CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: uebay...@cvs.openbsd.org2016/03/28 23:10:55 Modified files: devel/valgrind : Makefile distinfo Removed files: devel/valgrind/patches: patch-coregrind_m_syswrap_syswrap-openbsd_c patch-include_vki_vki-scnums-openbsd_h Log message: valgrind-3.10.1p9: Not only recognize pledge(2) but also skip it so Valgrind can call syscalls implicitly.
UPDATE: lang/go 1.6
I just ran "make test" on amd64, so any additional tests are welcome. OK? Index: Makefile === RCS file: /cvs/ports/lang/go/Makefile,v retrieving revision 1.28 diff -u -p -r1.28 Makefile --- Makefile14 Jan 2016 12:11:50 - 1.28 +++ Makefile29 Mar 2016 01:26:05 - @@ -4,7 +4,7 @@ ONLY_FOR_ARCHS =${GO_ARCHS} COMMENT = Go programming language -VERSION = 1.5.3 +VERSION = 1.6 EXTRACT_SUFX = .src.tar.gz DISTNAME = go${VERSION} PKGNAME = go-${VERSION} Index: distinfo === RCS file: /cvs/ports/lang/go/distinfo,v retrieving revision 1.15 diff -u -p -r1.15 distinfo --- distinfo14 Jan 2016 12:11:50 - 1.15 +++ distinfo29 Mar 2016 01:26:05 - @@ -1,2 +1,2 @@ -SHA256 (go1.5.3.src.tar.gz) = dU4G2rHDGrFo/J254yWWc0AV6p4kvETK5/I39BfOTv4= -SIZE (go1.5.3.src.tar.gz) = 12057623 +SHA256 (go1.6.src.tar.gz) = qWzOjOQ6m/mypMfUcLx+4MsAQQ2oFZgGgcg1Mhjc8UY= +SIZE (go1.6.src.tar.gz) = 12613308 Index: pkg/PLIST === RCS file: /cvs/ports/lang/go/pkg/PLIST,v retrieving revision 1.13 diff -u -p -r1.13 PLIST --- pkg/PLIST 5 Dec 2015 05:01:24 - 1.13 +++ pkg/PLIST 29 Mar 2016 01:26:05 - @@ -74,6 +74,7 @@ go/doc/go1.2.html go/doc/go1.3.html go/doc/go1.4.html go/doc/go1.5.html +go/doc/go1.6.html go/doc/go1.html go/doc/go1compat.html go/doc/go_faq.html @@ -182,8 +183,14 @@ go/misc/cgo/errors/ go/misc/cgo/errors/err1.go go/misc/cgo/errors/err2.go go/misc/cgo/errors/err3.go +go/misc/cgo/errors/issue11097a.go +go/misc/cgo/errors/issue11097b.go +go/misc/cgo/errors/issue13129.go +go/misc/cgo/errors/issue13423.go +go/misc/cgo/errors/issue13635.go go/misc/cgo/errors/issue7757.go go/misc/cgo/errors/issue8442.go +go/misc/cgo/errors/ptr.go go/misc/cgo/errors/test.bash go/misc/cgo/gmp/ go/misc/cgo/gmp/fib.go @@ -221,6 +228,7 @@ go/misc/cgo/test/callback_c_gccgo.c go/misc/cgo/test/cflags.go go/misc/cgo/test/cgo_linux_test.go go/misc/cgo/test/cgo_test.go +go/misc/cgo/test/cgo_unix_test.go go/misc/cgo/test/cthread.go go/misc/cgo/test/cthread_unix.c go/misc/cgo/test/cthread_windows.c @@ -228,12 +236,18 @@ go/misc/cgo/test/duplicate_symbol.go go/misc/cgo/test/env.go go/misc/cgo/test/exports.go go/misc/cgo/test/fpvar.go +go/misc/cgo/test/gcc68255/ +go/misc/cgo/test/gcc68255.go +go/misc/cgo/test/gcc68255/a.go +go/misc/cgo/test/gcc68255/c.c +go/misc/cgo/test/gcc68255/c.h go/misc/cgo/test/helpers.go go/misc/cgo/test/issue10303.go go/misc/cgo/test/issue11925.go go/misc/cgo/test/issue12030.go go/misc/cgo/test/issue1222.go go/misc/cgo/test/issue1328.go +go/misc/cgo/test/issue13402.go go/misc/cgo/test/issue1560.go go/misc/cgo/test/issue1635.go go/misc/cgo/test/issue2462.go @@ -245,6 +259,7 @@ go/misc/cgo/test/issue3729w.go go/misc/cgo/test/issue3741.go go/misc/cgo/test/issue3775.go go/misc/cgo/test/issue3945.go +go/misc/cgo/test/issue4029.c go/misc/cgo/test/issue4029.go go/misc/cgo/test/issue4029w.go