Re: WIP: GCC 6.2.0
On Wed, Sep 7, 2016 at 1:43 AM, Pascal Stumpfwrote: > On Tue, 6 Sep 2016 15:08:37 +0800, Dongsheng Song wrote: > > On Sun, Sep 4, 2016 at 8:30 PM, Pascal Stumpf wrote: > > > > > Here's a preliminary port of GCC 6.2.0, tested only on amd64, i386 and > > > sparc64 as I don't have access to anything else at the moment. > > > > > > Go ahead and play with it. > > > > > > > Thanks for your great work ! > > > > But when I install gcc 6.2, it conflicted with gcc 4.9.4, I must remove > gcc > > 4.9.4 ( then llvm 3.8.1) at first. > > > > Could you support them coexist by the version specific technology? > > > > That's quite a bit of work for questionable benefit, as we usually > switch our whole ports tree to one specific version at once and then get > rid of the old one. But I will not object to a working patch if you > have one. > Your patch block llvm 3.8.1, I'm not care about gcc 4.9.4, but llvm 3.8.1 depends gcc 4.
Re: WIP: GCC 6.2.0
On Tue, 6 Sep 2016 15:08:37 +0800, Dongsheng Song wrote: > On Sun, Sep 4, 2016 at 8:30 PM, Pascal Stumpfwrote: > > > Here's a preliminary port of GCC 6.2.0, tested only on amd64, i386 and > > sparc64 as I don't have access to anything else at the moment. > > > > Go ahead and play with it. > > > > Thanks for your great work ! > > But when I install gcc 6.2, it conflicted with gcc 4.9.4, I must remove gcc > 4.9.4 ( then llvm 3.8.1) at first. > > Could you support them coexist by the version specific technology? > That's quite a bit of work for questionable benefit, as we usually switch our whole ports tree to one specific version at once and then get rid of the old one. But I will not object to a working patch if you have one.
[NEW] x11/stumpw
Hello, I made a port for StumpWM. I would like some help to ensure that the port is fine. portscheck doesn't display any error at least. StumpWM needs some LISP libraries (cl-ppcre and clx) which are fetched by the port, I had to wrote a little wrapper (in patches/) named bootstrap-depends.lisp. Libraries are usually installed with the tool quicklisp, which could be compared to cpan for perl. There is absolutely no need to create a port for common lisp libraries, that's why I needed to create a script which basically tells sbcl where to look for the libraries during the build. There is a texinfo doc but I don't know what to do with it, it's not included in the package. I didn't used ${INSTALL_PROGRAM} macro because after stripping the binary, starting stumpwm would only start a vanilla sbcl (the binary should be around 60 Mb and after stripping it's only 620 kb). Thanks for reading regards PS : attachment missing in my previous mail stumpwm.tar.gz Description: GNU Zip compressed data
[NEW] x11/stumpw
Hello, I made a port for StumpWM. I would like some help to ensure that the port is fine. portscheck doesn't display any error at least. StumpWM needs some LISP libraries (cl-ppcre and clx) which are fetched by the port, I had to wrote a little wrapper (in patches/) named bootstrap-depends.lisp. Libraries are usually installed with the tool quicklisp, which could be compared to cpan for perl. There is absolutely no need to create a port for common lisp libraries, that's why I needed to create a script which basically tells sbcl where to look for the libraries during the build. There is a texinfo doc but I don't know what to do with it, it's not included in the package. I didn't used ${INSTALL_PROGRAM} macro because after stripping the binary, starting stumpwm would only start a vanilla sbcl (the binary should be around 60 Mb and after stripping it's only 620 kb). Thanks for reading regards
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: rob...@cvs.openbsd.org 2016/09/06 10:23:15 Removed files: editors/libreoffice/patches: patch-external_libpagemaker_UnpackedTarball_libpagemaker_mk patch-external_libpagemaker_patch-ldflags Log message: unbreak build
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 10:12:57 Modified files: net/bro: Makefile net/bro/patches: patch-aux_broctl_BroControl_options_py patch-src_broxygen_Manager_cc Added files: net/bro/patches: patch-aux_broccoli_test_broccoli-v6addrs_c patch-cmake_FindBIND_cmake Log message: Add a couple of build patches from Michael Shirk. While here, fix a warning.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jer...@cvs.openbsd.org 2016/09/06 09:38:41 Modified files: lang/ruby : Makefile Log message: Unhook ruby 2.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jer...@cvs.openbsd.org 2016/09/06 09:37:54 Modified files: devel/gitsh: Makefile mail/mew : Makefile misc/tpp : Makefile net/rrdtool: Makefile net/weechat: Makefile net/whatweb: Makefile x11/gnome/code-assistance: Makefile x11/kde4/korundum: Makefile x11/kde4/kross-interpreters: Makefile x11/kde4/ruby-qt: Makefile x11/kde4/webdev: Makefile x11/kde4/workspace: Makefile Log message: Bump ruby dependent ports after default ruby version switch from ruby 2.2 to ruby 2.3.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jer...@cvs.openbsd.org 2016/09/06 09:23:26 Modified files: lang/ruby : ruby.port.mk Log message: Make ruby 2.3 the default ruby version. Also, remove code allowing the building of ruby ports using ruby 2.0.
