake/Makefile,v retrieving revision 1.73 diff -u -p -r1.73 Makefile --- Makefile11 Dec 2012 11:51:39 - 1.73 +++ Makefile11 Dec 2012 19:06:41 - @@ -8,7 +8,7 @@ HOMEPAGE = http://www.cmake.org/ CATEGORIES = devel COMMENT = portable build system DISTNAME = cmake-2.8.10.2 -REVISION = 0 +REVISION = 1 MASTER_SITES = ${HOMEPAGE}files/v2.8/ MAINTAINER = David Coppa dco...@openbsd.org @@ -29,12 +29,16 @@ CONFIGURE_STYLE = simple CONFIGURE_ARGS = --prefix=${PREFIX} \ --datadir=/share/cmake \ --docdir=/share/doc/cmake \ - --system-libs + --system-libs \ + --parallel=${MAKE_JOBS} PORTHOME = ${WRKDIR} SEPARATE_BUILD = Yes USE_GROFF =Yes + +# gmake enables dpb parallel build +USE_GMAKE =Yes REGRESS_TARGET = test then use this in a .sh file (watch amd64/i386 etc... or use machine) and observe if all cores are used with top subst 4 with real cores you have i.e hw.ncpu rm -rf /usr/ports/logs/amd64/locks/* rm -rf /usr/ports/pobj/locks/* /usr/ports/infrastructure/bin/dpb -j 4 -p 4 -u -U -R -J 10 devel/cmake
Re: Parallel cmake (was: Re: boost 1.52.0 status)
On Tue, Dec 11, 2012 at 01:22:27PM -0600, Amit Kulkarni wrote: Which does absolutely nothing. MAKE_JOBS has no effect on cmake. The build is strictly sequential. You can parallelize the bootstrap phase with CONFIGURE_ARGS += --parallel=${MAKE_JOBS} but the main build remains sequential so the overall gain is negligible. I haven't been able to figure out how to get cmake to build things in parallel, but then I'm largely unfamiliar with cmake. in cmake based builds when you cd WRKOBJDIR and do a gmake -j4 it usually works in driving all cores to utilization. I tried with USE_GMAKE = Yes and it drives all cores so I think please add CONFIGURE_ARGS = --parallel=${MAKE_JOBS} and USE_GMAKE = Yes and then both configure and build will go parallel. Please just let me know what should I do, because I know nothing about DPB and related stuff (yet) Ok, apply this diff ? cmake.diff Index: Makefile === RCS file: /cvs/ports/devel/cmake/Makefile,v retrieving revision 1.73 diff -u -p -r1.73 Makefile --- Makefile 11 Dec 2012 11:51:39 - 1.73 +++ Makefile 11 Dec 2012 19:06:41 - @@ -8,7 +8,7 @@ HOMEPAGE =http://www.cmake.org/ CATEGORIES = devel COMMENT =portable build system DISTNAME = cmake-2.8.10.2 -REVISION = 0 +REVISION = 1 MASTER_SITES = ${HOMEPAGE}files/v2.8/ MAINTAINER = David Coppa dco...@openbsd.org @@ -29,12 +29,16 @@ CONFIGURE_STYLE = simple CONFIGURE_ARGS = --prefix=${PREFIX} \ --datadir=/share/cmake \ --docdir=/share/doc/cmake \ - --system-libs + --system-libs \ + --parallel=${MAKE_JOBS} PORTHOME = ${WRKDIR} SEPARATE_BUILD = Yes USE_GROFF = Yes + +# gmake enables dpb parallel build +USE_GMAKE = Yes REGRESS_TARGET = test The use of GNU make like this seems bogus. I take it CMake isn't doing something right. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
Re: Parallel cmake (was: Re: boost 1.52.0 status)
