Forcing reinstall of a port
It seems like I might have run across this, a way to uninstall a port and reinstall it without having to remove all of its dependencies and put them back in the process. This is related to my pasting from Abiword which at the moment seems to be a gettext/libiconv order thing. This is from the gnu libiconv page at http://www.gnu.org/software/libiconv/ and sounds ominous: - After installing GNU libiconv for the first time, it is recommended to recompile and reinstall GNU gettext, so that it can take advantage of libiconv. On systems other than GNU/Linux, the iconv program will be internationalized only if GNU gettext has been built and installed before GNU libiconv. This means that the first time GNU libiconv is installed, we have a circular dependency between the GNU libiconv and GNU gettext packages, which can be resolved by building and installing either first libiconv, then gettext, then libiconv again, or (on systems supporting shared libraries, excluding AIX) first gettext, then libiconv, then gettext again. Recall that before building a package for the second time, you need to erase the traces of the first build by running "make distclean". -- Both gettext and libiconv seem to have about 350 depencies on my box, in /var/db/pkg libiconv has the earlier date. I thought there was a "force" option somewhere. Abiword seems to be the only thing affected and I've got 924 ports installed (yes, ports, not packages, from distfiles). This is on OpenBSD 5.7 but I have the same pasting problem on 5.6 and 5.8 uses the same Abiword version (3.0.1). -- Credit is the root of all evil. - AB1JX
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2015/10/23 03:28:58 Log message: import ports/net/mlvpn, ok ajacoutot@ MLVPN is VPN software that encrypts data with the Salsa20 stream cipher and uses Ed25519 authentication. It can either run with a single ISP connection, or can aggregate multiple circuits with active monitoring to identify and remove faulty ones without dropping TCP connections. To avoid problems with out-of-order packet delivery, it can optionally add delays per-connection or buffer packets before delivering to the tun/tap interface. MLVPN uses privilege separation and supports Linux, OpenBSD and FreeBSD. Status: Vendor Tag: sthen Release Tags: sthen_20151023 N ports/net/mlvpn/Makefile N ports/net/mlvpn/distinfo N ports/net/mlvpn/pkg/PLIST N ports/net/mlvpn/pkg/DESCR N ports/net/mlvpn/pkg/mlvpn.rc No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: z...@cvs.openbsd.org2015/10/23 04:04:47 Modified files: www/netsurf/browser: Makefile Added files: www/netsurf/browser/patches: patch-gtk_res_netsurf-gtk_desktop Log message: Fix desktop file, allowing to actually run netsurf as we call it. Original report from Roman Yakovlev. okay bentley@ (maintainer).
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2015/10/23 05:52:31 Modified files: devel/py-pathlib: Makefile Log message: s/sphinx/py-sphinx
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2015/10/23 08:21:16 Modified files: textproc/libxml: Makefile textproc/libxml/patches: patch-parser_c Log message: Security fixes for CVE-2015-7941 and CVE-2015-7942
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2015/10/23 08:22:13 Modified files: textproc/libxml: Tag: OPENBSD_5_8 Makefile textproc/libxml/patches: Tag: OPENBSD_5_8 patch-parser_c Log message: Security fixes for CVE-2015-7941 and CVE-2015-7942
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2015/10/23 09:15:28 Modified files: emulators/qemu : Makefile Added files: emulators/qemu/patches: patch-net_tap-bsd_c Log message: Adjust Qemu following the split of tun(4)'s "link0" mode into tap(4).