go/misc/cgo/test/issue4054a.go @@ -311,8 +326,14 @@ go/misc/cgo/test/issue9400/asm_ppc64x.s go/misc/cgo/test/issue9400/gccgo.go go/misc/cgo/test/issue9400/stubs.go go/misc/cgo/test/issue9400_linux.go +go/misc/cgo/test/issue9510.go +go/misc/cgo/test/issue9510a/ +go/misc/cgo/test/issue9510a/a.go +go/misc/cgo/test/issue9510b/ +go/misc/cgo/test/issue9510b/b.go go/misc/cgo/test/issue9557.go go/misc/cgo/test/setgid_linux.go +go/misc/cgo/test/sigaltstack.go go/misc/cgo/test/sigprocmask_linux.c go/misc/cgo/test/sigprocmask_linux.go go/misc/cgo/test/sleep_windows_386.go @@ -320,9 +341,18 @@ go/misc/cgo/testasan/ go/misc/cgo/testasan/main.go go/misc/cgo/testcarchive/ go/misc/cgo/testcarchive/main.c +go/misc/cgo/testcarchive/main2.c +go/misc/cgo/testcarchive/main3.c +go/misc/cgo/testcarchive/main4.c go/misc/cgo/testcarchive/src/ go/misc/cgo/testcarchive/src/libgo/ go/misc/cgo/testcarchive/src/libgo/libgo.go +go/misc/cgo/testcarchive/src/libgo2/ +go/misc/cgo/testcarchive/src/libgo2/libgo2.go +go/misc/cgo/testcarchive/src/libgo3/ +go/misc/cgo/testcarchive/src/libgo3/libgo3.go +go/misc/cgo/testcarchive/src/libgo4/ +go/misc/cgo/testcarchive/src/libgo4/libgo4.go go/misc/cgo/testcarchive/src/p/ go/misc/cgo/testcarchive/src/p/p.go go/misc/cgo/testcarchive/test.bash @@ -331,11 +361,19 @@ go/misc/cgo/testcshared/main0.c go/misc/cgo/testcshared/main1.c go/misc/cgo/testcshared/main2.c go/misc/cgo/testcshared/main3.c +go/misc/cgo/testcshared/main4.c +go/misc/cgo/testcshared/main5.c go/misc/cgo/testcshared/src/ go/misc/cgo/testcshared/src/libgo/ go/misc/cgo/testcshared/src/libgo/libgo.go go/misc/cgo/testcshared/src/libgo2/ +go/misc/cgo/testcshared/src/libgo2/dup2.go +go/misc/cgo/testcshared/src/libgo2/dup3.go go/misc/cgo/testcshared/src/libgo2/libgo2.go +go/misc/cgo/testcshared/src/libgo4/
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: k...@cvs.openbsd.org2016/03/28 16:42:13 Modified files: lang/hugs : Makefile lang/hugs/pkg : PLIST Removed files: lang/hugs/pkg : PFRAG.shared Log message: Merge PFRAG.ahared into PLIST. Reminded by naddy@.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2016/03/28 15:37:25 Modified files: security/sslsplit: Makefile distinfo security/sslsplit/patches: patch-sslsplit_1 Added files: security/sslsplit/patches: patch-defaults_h Removed files: security/sslsplit/patches: patch-main_c patch-opts_c patch-ssl_c Log message: update to sslsplit-0.5.0
Re: gettext and libiconv don't need modules
2016-03-28 17:39 GMT+03:00 Christian Weisgerber: > TL;DR: We can and should kill the gettext and libiconv MODULES and > replace them with ordinary LIB_DEPENDS. > > > If you look at the libiconv module, it specifies both a lib and a > run dependency on libconv. That's because the iconv library also > needs some data file. On static archs, lib depends turned into > pure build depends, so we needed an extra run depends to make sure > that the data was installed. Without static archs, the lib depends > and wantlib is enough. > > The same applies to the gettext module, although that one has some > additional cruft which continues historical dependencies so we > didn't have to bump the ports that used the module when gettext was > updated. Again, a lib depends and wantlib