Re: Too many open files with graphics/shotwell
On Tue, Sep 06, 2016 at 12:36:54PM +0100, Stuart Henderson wrote: > I don't know shotwell but based on the messages I guess it is using some > kqueue-based monitoring to look for file changes, which uses a file > descriptor for each file monitored. If so, you will need a whole stack > of FDs available, or maybe you can point it at a smaller tree.. I'm not pointing to any tree. The fact of going to File -> Import From Folder... is enough to generate a core dump. No big dir selected. Sometimes, even Help -> About is able to trigger this. Thank you. -- db
Re: port for turnserver
On 2016/09/04 10:43, Stephen Graf wrote: > I have built a port for turnserver-4.5.0.3, > https://github.com/coturn/coturn. > > > > It needs an experienced maintainer to complete and verify. > > > > I have tested the port and am running the results on openBSD 5.9 stable. > There is a reported dependency problem but it does not seem to affect the > operation of the sub app involved. The license seems to be a scaled down > version of some other license, but requires the license file to be included > in any distribution. This app has been ported on FreeBSD. > Port submissions need to be for -current not -stable of an old release. Looking at the PLIST /usr/local/etc is wrong, you need to install to an examples dir, @sample into ${SYSCONFDIR}, and teach the software to look in ${SYSCONFDIR}. /usr/local/var(/db/turndb) is wrong. rc script is hardcoding paths which should come from variables (/etc -> ${SYSCONFDIR}, /usr/local -> ${PREFIX}, /var -> ${LOCALSTATEDIR}. Missing "# $OpenBSD$" rcsid comment in turnserver.rc. The uid you chose is already taken. Wouldn't telephony be better than net? The configure script is picking up various libraries if installed (postgresql, hiredis, mysqlclient, gettext) these need to be listed properly in WANTLIB/LIB_DEPENDS or disabled. DESCR shouldn't have the WWW line, that comes automatically from HOMEPAGE. The LICENSE/LICENSE_FILE stuff aren't OpenBSD, look at our other ports to see how this is done. The MASTER_SITES0/1 construct aren't useful here, it doesn't even use the files from extradocs. Zap the CVS dir in the tar.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2016/09/06 07:44:22 Modified files: mail/mailman : Makefile distinfo mail/mailman/patches: patch-Mailman_Defaults_py_in patch-Mailman_htmlformat_py mail/mailman/pkg: PLIST Log message: Update to 2.1.23; includes CVE-2016-6893 fix.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 07:33:35 Modified files: devel/quirks : Makefile devel/quirks/files: Quirks.pm Log message: Register rygel and vino removal.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 07:31:15 Modified files: meta/gnome : Makefile Log message: x11/gnome/vino and x11/gnome/rygel are gone.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 07:29:37 Modified files: x11/gnome/tracker: Makefile Log message: x11/gnome/rygel is no more.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 07:29:02 Modified files: x11/gnome : Makefile Log message: -rygel -vino
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 07:28:27 Removed files: x11/gnome/rygel: Makefile distinfo x11/gnome/rygel/pkg: DESCR PLIST x11/gnome/vino : Makefile distinfo x11/gnome/vino/pkg: DESCR PLIST Log message: Remove rygel and vino. Similarly to gnome-user-share, these will not work without systemd in the next GNOME release which is just around the corner. discussed with/ok jasper@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 07:18:42 Modified files: devel/libslang/patches: patch-autoconf_aclocal_m4 Log message: remove a sparc leftover