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2015/10/23 09:12:43 Modified files: x11/gnome/eog-plugins: Makefile distinfo Log message: update to eog-plugins-3.16.2
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2015/10/23 09:12:45 Modified files: x11/gnome/eog : Makefile distinfo x11/gnome/eog/pkg: PLIST Log message: update to eog-3.18.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2015/10/23 09:34:57 Modified files: x11/gnome/empathy: Makefile distinfo x11/gnome/empathy/pkg: PLIST Log message: update to empathy-3.12.11
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: feine...@cvs.openbsd.org2015/10/23 09:27:16 Modified files: print/gl2ps: Makefile distinfo Log message: Update to GL2PS 1.3.9 "Diff reads fine" pirofti@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: giova...@cvs.openbsd.org2015/10/23 09:48:32 Modified files: devel/p5-Log-Any: Makefile distinfo devel/p5-Log-Any/pkg: PLIST Log message: Update to 1.032 and take maintainership ok David Hill (old maintainer)
Re: Firefox 41.0.2 with W^X
On 22/10/15(Thu) 21:40, Amit Kulkarni wrote: > On Thu, Oct 22, 2015 at 12:26 PM, David Coppawrote: > > > On Thu, Oct 22, 2015 at 3:45 PM, Ed Ahlsen-Girard > > wrote: > > > I have noticed a performance hit since the switch was flipped. Firefox > > > stays at the top of top most of the time, and its CPU percentages have > > > spiked to 175% if multiple tabs were being opened. dmesg below the sig. > > > > Can you try if the attached patch is an improvement? > > > > Hi, > > This CPU spike is present with October 11 packages (Firefox 41.0.1) on > amd64, so it will be difficult to isolate the performance impact of the W > ^X vs the existing situation. FWIW I found that firefox is (ab)using pthread_mutex_trylock(3) a lot resulting in a storm of sched_yield(2) triggering a lot (dozen to hundreds of thousands) of IPIs on my x220. I tried to look at the source code but couldn't figure out where the call of pthread_mutex_trylock(3) are coming from. Firefox is just a monster. I'm sorry but I agree that if nobody is taking care of this regression it will be really hard to measure the impact of the W^X change.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2015/10/23 03:29:40 Modified files: infrastructure/db: user.list Log message: +mlvpn
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2015/10/23 03:30:26 Modified files: net: Makefile Log message: +mlvpn
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2015/10/23 05:21:09 Modified files: www/asql : Makefile distinfo Log message: update to asql-1.7, remove pea@ as maintainer at his request
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: bl...@cvs.openbsd.org 2015/10/23 04:45:39 Modified files: devel/p5-BSD-Resource: Makefile distinfo Log message: update p5-BSD-Resource to 1.2908
NEW: devel/p5-Search-Elasticsearch
Search::Elasticsearch - The official client for Elasticsearch The perl module and some new dependencies, comments ? ok ? Cheers Giovanni p5-Search-Elasticsearch-1.99.tgz Description: application/compressed-tar p5-Any-URI-Escape-0.01.tgz Description: application/compressed-tar p5-Hijk-0.24.tgz Description: application/compressed-tar
Re: Forcing reinstall of a port
When I set LC_CTYPE to en-US.UTF-8 as has been mentioned here and in the OpenBSD FAQ, then do something like type uxterm into an rxvt window I get the warning: Warning: locale not supported by Xlib, locale set to C So could the Xlib version be causing this? I'm floundering. Maybe uxterm's trying to set a locale on startup. On 10/23/15, Stuart Hendersonwrote: > On 2015/10/23 14:48, Alan Corey wrote: >> This is related to my pasting from Abiword which at the moment seems >> to be a gettext/libiconv order thing. > > I think this is probably an Abiword bug. They've had problems in this > area before, and even just copying and pasting entirely within Abiword > doesn't work for me (-current amd64). > > -- Credit is the root of all evil. - AB1JX
Re: Forcing reinstall of a port