would be enough now. > > The gettext module also includes a build dependency on gettext-tools > from last summer when we split the gettext runtime and tools into > separate ports. Only a small subset of gettext-using ports actually > need the tools; principally those that run msgfmt. Some people > (well, czarkoff@) have wanted to remove the general dependency on > the tools and only have an explicit build dependency in those ports > that actually need it. I've been hesitant about this, but now it > fits in with the idea of getting rid of the module. > > Killing the modules and replacing them with conventional dependencies > would put an end to the weird special case handling of the gettext > and libiconv libraries (MODULES, MOD*_LIB_DEPENDS, MOD*_WANTLIB), > making things simpler. Having to pay attention which ports actually > need to depend on gettext-tools is not simpler, but then this isn't > different from any other dependency. > > Opinions so far? So new rules for gettext dependency would be: 1. If PLIST contains ${PREFIX}/share/locale/*/*/*.mo files, the package should have, either directly or indirectly, a run-time dependency (LDEP/RDEP) on devel/gettext,-main. 2. If PLIST doesn't have such files, but package have direct RDEP on devel/gettext, this deserves a warning. Also, we drop all rules regarding libiconv: now we just add a dependency on either it or gettext, when needed, right? I want to formailize those rules as strict as possible, to adjust portcheck. > How would we go about killing the modules? As far as I can tell, > nothing prevents us from replacing MODULES with the normal > LIB_DEPENDS/WANTLIB entries, starting immediately, as part of normal > updates and maintenance work. For the libiconv module, this doesn't > even require a bump. For the gettext module it does if not part > of another update. > > Regarding the build dependency on gettext-tools, my plan would be > to > * identify ports that use the tools from bulk build logs, > * add explicit BUILD_DEPENDS += devel/gettext-tools, > * then do a test build with a default of MODGETTEXT_TOOLS ?= No, > * fix any fallout, > * repeat, > * and eventually kill MODGETTEXT_TOOLS. > > Comments? I think this is The Only Right Way. -- WBR, Vadim Zhukov
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: k...@cvs.openbsd.org2016/03/28 14:08:56 Modified files: sysutils/e2fsprogs: Makefile Added files: sysutils/e2fsprogs/patches: patch-lib_ext2fs_ext2_fs_h Log message: Protect us from #defining linux specific ioctls in ext2_fs.h which may be picked up by autoconfigure scripts and even used at runtime. Found by brynet@ when building archivers/libarchive. People agree (at least brynet@, zhuk@, nadd6@ iirc), the only bikeshed was wether to remote the #defines completely, used #ifdef __linux__ or #ifndef __OpenBSD__. I choose the latter for now, because I think i've seen some consumers of this port where some non-linux systems apparently *do* support the same ioctls.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: juan...@cvs.openbsd.org 2016/03/28 13:08:18 Modified files: lang/racket-minimal: Makefile lang/racket-minimal/pkg: README Log message: Add a warning about to run raco commands as root.