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 06:37:52 Modified files: x11/mplayer: Makefile Log message: Sync FFMPEG_V. from Brad (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 06:37:11 Modified files: graphics/ffmpeg: Makefile distinfo graphics/ffmpeg/patches: patch-configure patch-libavcodec_aaccoder_c patch-libavcodec_aacenc_c Log message: Update to ffmpeg-20160903. from Brad (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2016/09/06 06:10:05 Modified files: mail/courier-authlib: Makefile Log message: fix REVISION-userdb in courier-authlib
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 06:08:18 Modified files: www/puppetboard: Makefile Log message: another port sebastia@ takes over
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 06:07:42 Modified files: devel/ctftools : Makefile Added files: devel/ctftools/patches: patch-contrib_opensolaris_tools_ctf_cvt_ctfmerge_c Log message: Allow CTFMERGE_TERMINATE_NO_UNLINK=1 to keep the output file
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 05:55:15 Modified files: sysutils/ansible/patches: patch-lib_ansible_plugins_action_copy_py Log message: regen patch
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: dco...@cvs.openbsd.org 2016/09/06 05:50:52 Modified files: x11/xmobar : Makefile distinfo Log message: Update to xmobar-0.24.3
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2016/09/06 05:45:12 Modified files: graphics/png : Makefile distinfo graphics/png/pkg: PLIST Log message: update to libpng-1.6.25
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2016/09/06 05:43:12 Modified files: security/libsodium: Makefile Log message: WANTLIB+=pthread for libsodium, pointed out by nigel@
Re: Patch for fonts/terminus-font/Makefile
On 2016/09/04 22:29, Aioi Yuuko wrote: > I was actually considering that naming convention, with only one minor > caveat: it would break the current flavors. I see four possibilities: > a) keep symquotes and center_tilde, then add in flavors named directly after > the remaining upstream patches > b) custom names for everything > c) all names directly after the upstream patches > d) leave it alone and if anyone wants to really customize, let them edit the > makefile themselves I like option a), with the addition "add the new flavours to pkg/DESCR". The only thing is that then it's necessary to check that the various flavour combinations don't conflict (the existing two flavours are OK), and this would need to be re-checked for future updates.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 05:40:22 Modified files: sysutils/ansible: Makefile Log message: take maintainer suggested by and ok rpe@ (previous MAINTAINER)
Re: Too many open files with graphics/shotwell
On 2016/09/04 19:59, Daniel Bolgheroni wrote: > Running shotwell, it closes with whatever action tried inside the > program with: > > $ shotwell > > (shotwell:25255): Gtk-WARNING **: Theme parsing error: :2:38: The style > property GtkPaned:handle-size is deprecated and shouldn't be used anymore. It > will be removed in a future version > > (shotwell:25255): GLib-GIO-CRITICAL **: g_file_monitor_emit_event: assertion > 'G_IS_FILE_MONITOR (monitor)' failed > > (shotwell:25255): Gtk-WARNING **: Attempting to read the recently used > resources file at '/home/dbolgheroni/.local/share/recently-used.xbel', but > the parser failed: Failed to open file > '/home/dbolgheroni/.local/share/recently-used.xbel': Too many open files. > > (shotwell:25255): GLib-ERROR **: Creating pipes for GWakeup: Too many open > files I don't know shotwell but based on the messages I guess it is using some kqueue-based monitoring to look for file changes, which uses a file descriptor for each file monitored. If so, you will need a whole stack of FDs available, or maybe you can point it at a smaller tree..