No, I get that if I type rxvt into rxvt. I'm not using xconsole but it might show there too. But I don't get it if I type firefox or abiword, ufraw, blender, filezilla, etc. into rxvt. On 10/23/15, Alan Coreywrote: > When I set LC_CTYPE to en-US.UTF-8 as has been mentioned here and in > the OpenBSD FAQ, then do something like type uxterm into an rxvt > window I get the warning: > > Warning: locale not supported by Xlib, locale set to C > > So could the Xlib version be causing this? I'm floundering. Maybe > uxterm's trying to set a locale on startup. > > On 10/23/15, Stuart Henderson wrote: >> On 2015/10/23 14:48, Alan Corey wrote: >>> This is related to my pasting from Abiword which at the moment seems >>> to be a gettext/libiconv order thing. >> >> I think this is probably an Abiword bug. They've had problems in this >> area before, and even just copying and pasting entirely within Abiword >> doesn't work for me (-current amd64). >> >> > > > -- > Credit is the root of all evil. - AB1JX > -- Credit is the root of all evil. - AB1JX
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: abie...@cvs.openbsd.org 2015/10/23 20:46:47 Modified files: lang/node : Makefile distinfo lang/node/patches: patch-common_gypi patch-deps_npm_node_modules_node-gyp_lib_configure_js patch-deps_npm_node_modules_node-gyp_lib_install_js patch-deps_v8_Makefile patch-deps_v8_tools_gyp_v8_gyp patch-node_gyp lang/node/pkg : PLIST Added files: lang/node/patches: patch-tools_test_py Removed files: lang/node : node.port.mk lang/node/patches: patch-deps_npm_bin_node-gyp-bin_node-gyp patch-deps_uv_src_unix_openbsd_c patch-deps_v8_build_common_gypi patch-deps_v8_src_heap_cc patch-lib_module_js patch-tools_gyp_node_py Log message: Finally update node to the latest version, lots of changes: - Use internal OpenSSL (no LibreSSL :( ) - Remove node.port.mk as there are no longer any modules in ports. - Remove various patches that are now upstream. - Fix some tests by telling test.py the fully qualified path to node - Add gcc as a RUN dep for building native node modules. - Clean up Makefile a bit. OK juanfra@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2015/10/23 23:17:33 Modified files: multimedia/libmatroska: Makefile distinfo Log message: Update to libmatroska-1.4.3. from Brad (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2015/10/23 23:16:41 Modified files: textproc/libebml: Makefile distinfo Log message: Update to libebml-1.3.2. from Brad (maintainer)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2015/10/23 22:22:06 Modified files: misc/hwdata: Makefile distinfo Log message: Update to hwdata-0.283.
NEW: textproc/wkhtmltopdf
This is my first attempt at writing a port. wkthmltopdf is a program that uses Webkit to convert an HTML file to a PDF or image file. More information can be found on their website: http://wkhtmltopdf.org I use it in a number of Ruby on Rails applications with the wickedpdf gem to generate PDF invoices for example. Note that upstream advises to build it using their patched QT in order to support the following options: * Printing more then one HTML document into a PDF file. * Running without an X11 server. * Adding a document outline to the PDF file. * Adding headers and footers to the PDF file. * Generating a table of contents. * Adding links in the generated PDF file. * Printing using the screen media-type. * Disabling the smart shrink feature of webkit. My port doesn't use their patched QT as I find it quite ugly. However, some of these options are pretty handy on servers, especially running without an X11 server. What are your thoughts and comments on this port? Frank wkhtmltopdf-port-0.12.2.4p0.tar.gz Description: application/gzip
Re: Forcing reinstall of a port
On 2015-10-23, Alan Coreywrote: > This is related to my pasting from Abiword which at the moment seems > to be a gettext/libiconv order thing. That seems unlikely. > On systems other than GNU/Linux, the iconv program will be > internationalized only if GNU gettext has been built and installed > before GNU libiconv. Yep, so the iconv(1) from our port does not have internationalization. That means no German, French, etc. error messages, but it doesn't affect iconv(1)'s or iconv(3)'s actual functionality. > Both gettext and libiconv seem to have about 350 depencies on my box, You are using the term "dependencies" wrong. Those 350 packages have gettext and libiconv as dependencies. I don't think we have a well-established term for the other direction--"dependents" would make sense. -- Christian "naddy" Weisgerber na...@mips.inka.de