Re: [wip] Firefox 46.0beta5 & Thunderbird 45.0beta3
On Mon, Mar 28, 2016 at 07:49:39PM +0200, Landry Breuil wrote: > Hi, > > Firefox 46 will be live in some weeks, and hopefully this time with Gtk3 > by default. I also plan to default webrtc to build and ship, as it is > upstream since some releases - it needs dogfooding and testing, webcam > works, audio/sndio is flakey (some patches need backporting), data > channels demo works. > If you want to disable webrtc, just toggle media.peerconnection.enabled > to false in about:config. > > Upstream also plans to *maybe* enable e10s (separate rendering process) > in some conditions (some addons are still problematic, a11y things too), > i'm not sure how it will go, but to test it i think you need to toggle a > browser.tabs.remote.x pref in about:config too. Last i tried it in > nightly, it just worked. Apparently, to force-enable it at runtime (if e10s is enabled during the build, which should be the case by default in beta, iiuic) you need to set browser.tabs.remove.force-enable to true. Knobs, knobs, knobs everywhere! Landry
[wip] Firefox 46.0beta5 & Thunderbird 45.0beta3
Hi, Firefox 46 will be live in some weeks, and hopefully this time with Gtk3 by default. I also plan to default webrtc to build and ship, as it is upstream since some releases - it needs dogfooding and testing, webcam works, audio/sndio is flakey (some patches need backporting), data channels demo works. If you want to disable webrtc, just toggle media.peerconnection.enabled to false in about:config. Upstream also plans to *maybe* enable e10s (separate rendering process) in some conditions (some addons are still problematic, a11y things too), i'm not sure how it will go, but to test it i think you need to toggle a browser.tabs.remote.x pref in about:config too. Last i tried it in nightly, it just worked. While here, might aswell test Thunderbird 45.0b3, which should be very close to the final release which will replace Tb 3.8.x.x. https://cgit.rhaalovely.net/mozilla-firefox/?h=beta git clone -b beta https://rhaalovely.net/git/mozilla-firefox https://cgit.rhaalovely.net/mozilla-thunderbird/?h=beta git clone -b beta https://rhaalovely.net/git/mozilla-thunderbird If you want to build it you probably need the attached mozilla.port.mk diff. Provided amd64 pkgs should be 'compatible' with latest snaps, barring more libc/libpthread bumps.. doas env PKG_PATH=https://rhaalovely.net/stuff/amd64/ pkg_add -u firefox doas env PKG_PATH=https://rhaalovely.net/stuff/amd64/ pkg_add -u thunderbird my i386 builder has issues again. Landry ? 43gtk3.diff ? bundled.diff ? mozilla.port.mk.all Index: mozilla.port.mk === RCS file: /cvs/ports/www/mozilla/mozilla.port.mk,v retrieving revision 1.88 diff -u -r1.88 mozilla.port.mk --- mozilla.port.mk 20 Mar 2016 00:02:31 - 1.88 +++ mozilla.port.mk 28 Mar 2016 17:48:18 - @@ -21,6 +21,8 @@ MOZILLA_DIST ?=${MOZILLA_PROJECT} MOZILLA_DIST_VERSION ?=${MOZILLA_VERSION:C/rc.//} +# needed for PLIST +MOZILLA_VER = ${MOZILLA_VERSION:C/b[0-9]*//:C/esr//:C/rc.$//} HOMEPAGE ?=http://www.mozilla.org/projects/${MOZILLA_DIST} @@ -123,7 +125,8 @@ INSTALL_STRIP = .endif -.if ${FLAVOR:Mgtk3} +#.if ${FLAVOR:Mgtk3} +.if ${PKGPATH} == "www/mozilla-firefox" || (${MOZILLA_PROJECT} == "thunderbird" && ${MOZILLA_BRANCH} == "beta") # https://bugzilla.mozilla.org/show_bug.cgi?id=983843 CONFIGURE_ARGS += --with-system-cairo CONFIGURE_ARGS += --enable-default-toolkit=cairo-gtk3 @@ -142,16 +145,14 @@ CONFIGURE_ARGS +=--enable-application=${MOZILLA_CODENAME} # starting with esr45, only xulrunner will be special -.if ${MOZILLA_PROJECT} == "thunderbird" +.if ${MOZILLA_PROJECT} == "thunderbird" && ${MOZILLA_BRANCH} != "beta" WRKDIST ?= ${WRKDIR}/comm-${MOZILLA_BRANCH} .elif ${MOZILLA_PROJECT} == "xulrunner" || ${PKGPATH} == "www/firefox-esr" WRKDIST ?= ${WRKDIR}/mozilla-${MOZILLA_BRANCH} .else -WRKDIST ?= ${WRKDIR}/${MOZILLA_DIST}-${MOZILLA_DIST_VERSION} +WRKDIST ?= ${WRKDIR}/${MOZILLA_DIST}-${MOZILLA_VERSION} .endif -# needed for PLIST -MOZILLA_VER = ${MOZILLA_VERSION:C/b[0-9]*//:C/esr//:C/rc.$//} SUBST_VARS += MOZILLA_PROJECT MOZILLA_VER MOZILLA_VERSION MAKE_ENV +=MOZILLA_OFFICIAL=1 \
Re: gettext and libiconv don't need modules