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 05:34:11 Modified files: sysutils/mcollective-plugins/filemgr-agent: Makefile distinfo sysutils/mcollective-plugins/service-agent: Makefile distinfo sysutils/mcollective-plugins/sysctl-data: Makefile distinfo Log message: use GH_COMMIT instead of a mirrored tarball; no actual file changes
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2016/09/06 05:33:54 Modified files: www/syweb : Makefile www/syweb/pkg : PLIST Removed files: www/syweb/pkg : MESSAGE Log message: Fix syweb's path to rrdtool and mention PHP's use of /bin/sh in running programs. Prompted by a mail from thuban at yeuxdelibad.net and an old diff from Ralf Horstmann that I probably missed because it was sent to ports@ rather than the maintainer. Convert MESSAGE to README while there.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 05:31:24 Modified files: x11/gtk2mm : Makefile distinfo x11/gtk2mm/pkg : PLIST Removed files: x11/gtk2mm/patches: patch-configure_ac patch-gdk_src_color_hg patch-gdk_src_pixbuf_ccg patch-gdk_src_pixbuf_hg patch-gdk_src_pixbufanimationiter_hg patch-gdk_src_pixbufformat_hg patch-gdk_src_rectangle_hg patch-gdk_src_rgbcmap_hg patch-gdk_src_types_hg patch-gtk_gtkmm_border_h patch-gtk_src_entrycompletion_hg patch-gtk_src_enums_hg patch-gtk_src_main_hg patch-gtk_src_papersize_hg patch-gtk_src_printunixdialog_hg patch-gtk_src_textattributes_hg patch-gtk_src_treemodel_hg Log message: Update to gtk2mm-2.24.5.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 05:27:26 Modified files: databases/puppetdb: Makefile databases/ruby-hiera: Makefile Log message: sebastia@ takes over these too
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 05:27:50 Modified files: productivity/tryton/stock: Makefile distinfo Log message: Update to trytond-module-stock-3.2.9.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 05:27:00 Modified files: productivity/tryton/proteus: Makefile distinfo Log message: Update to proteus-3.2.8.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 05:25:58 Modified files: productivity/tryton/account_statement: Makefile distinfo Log message: Update to trytond-module-account_statement-3.2.3.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 05:26:08 Modified files: sysutils/ruby-facter: Makefile sysutils/ruby-puppet: Makefile.inc sysutils/ruby-puppet/4: Makefile databases/py-puppetdb: Makefile net/ruby-msgpack: Makefile net/ruby-stomp : Makefile sysutils/mcollective: Makefile sysutils/mcollective-plugins: Makefile.inc sysutils/mcollective-plugins/filemgr-agent: Makefile sysutils/mcollective-plugins/nrpe-agent: Makefile sysutils/mcollective-plugins/package-agent: Makefile sysutils/mcollective-plugins/puppet-agent: Makefile sysutils/mcollective-plugins/service-agent: Makefile sysutils/mcollective-plugins/sysctl-data: Makefile databases/puppetdb4: Makefile databases/ruby-hiera3: Makefile Log message: sebastia@ is the new puppetmaster
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 05:23:11 Modified files: productivity/tryton/account: Makefile distinfo Log message: Update to trytond-module-account-3.2.10.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 05:22:01 Modified files: productivity/tryton/tryton: Makefile distinfo productivity/tryton/tryton/pkg: PLIST Log message: Update to tryton-3.2.17.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: dco...@cvs.openbsd.org 2016/09/06 05:18:18 Modified files: audio/pianobar : Makefile audio/pianobar/files: cert.pem Log message: Sync with src/lib/libcrypto/cert.pem
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 05:16:20 Modified files: databases : Makefile sysutils : Makefile Log message: unhook puppet3 ports (ruby-hiera, puppetdb and ruby-facter); this also resolves conflicts with their puppet4 counterparts during dbp builds. these unhooked ports are scheduled for removal when puppet 3 itself leaves the stage