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2015/10/23 14:34:55 Modified files: databases/puppetdb: Makefile distinfo Log message: update to puppetdb-2.3.8
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2015/10/23 14:36:05 Modified files: devel/ruby-rspec/specinfra: Makefile distinfo Log message: update to specinfra-2.44.1
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2015/10/23 14:48:34 Modified files: sysutils/entr : Makefile distinfo sysutils/entr/pkg: DESCR Log message: update to entr-3.3, from Fred (and similar to my 3.2 update diff which I had an OK from the maintainer for earlier but forgot to commit)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2015/10/23 08:59:29 Modified files: x11/gnome/online-accounts: Makefile distinfo Log message: update to gnome-online-accounts-3.18.1
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2015/10/23 08:38:55 ports/www/zendframework/patches Update of /cvs/ports/www/zendframework/patches In directory cvs.openbsd.org:/tmp/cvs-serv7026/patches Log Message: Directory /cvs/ports/www/zendframework/patches added to the repository
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2015/10/23 09:10:37 Modified files: net/openvpn: Makefile net/openvpn/patches: patch-src_openvpn_tun_c Log message: Adjust OpenVPN following the split of tun(4)'s "link0" mode into tap(4).
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2015/10/23 09:21:43 Modified files: x11/gnome/grilo-plugins: Makefile distinfo Log message: update to grilo-plugins-0.2.16
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2015/10/23 09:24:47 Modified files: x11/gnome/system-monitor: Makefile distinfo x11/gnome/system-monitor/pkg: PLIST Removed files: x11/gnome/system-monitor/patches: patch-configure Log message: update to gnome-system-monitor-3.18.0.1
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2015/10/23 09:23:16 Modified files: graphics/evince: Makefile distinfo graphics/evince/pkg: PLIST Removed files: graphics/evince/patches: patch-libview_ev-view_c Log message: update to evince-3.18.1
Re: NEW: audio/moc
On Mon, Oct 19, 2015 at 12:57:39PM +0200, Dmitrij D. Czarkoff wrote: > Michael Seyfert said: > > This is a port for the music on console (moc) music player. > > I have tested it on amd64, using it as my primary music player. > > > > I patched out most of the strcpy and scanf. > > Please let me know what you think of it. > > Several comments follow. I didn't try building this port, and didn't > look closely into patches. > > In moc/Makefile: > > | # License: GPLv3+ > | PERMIT_PACKAGE_CDROM = Yes > | PERMIT_PACKAGE_FTP = Yes > | PERMIT_DISTFILES_FTP = Yes > > PERMIT_PACKAGE_CDROM=Yes implies PERMIT_PACKAGE_FTP=Yes and > PERMIT_DISTFILES_FTP=Yes. OK, fixed this. > > | CONFIGURE_STYLE = gnu autoconf > > autoconf implies gnu. > changed to just 'gnu'. > | CONFIGURE_ENV = \ > | CFLAGS="-I${LOCALBASE}/include > -I${LOCALBASE}/include/db4" \ > | LDFLAGS=-L${LOCALBASE}/lib > > You should use > > CFLAGS += -I${LOCALBASE}/include -I${LOCALBASE}/include/db4 > LDFLAGS += -L${LOCALBASE}/lib > > The way you do it shadows user settings. > Looking at the other ports they all modify LDFLAGS the way Stuart said, so I changed it following his suggestion: CONFIGURE_ENV = \ CPPFLAGS="-I${LOCALBASE}/include -I${LOCALBASE}/include/db4" \ LDFLAGS="-L${LOCALBASE}/lib" > In moc/patches/patch-mocp_1 you should use ${LOCALBASE} here and add > something like this to makefile: > > post-build: > ${SUBST_CMD} ${WRKSRC}/mocp.1 > Fixed. > -- > Dmitrij D. Czarkoff > Attached is another try at this port. I patched every strcpy / sprintf that was warned about by the compiler regardless of triviality. Is this a bad idea? It probably doesn't hurt anything, but gives extra complexity to the patches/ folder. This should go upstream to the official code for the next moc release. -- Michael Seyfertmoc-openbsd-port-try2-2.5.0.tgz Description: application/tar-gz
Re: Forcing reinstall of a port
On 2015/10/23 14:48, Alan Corey wrote: > This is related to my pasting from Abiword which at the moment seems > to be a gettext/libiconv order thing. I think this is probably an Abiword bug. They've had problems in this area before, and even just copying and pasting entirely within Abiword doesn't work for me (-current amd64).