Marc Espie: > > If you look at the libiconv module, it specifies both a lib and a > > run dependency on libconv. That's because the iconv library also > > needs some data file. On static archs, lib depends turned into > > pure build depends, so we needed an extra run depends to make sure > > that the data was installed. Without static archs, the lib depends > > and wantlib is enough. > > Did you run some kind of check to make sure it's indeed enough ? > I'm worried about stunts like gtk+3 magic dlopens I haven't and I don't see how I could. I think that concern is far-fetched. One issue I thought of: The handful of ports that are linked statically (CGI bins and the like) will need special attention. -- Christian "naddy" Weisgerber na...@mips.inka.de
Re: gettext and libiconv don't need modules
On Mon, Mar 28, 2016 at 04:39:14PM +0200, Christian Weisgerber wrote: > TL;DR: We can and should kill the gettext and libiconv MODULES and > replace them with ordinary LIB_DEPENDS. > > > If you look at the libiconv module, it specifies both a lib and a > run dependency on libconv. That's because the iconv library also > needs some data file. On static archs, lib depends turned into > pure build depends, so we needed an extra run depends to make sure > that the data was installed. Without static archs, the lib depends > and wantlib is enough. Did you run some kind of check to make sure it's indeed enough ? I'm worried about stunts like gtk+3 magic dlopens
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: z...@cvs.openbsd.org2016/03/28 09:38:21 Modified files: infrastructure/bin: portcheck Log message: Fix the case when WANTLIB-= didn't silence portcheck.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: z...@cvs.openbsd.org2016/03/28 09:37:12 Modified files: tests/portcheck: Makefile Log message: Hook up recently added test.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: z...@cvs.openbsd.org2016/03/28 09:36:18 Log message: Import (currently failing) test of false-positive in portcheck: when you have a PKG_ARCH=* port and, to avoid WANTLIBs (coming from MODULES, for example) getting in package, specify WANTLIB-= (empty), portcheck misses this and still warns about non-empty WANTLIB. The fix to be comitted separately in the next commit. Status: Vendor Tag: zhuk Release Tags: zhuk_20160328 N ports/tests/portcheck/t18/Makefile N ports/tests/portcheck/t18/pkg/PLIST No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: na...@cvs.openbsd.org 2016/03/28 08:50:12 Modified files: audio/mpg123 : Makefile distinfo Removed files: audio/mpg123/patches: patch-configure Log message: maintenance update to 1.23.3
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2016/03/28 08:49:52 Modified files: textproc/p5-XML-Smart: Makefile distinfo textproc/p5-XML-Smart/pkg: DESCR Log message: update to 1.79.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2016/03/28 08:48:02 Modified files: productivity/wyrd: Makefile distinfo Removed files: productivity/wyrd/patches: patch-Makefile_in patch-configure Log message: update to 1.4.6; drops patches no longer required.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2016/03/28 08:46:07 Modified files: devel/p5-Module-Which: Makefile distinfo Log message: update to 0.05. ok bluhm@
gettext and libiconv don't need modules
TL;DR: We can and should kill the gettext and libiconv MODULES and replace them with ordinary LIB_DEPENDS. If you look at the libiconv module, it specifies both a lib and a run dependency on libconv. That's because the iconv library also needs some data file. On static archs, lib depends turned into pure build depends, so we needed an extra run depends to make sure that the data was installed. Without static archs, the lib depends and wantlib is enough. The same applies to the gettext module, although that one has some additional cruft which continues historical dependencies so we didn't have to bump the ports that used the module when gettext was updated. Again, a lib depends and wantlib would be enough now. The gettext module also includes a build dependency on gettext-tools from last summer when we split the gettext runtime and tools into separate ports. Only a small subset of gettext-using ports actually need the tools; principally those that run msgfmt. Some people (well, czarkoff@) have