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 05:14:51 Modified files: sysutils/ruby-puppet: Makefile Makefile.inc Log message: stop building puppet 3, effectively making puppet 4 the default on OpenBSD. puppet 3 is scheduled to be removed from the tree after 6.1. for more information on migrating to Puppet 4, please refer to [1] as a starting point. however note we don't have a port for puppetserver yet, it's a work in progress, see upstream SERVER-14 for more information. However the ruby puppetmaster still works fine for us. 1: https://docs.puppet.com/puppet/latest/reference/upgrade_major_pre.html as discussed with and OK sebastia@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 05:14:04 Modified files: www/webkitgtk4 : Tag: OPENBSD_6_0 Makefile distinfo Log message: Update to webkitgtk4-2.12.5 (sync with their latest stable release). - fixes several crashes and hangs
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 05:04:58 Modified files: www/webkitgtk4 : Makefile distinfo Log message: Update to webkitgtk4-2.12.5. - fixes several crashes and hangs
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 05:02:00 Modified files: databases/ruby-hiera-eyaml: Makefile databases/ruby-hiera-file: Makefile Log message: puppet 4 will soon be the default puppet version, so switch these ports to the matching hiera port. discussed with and ok sebastia@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2016/09/06 04:31:12 Modified files: infrastructure/mk: bsd.port.mk Log message: dismantle SIGNING_PARAMETERS, no longer possible nor advisable to sign stuff on the fly
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: pas...@cvs.openbsd.org 2016/09/06 04:14:24 Modified files: lang/gcc/4.9 : Makefile distinfo Log message: regen i386 bootstrap
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 03:50:24 Modified files: sysutils/ruby-puppet/4: Makefile sysutils/ruby-puppet/4/pkg: PLIST Log message: @sample /var/puppetlabs/, from semarie@
Re: MariaBD 10.0.27 OpenBSD_6_0 update
Il giorno Martedi, 6 Settembre, 2016 04:19 CEST, Ian McWilliamha scritto: > The latest update to MariaDB in OpenBSD_6_0 stable tree fails to build. > > Need to remove one patch file. > yes, please unbreak. ok giovanni@ Cheers Giovanni > ===> Extracting for mariadb-10.0.27 > ===> Patching for mariadb-10.0.27 > Ignoring previously applied (or reversed) patch. > 1 out of 1 hunks ignored--saving rejects to > storage/innobase/include/os0sync.h.rej > ***> patch-storage_innobase_include_os0sync_h did not apply cleanly > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2596 > '/usr/ports/pobj/mariadb-10.0.27/.patch_done': @if cd > /usr/ports/databases/m...) > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:1891 > '/usr/packages/amd64/all/mariadb-client-10.0.27v1.tgz') > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2416 > '_internal-package') > *** Error 1 in /usr/ports/databases/mariadb > (/usr/ports/infrastructure/mk/bsd.port.mk:2396 'package') > > > Ian McWilliam > > Index: patches/patch-storage_innobase_include_os0sync_h > === > RCS file: patches/patch-storage_innobase_include_os0sync_h > diff -N patches/patch-storage_innobase_include_os0sync_h > --- patches/patch-storage_innobase_include_os0sync_h4 May 2016 07:37:22 > - 1.4 > +++ /dev/null 1 Jan 1970 00:00:00 - > @@ -1,17 +0,0 @@ > -$OpenBSD: patch-storage_innobase_include_os0sync_h,v 1.4 2016/05/04 07:37:22 > sthen Exp $ > - > -InnoDB attempted to verify if __sync_lock_test_and_set() is available by > -checking IB_STRONG_MEMORY_MODEL. This macro has nothing to do with > availability > -of __sync_lock_test_and_set(), the right one is HAVE_ATOMIC_BUILTINS. > - > storage/innobase/include/os0sync.h.origMon May 2 06:19:43 2016 > -+++ storage/innobase/include/os0sync.h Mon May 2 06:20:19 2016 > -@@ -466,7 +466,7 @@ amount to decrement. */ > - # define os_atomic_decrement_uint64(ptr, amount) \ > - os_atomic_decrement(ptr, amount) > - > --# if defined(IB_STRONG_MEMORY_MODEL) > -+# if defined(HAVE_ATOMIC_BUILTINS) > - > - /** Do an atomic test and set. > - @param[in,out]ptr Memory location to set to non-zero