I tried with USE_GMAKE = Yes and it drives all cores so I think please add CONFIGURE_ARGS = --parallel=${MAKE_JOBS} and USE_GMAKE = Yes and then both configure and build will go parallel. Please just let me know what should I do, because I know nothing about DPB and related stuff (yet) Ok, apply this diff ? cmake.diff Index: Makefile === RCS file: /cvs/ports/devel/cmake/Makefile,v retrieving revision 1.73 diff -u -p -r1.73 Makefile --- Makefile 11 Dec 2012 11:51:39 - 1.73 +++ Makefile 11 Dec 2012 19:06:41 - @@ -8,7 +8,7 @@ HOMEPAGE =http://www.cmake.org/ CATEGORIES = devel COMMENT =portable build system DISTNAME = cmake-2.8.10.2 -REVISION = 0 +REVISION = 1 MASTER_SITES = ${HOMEPAGE}files/v2.8/ MAINTAINER = David Coppa dco...@openbsd.org @@ -29,12 +29,16 @@ CONFIGURE_STYLE = simple CONFIGURE_ARGS = --prefix=${PREFIX} \ --datadir=/share/cmake \ --docdir=/share/doc/cmake \ - --system-libs + --system-libs \ + --parallel=${MAKE_JOBS} PORTHOME = ${WRKDIR} SEPARATE_BUILD = Yes USE_GROFF = Yes + +# gmake enables dpb parallel build +USE_GMAKE = Yes REGRESS_TARGET = test The use of GNU make like this seems bogus. I take it CMake isn't doing something right. as far back as 2008 I seem to remember cmake having implicit dependency on gmake (i tried it on on Solaris,FreeBSD and then OpenBSD). my builds would always parallel build with gmake but not with plain os make. i still haven't figured out why. maybe david can chime in with more... perhaps http://cmake.3232098.n2.nabble.com/parallel-make-j-N-td7193296.html which leads to http://public.kitware.com/Bug/view.php?id=12882
Re: boost 1.52.0 status
what about having DPB_PROPERTIES=parallel ? Not sure about that. Someone else who is familiar with this feature should comment as to whether it should be added or not. You run a build with MAKE_JOBS=1 and one with MAKE_JOBS=2 (or better 4 if you have the hardware) and compare the build times. If it sorta, kinda scales, we enable DPB_PROPERTIES=parallel. The last time I checked boost it scarled poorly enough that I thought enabling the parallel build wasn't worth it, but I don't remember the figures. results are in... MAKE_JOBS=1 22m44.71s real18m7.76s user 2m43.81s system === Cleaning for boost-1.52.0p1 MAKE_JOBS=2 12m52.11s real18m24.61s user 3m51.25s system === Cleaning for boost-1.52.0p1 MAKE_JOBS=4 8m17.98s real19m8.27s user 6m3.54s system === Cleaning for boost-1.52.0p1 so yes, it does provide a boost :-) this is with a cpu0: Intel(R) Xeon(R) CPU E3-1220 V2 @ 3.10GHz, 3093.39 MHz
Re: UPDATE: net/nmap
On 2012/12/11 19:54, Giovanni Bechis wrote: On 12/11/12 19:49, David Coppa wrote: On Tue, Dec 11, 2012 at 7:39 PM, Giovanni Bechis giova...@bigio.snb.it wrote: Update to latest version, a lua.port.mk diff is needed to build it because it needs at least lua 5.2. Comments ? Hints ? Ok ? Cheers Giovanni imho this is not right. Just set MODLUA_VERSION=5.2 into net/nmap/Makefile This does not work, lua.port.mk still thinks that lua-5.1 is needed. That should be fixed so it works then, overriding MODLUA_DEFAULT_VERSION like this is definitely wrong. OK? Index: lua.port.mk === RCS file: /cvs/ports/lang/lua/lua.port.mk,v retrieving revision 1.23 diff -u -p -r1.23 lua.port.mk --- lua.port.mk 31 Oct 2012 13:39:32 - 1.23 +++ lua.port.mk 11 Dec 2012 20:57:17 - @@ -24,7 +24,7 @@ FLAVOR?= # empty .if ${FLAVOR:L:Mlua52} MODLUA_VERSION=5.2 .else -MODLUA_VERSION=${MODLUA_DEFAULT_VERSION} +MODLUA_VERSION?= ${MODLUA_DEFAULT_VERSION} .endif .if ${MODLUA_VERSION} == 5.1