wanted to remove the general dependency on the tools and only have an explicit build dependency in those ports that actually need it. I've been hesitant about this, but now it fits in with the idea of getting rid of the module. Killing the modules and replacing them with conventional dependencies would put an end to the weird special case handling of the gettext and libiconv libraries (MODULES, MOD*_LIB_DEPENDS, MOD*_WANTLIB), making things simpler. Having to pay attention which ports actually need to depend on gettext-tools is not simpler, but then this isn't different from any other dependency. Opinions so far? How would we go about killing the modules? As far as I can tell, nothing prevents us from replacing MODULES with the normal LIB_DEPENDS/WANTLIB entries, starting immediately, as part of normal updates and maintenance work. For the libiconv module, this doesn't even require a bump. For the gettext module it does if not part of another update. Regarding the build dependency on gettext-tools, my plan would be to * identify ports that use the tools from bulk build logs, * add explicit BUILD_DEPENDS += devel/gettext-tools, * then do a test build with a default of MODGETTEXT_TOOLS ?= No, * fix any fallout, * repeat, * and eventually kill MODGETTEXT_TOOLS. Comments? -- Christian "naddy" Weisgerber na...@mips.inka.de
ECC support for sendmail
Hi, Please consider adding the following to the OpenBSD sendmail port in order to add ECC support to STARTTLS (-D_FFR_TLS_EC), and hopefully to add a little more granular control of TLS (-D_FFR_TLS_SE_OPTS) as well (but at least the former seems a quite reasonable default in CE 2016). --- sendmail/files/site.OS.m4.dist Mon Mar 28 06:39:40 2016 +++ sendmail/files/site.OS.m4 Mon Mar 28 06:50:33 2016 @@ -32,6 +32,8 @@ APPENDDEF(`conf_sendmail_ENVDEF', `-DMILTER') dnl (START)TLS APPENDDEF(`confENVDEF', `-DSTARTTLS')dnl +APPENDDEF(`confENVDEF', `-D_FFR_TLS_EC')dnl +APPENDDEF(`confENVDEF', `-D_FFR_TLS_SE_OPTS')dnl APPENDDEF(`confLIBS', `-lssl -lcrypto')dnl dnl Flavors dnl === Thanks and Best Regards, --Kyle P.S. Also, please note that I'm not on the ports mailing list. -- CA +1-778-819-UNIX BackWatcher, Inc. US +1-425-584-UNIX Information Security Solutions SIP am...@backwatcher.comwww.backwatcher.ca INUM +883-5100-0990-1657 / ISN UNIX*1917 / C*NET 1-731-UNIX GPG ed25519/F57091DBD60FBBB8 [ed25519/D60FBBB8] 985C 5B61 4ACE C89A 0DEE ECCD F570 91DB D60F BBB8 OTR E1A46361 9FD0D801 0132D21A FE2E96BE 39E3F069 : am...@backwatcher.com 5AB3E0B8 31F6ADB4 9A7D2FC2 A8235281 5776701E : silcnet pgpW2BykBgzC0.pgp Description: OpenPGP digital signature
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: sema...@cvs.openbsd.org 2016/03/28 08:25:02 Modified files: www/privoxy: Makefile www/privoxy/pkg: PLIST Log message: www/privoxy: annotate all templates with @sample OK sthen@
Re: [NEW] security/pdfid
On Mon, Mar 28, 2016 at 12:44:33PM +0200, Remi Pointel wrote: > attached is the new port pdfid. I would recommend to install the binary without the .py suffix, the user does not care that it is written in python. You should patch the #! line. $ /usr/local/bin/pdfid.py env: python: No such file or directory There are 3 plugins which were not installed. The script searches for them in the bin directory. This diff fixes the issues, but perhaps the plugins should be installed somewhere in /usr/local/lib and the script's search routine patched. bluhm --- Makefile.orig Mon Mar 28 15:33:36 2016 +++ MakefileMon Mar 28 15:52:13 2016 @@ -21,12 +21,17 @@ EXTRACT_SUFX = .zip MODULES = lang/python -NO_BUILD = Yes NO_TEST = Yes WRKDIST = ${WRKDIR} +do-build: + echo 'H\n1s,^#!.*,#!${MODPY_BIN},\nwq' | ed -s ${WRKSRC}/pdfid.py + do-install: - ${INSTALL_SCRIPT} ${WRKSRC}/pdfid.py ${PREFIX}/bin/ + ${INSTALL_SCRIPT} ${WRKSRC}/pdfid.py ${PREFIX}/bin/pdfid + for plugin in `cat ${WRKSRC}/plugin_list`; do \ + ${INSTALL_DATA} ${WRKSRC}/$$plugin ${PREFIX}/bin/; \ + done .include
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gonz...@cvs.openbsd.org 2016/03/28 07:25:25 Modified files: education/anki : Makefile distinfo education/anki/patches: patch-anki_anki education/anki/pkg: PLIST Log message: Update for Anki to 2.0.33: http://ankisrs.net/docs/changes.html Maintener timout.