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: sebas...@cvs.openbsd.org2015/10/23 01:53:46 Modified files: www/sogo : Makefile www/sogo/patches: patch-Scripts_sql-update-2_2_17_to_2_3_0-mysql_sh patch-Scripts_sql-update-2_2_17_to_2_3_0_sh Log message: Fix the DB schema update scripts, that somehow got lost with the update to 2.3.2
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: dco...@cvs.openbsd.org 2015/10/23 01:53:06 Modified files: audio/mpd : Makefile Added files: audio/mpd/patches: patch-src_decoder_plugins_FfmpegDecoderPlugin_cxx Log message: ffmpeg decoder: avoid a double free error while decoding a stream bug reported and fix tested by "J. Scott Heppler"
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ajacou...@cvs.openbsd.org 2015/10/23 01:55:32 Modified files: x11/gnome/initial-setup: Makefile distinfo x11/gnome/initial-setup/patches: patch-gnome-initial-setup_Makefile_in patch-gnome-initial-setup_pages_Makefile_in patch-gnome-initial-setup_pages_account_Makefile_in patch-gnome-initial-setup_pages_password_Makefile_in x11/gnome/initial-setup/pkg: PLIST Log message: Update to gnome-initial-setup-3.18.0.
Re: Web content management stack
On 2015/10/23 13:08, Jason Tubnor wrote: > Hi, > > Does anyone have a combination of packages/ports to build a web content > management solution for business internet that they could recommend? I'd > like to use relayd if possible but that is not something set in stone. > Squid, squidguard and clamav have come to mind but wonder what else others > use. > > Scope of the project is to filter internet traffic for users, reducing > malicious downloads coming in and preventing access to adult and gambling > content. I do have access to push certificates to end-users machines. > > Any software pointers or different FAQ guides will be greatly appreciated. I'm using squid, c-icap, squidclamav and squid ACLs. I haven't got round to doing TLS MITM in production yet but really need to as a significant percentage of traffic is now encrypted, the implementation of this in squid 3.5 is quite well-featured and can be configured to do things like avoid MITMing certain sites, for example you might want to do this for online banking sites. It can also copy attributes where possible from the real server certificate so that in many cases browsers will use their normal UI for presenting cert errors. (obviously not for CA issues, but for things like bad dates). Squid native ACLs have improved a lot since the time squidguard was written, you're better off using these rather than squidguard if possible. If you need more than can be done with these then better to look at ufdbGuard than squidguard, it's not committed but I have a port lying around aomewhere.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: dco...@cvs.openbsd.org 2015/10/23 02:15:51 Modified files: audio/mpd : Tag: OPENBSD_5_8 Makefile Added files: audio/mpd/patches: Tag: OPENBSD_5_8 patch-src_decoder_plugins_FfmpegDecoderPlugin_cxx patch-src_encoder_plugins_FlacEncoderPlugin_cxx patch-src_tag_ApeLoader_cxx Log message: MFC: flac encoder: fix crash with 32 bit playback (upstream git commit 4e3d1821895c92d060fc3a9059e83ac58c8c79a3) tag/ApeLoader: fix buffer overflow after unterminated key (upstream git commit 205fba74cb9df985cdf928101633ffc41772) ffmpeg decoder: avoid a double free error while decoding a stream