Re: UPDATE: net/nmap
On Tue, Dec 11, 2012 at 08:58:05PM +, Stuart Henderson wrote: On 2012/12/11 19:54, Giovanni Bechis wrote: On 12/11/12 19:49, David Coppa wrote: On Tue, Dec 11, 2012 at 7:39 PM, Giovanni Bechis giova...@bigio.snb.it wrote: Update to latest version, a lua.port.mk diff is needed to build it because it needs at least lua 5.2. Comments ? Hints ? Ok ? Cheers Giovanni imho this is not right. Just set MODLUA_VERSION=5.2 into net/nmap/Makefile This does not work, lua.port.mk still thinks that lua-5.1 is needed. That should be fixed so it works then, overriding MODLUA_DEFAULT_VERSION like this is definitely wrong. OK? Thanks; certainly OK. Index: lua.port.mk === RCS file: /cvs/ports/lang/lua/lua.port.mk,v retrieving revision 1.23 diff -u -p -r1.23 lua.port.mk --- lua.port.mk 31 Oct 2012 13:39:32 - 1.23 +++ lua.port.mk 11 Dec 2012 20:57:17 - @@ -24,7 +24,7 @@ FLAVOR?=# empty .if ${FLAVOR:L:Mlua52} MODLUA_VERSION= 5.2 .else -MODLUA_VERSION= ${MODLUA_DEFAULT_VERSION} +MODLUA_VERSION?= ${MODLUA_DEFAULT_VERSION} .endif .if ${MODLUA_VERSION} == 5.1 -- Cheers, Jasper Stay Hungry. Stay Foolish
Re: Parallel cmake (was: Re: boost 1.52.0 status)
On Tue, 11 Dec 2012, Brad Smith wrote: The use of GNU make like this seems bogus. I take it CMake isn't doing something right. Hi, I've found another (related) problem. $MAKE_PROGRAM is not honored and cmake picks gmake up if it's installed (even without USE_GMAKE=Yes). From cmake-2.8.10.2/bootstrap: ---8--- CMAKE_KNOWN_MAKE_PROCESSORS=gmake make ... # If MAKE is set, use that for make processor, otherwise use list of known make if [ -n ${MAKE} ]; then cmake_make_processors=${MAKE} else cmake_make_processors=${CMAKE_KNOWN_MAKE_PROCESSORS} fi ---8--- Ok for the diff below? I've also added --parallel=${MAKE_JOBS} to CONFIGURE_ARGS. I still need to investigate the make vs. gmake parallelization issues... cheers, David Index: Makefile === RCS file: /cvs/ports/devel/cmake/Makefile,v retrieving revision 1.73 diff -u -p -r1.73 Makefile --- Makefile11 Dec 2012 11:51:39 - 1.73 +++ Makefile12 Dec 2012 05:52:46 - @@ -8,7 +8,7 @@ HOMEPAGE = http://www.cmake.org/ CATEGORIES = devel COMMENT = portable build system DISTNAME = cmake-2.8.10.2 -REVISION = 0 +REVISION = 1 MASTER_SITES = ${HOMEPAGE}files/v2.8/ MAINTAINER = David Coppa dco...@openbsd.org @@ -29,7 +29,9 @@ CONFIGURE_STYLE = simple CONFIGURE_ARGS = --prefix=${PREFIX} \ --datadir=/share/cmake \ --docdir=/share/doc/cmake \ - --system-libs + --system-libs \ + --parallel=${MAKE_JOBS} +CONFIGURE_ENV =MAKE=${MAKE_PROGRAM} PORTHOME = ${WRKDIR}