Re: [UPDATE] math/py-scipy
On Sat, Mar 26, 2016 at 04:45:19PM -0400, Daniel Dickman wrote: > Here's an update of scipy that also adds a python3 flavour. Both the > update and the python3 flavour were requested by Anton Kasimov. A few > things to note. > > First, weave is only packaged under python2. I added MODPY_COMMENT3 in the > Makefile which is like MODPY_COMMENT except in reverse. (i.e. it's a > comment under python3 instead of python2). If anyone wants me to handle > differently let me know. > > Next, one of the tests core dumps: test_economic_1_col_bad_update. Since I > haven't seen any bugs filed on github for this issue, I'm wondering if > there's an error with OpenBSD complex math. A reproducible test case is > included in the patch for anyone that wants to dig further, but I don't > consider this to be something to block the update. > > Finally, while some regress tests fail, it's no worse than the failures > with the current version in ports. > > Diff below and online: > http://dickman.org/openbsd/ports/python/scipy.diff > > ok? > ok shadchin@ -- Alexandr Shadchin
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: rpoin...@cvs.openbsd.org2016/03/28 06:40:06 Log message: import pdfid ok sthen@. PDFiD is not a PDF parser, but it will scan a file to look for certain PDF keywords, allowing you to identify PDF documents that contain (for example) JavaScript or execute an action when opened. PDFiD will also handle name obfuscation. Status: Vendor Tag: rpointel Release Tags: rpointel_20160328 N ports/security/pdfid/Makefile N ports/security/pdfid/distinfo N ports/security/pdfid/pkg/PLIST N ports/security/pdfid/pkg/DESCR No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: z...@cvs.openbsd.org2016/03/28 06:40:06 Modified files: infrastructure/bin: portcheck Log message: Compact checks lines for dbus,-suid.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: rpoin...@cvs.openbsd.org2016/03/28 06:40:49 Modified files: security : Makefile Log message: SUBDIR += pdfid
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: d...@cvs.openbsd.org2016/03/28 06:20:33 Modified files: www/py-werkzeug: Makefile distinfo Log message: update to py-werkzeug-0.11.5 ok jca@ shadchin@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2016/03/28 05:17:06 Added files: databases/p5-sybperl/patches: patch-DBlib_DBlib_xs Log message: fix following freetds update, which moved DBMAXNAME defn to a non-installed header. spotted by nigel@, https://github.com/FreeTDS/freetds/issues/57
Re: NEW: devel/py-wrapt 1.10.6
On Mon, Mar 28, 2016 at 12:18:16PM +0200, Daniel Jakots wrote: > On Sun, 27 Mar 2016 23:06:00 +0200, Daniel Jakots> wrote: > > > - enable test > > I did the same omission than in the update of py-rsa, it needs: > > TEST_DEPENDS = devel/py-wrapt${MODPY_FLAVOR} \ > devel/py-test${MODPY_FLAVOR} Instead devel/py-wrapt${MODPY_FLAVOR} better to use ${FULLPKGNAME}:${FULLPKGPATH} Example: In system installed py-wrapt-1.0.0 Check depends will ok, but need install new py-wrapt-1.10.6 Thanks for review and rework. -- Alexandr Shadchin
[NEW] security/pdfid