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2016/09/06 03:04:48 Modified files: databases/ruby-hiera-eyaml: Makefile databases/ruby-hiera-file: Makefile databases/ruby-hiera: Makefile databases/ruby-hiera3: Makefile devel/ruby-rgen: Makefile net/ruby-msgpack: Makefile sysutils/ruby-augeas: Makefile sysutils/ruby-shadow: Makefile Log message: drop MODRUBY_REV on these ports; jeremy@ pointed out that with 'CONFIGURE_STYLE=ruby gem' it's not needed and dependant ports will require the correct flavor/version
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: dco...@cvs.openbsd.org 2016/09/06 03:00:19 Modified files: x11/qt4: Makefile x11/qt5: Makefile Log message: USE_WXNEEDED, add "-Wl,-z,wxneeded" to Qt5WebKit and QtWebKit pkg-config files. Similar commit to what was done to www/webkit, and www/webkitgtk4. OK ajacoutot@, landry@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: dco...@cvs.openbsd.org 2016/09/06 02:50:50 Modified files: mail/trojita : Makefile www/otter-browser: Makefile Log message: Remove USE_WXNEEDED, it's handled by cmake now
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: dco...@cvs.openbsd.org 2016/09/06 02:43:54 Modified files: devel/cmake: Makefile devel/cmake/patches: patch-Source_cmComputeLinkInformation_cxx Log message: Hack cmake so that if webkit is in the list of shared libraries that are passed to the linker, automatically add '-Wl,-z,wxneeded' to the linker flags. Works for libQt5WebKit, libQtWebKit, libwebkit2gtk, libwebkitgtk. Requested by, and ok, ajacoutot@ Also ok landry@
Re: cmake, webkit and '-Wl,-z,wxneeded'
On Tue, Sep 6, 2016 at 10:17 AM, Landry Breuilwrote: > On Mon, Sep 05, 2016 at 03:13:21PM +0100, Antoine Jacoutot wrote: >> On September 5, 2016 12:50:44 PM GMT+01:00, David Coppa >> wrote: >> > >> >Hi! >> > >> >Unfortunately, I did not found a single point where to insert my >> >hack without touching the C++ code. >> > >> >So here's the "fix" I came up with. I've hacked it inside the >> >function used by CMake to compute the list of shared libraries >> >that are passed to the linker. > > Ugh, that's a bit gross, but if it works there's really no way to > fix that via the Find.cmake files ? Started with this, ended up patching a handful of FindXXX.cmake files... It was not reliable. And what about local FindXXX.cmake modules from ${WRKSRC}/cmake/ ? >> >This works in the same way for QtWebKit, Qt5WebKit, and webkitgtk. >> > >> >So far I've tested audio/mscore, cad/xtrkcad, mail/geary, mail/trojita, >> >net/owncloudclient, www/midori, and www/otter-browser and the trick >> >seems to work... > > Seems like enough testing for me, ok. We can still devise a better fix > someday... Sure, I'm still thinking about a better way to achieve this goal. This works reasonably well for the moment... Ciao! David -- "If you try a few times and give up, you'll never get there. But if you keep at it... There's a lot of problems in the world which can really be solved by applying two or three times the persistence that other people will." -- Stewart Nelson
Re: Fix QtWebKit.pc
On Mon, Sep 05, 2016 at 04:37:50PM +0200, David Coppa wrote: > On Mon, 05 Sep 2016, David Coppa wrote: > > > On Mon, Sep 5, 2016 at 4:12 PM, Antoine Jacoutot> > wrote: > > > On September 5, 2016 2:33:30 PM GMT+01:00, David Coppa > > > wrote: > > >> > > >> > > >> For ports linking to QtWebKit via pkg-config (not related to cmake), > > >> if they exist... > > >> > > >> Index: Makefile > > >> > > >> > > >> RCS file: /cvs/ports/x11/qt4/Makefile,v > > >> retrieving revision 1.135 > > >> diff -u -p -r1.135 Makefile > > >> --- Makefile 18 Mar 2016 21:38:26 - 1.135 > > >> +++ Makefile 5 Sep 2016 13:32:17 - > > >> @@ -23,7 +23,7 @@ PKGNAME-main = qt4-${PKGVERSION} > > >> PKGNAME-debug = qt4-debug-${PKGVERSION} > > >> FULLPKGNAME-html = qt4-html-${PKGVERSION} > > >> FULLPKGPATH-html = ${BASE_PKGPATH},-html > > >> -REVISION-main = 6 > > >> +REVISION-main = 7 > > >> > > >> # XXX qmake include parser is bogus > > >> DPB_PROPERTIES = parallel nojunk > > >> @@ -262,6 +262,8 @@ post-install: > > >> perl -pi.bak -e 's,-L${WRKBUILD}/lib,,g;' $$i; \ > > >> rm $$i.bak; \ > > >> done > > >> + @sed -i -e s/"-lQtWebKit"/"-lQtWebKit -Wl,-z,wxneeded"/ \ > > >> + ${QT_PKGCFGDIR}/QtWebKit.pc > > >> # this helps qdbuscpp2xml > > >> cd ${QT_BINDIR} > > >> && ln -sf moc moc4 > > >> .for l v in ${SHARED_LIBS} > > >> > > > > > > I'm ok with it but don't you also want to set USE_WXNEEDED ? > > Index: Makefile > === > RCS file: /cvs/ports/x11/qt4/Makefile,v > retrieving revision 1.135 > diff -u -p -r1.135 Makefile > --- Makefile 18 Mar 2016 21:38:26 - 1.135 > +++ Makefile 5 Sep 2016 14:37:33 - > @@ -1,5 +1,7 @@ > # $OpenBSD: Makefile,v 1.135 2016/03/18 21:38:26 naddy Exp $ > > +USE_WXNEEDED = Yes > + > ONLY_FOR_ARCHS = ${GCC4_ARCHS} > BROKEN-hppa =libQtWebKit.so.3.0: undefined reference to > __sync_sub_and_fetch_4 > > @@ -23,7 +25,7 @@ PKGNAME-main = qt4-${PKGVERSION} > PKGNAME-debug = qt4-debug-${PKGVERSION} > FULLPKGNAME-html = qt4-html-${PKGVERSION} > FULLPKGPATH-html = ${BASE_PKGPATH},-html > -REVISION-main = 6 > +REVISION-main = 7 > > # XXX qmake include parser is bogus > DPB_PROPERTIES = parallelnojunk > @@ -262,6 +264,8 @@ post-install: > perl -pi.bak -e 's,-L${WRKBUILD}/lib,,g;' $$i; \ > rm $$i.bak; \ > done > + @sed -i -e s/"-lQtWebKit"/"-lQtWebKit -Wl,-z,wxneeded"/ \ > + ${QT_PKGCFGDIR}/QtWebKit.pc > # this helps qdbuscpp2xml > cd ${QT_BINDIR} && ln -sf moc moc4 > .for l v in ${SHARED_LIBS} Ok for both qt4/qt5 fixes, thanks for working on it!
Re: cmake, webkit and '-Wl,-z,wxneeded'
On Mon, Sep 05, 2016 at 03:13:21PM +0100, Antoine Jacoutot wrote: > On September 5, 2016 12:50:44 PM GMT+01:00, David Coppa> wrote: > > > >Hi! > > > >Unfortunately, I did not found a single point where to insert my > >hack without touching the C++ code. > > > >So here's the "fix" I came up with. I've hacked it inside the > >function used by CMake to compute the list of shared libraries > >that are passed to the linker. Ugh, that's a bit gross, but if it works there's really no way to fix that via the Find.cmake files ? > >This works in the same way for QtWebKit, Qt5WebKit, and webkitgtk. > > > >So far I've tested audio/mscore, cad/xtrkcad, mail/geary, mail/trojita, > >net/owncloudclient, www/midori, and www/otter-browser and the trick > >seems to work... Seems like enough testing for me, ok. We can still devise a better fix someday...
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2016/09/06 01:50:43 Modified files: misc/shared-mime-info: Makefile distinfo misc/shared-mime-info/pkg: PLIST Log message: Update to shared-mime-info-1.7.
Re: WIP: GCC 6.2.0
On Sun, Sep 4, 2016 at 8:30 PM, Pascal Stumpfwrote: > Here's a preliminary port of GCC 6.2.0, tested only on amd64, i386 and > sparc64 as I don't have access to anything else at the moment. > > Go ahead and play with it. > Thanks for your great work ! But when I install gcc 6.2, it conflicted with gcc 4.9.4, I must remove gcc 4.9.4 ( then llvm 3.8.1) at first. Could you support them coexist by the version specific technology?