Hi, attached is the new port pdfid. == $ pkg_info pdfid Information for inst:pdfid-0.2.1 Comment: tool to test a PDF file Description: PDFiD is not a PDF parser, but it will scan a file to look for certain PDF keywords, allowing you to identify PDF documents that contain (for example) JavaScript or execute an action when opened. PDFiD will also handle name obfuscation. Maintainer: Remi PointelWWW: http://blog.didierstevens.com/programs/pdf-tools/ == Ok? Cheers, Remi. pdfid-0.2.1.tar.gz Description: application/gzip
www/privoxy: annotate all templates with @sample
Hi, I noted that several templates of privoxy are installed in ${LOCALBASE}/share/examples/privoxy/templates/, but not copied as samples in /etc/privoxy/templates/ It results 500 Internal Privoxy Error when privoxy needs the missing template. There are 3 files in examples but with @sample entries: privoxy/templates/connection-timeout privoxy/templates/no-server-data privoxy/templates/url-info-osd.xml The following diff annotate them as @sample, in order to let pkg_add installs them in SYSCONFDIR. Comments or OK ? -- Sebastien Marie Index: Makefile === RCS file: /cvs/ports/www/privoxy/Makefile,v retrieving revision 1.32 diff -u -p -r1.32 Makefile --- Makefile26 Jan 2016 21:52:17 - 1.32 +++ Makefile28 Mar 2016 10:18:16 - @@ -6,6 +6,7 @@ V= 3.0.24 DISTNAME= privoxy-${V}-stable PKGNAME= privoxy-${V} CATEGORIES=www +REVISION= 0 HOMEPAGE= http://www.privoxy.org/ Index: pkg/PLIST === RCS file: /cvs/ports/www/privoxy/pkg/PLIST,v retrieving revision 1.10 diff -u -p -r1.10 PLIST --- pkg/PLIST 2 May 2011 17:43:43 - 1.10 +++ pkg/PLIST 28 Mar 2016 10:18:16 - @@ -111,6 +111,9 @@ share/examples/privoxy/templates/connect @owner @group share/examples/privoxy/templates/connection-timeout +@owner _privoxy +@group _privoxy +@sample ${SYSCONFDIR}/privoxy/templates/connection-timeout share/examples/privoxy/templates/default @owner _privoxy @group _privoxy @@ -202,6 +205,9 @@ share/examples/privoxy/templates/mod-uns @owner @group share/examples/privoxy/templates/no-server-data +@owner _privoxy +@group _privoxy +@sample ${SYSCONFDIR}/privoxy/templates/no-server-data share/examples/privoxy/templates/no-such-domain @owner _privoxy @group _privoxy @@ -257,6 +263,9 @@ share/examples/privoxy/templates/untrust @owner @group share/examples/privoxy/templates/url-info-osd.xml +@owner _privoxy +@group _privoxy +@sample ${SYSCONFDIR}/privoxy/templates/url-info-osd.xml share/examples/privoxy/user.action @owner _privoxy @group _privoxy
Re: NEW: devel/py-wrapt 1.10.6
On Sun, 27 Mar 2016 23:06:00 +0200, Daniel Jakotswrote: > - enable test I did the same omission than in the update of py-rsa, it needs: TEST_DEPENDS = devel/py-wrapt${MODPY_FLAVOR} \ devel/py-test${MODPY_FLAVOR}
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: z...@cvs.openbsd.org2016/03/28 03:29:00 Modified files: games/gottcode : Makefile.inc Log message: Unbreak those as well: it's bad idea to commit subdirectory when the fix was done in Makefile.inc. Spotted by aja@ as well
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: z...@cvs.openbsd.org2016/03/28 03:27:10 Modified files: graphics/cstitch: Makefile Log message: Unbreak after recent qommit. Spotted by aja@
Re: [update] textproc/multimarkdown
On Tue, 22 Mar 2016 12:58:32 -0600 attilawrote: > Stuart Henderson writes: > > > On 2016/03/19 15:38, Michael McConville wrote: > >> > +ALL_TARGET =deprecated > >> > + > >> > +# golf MAKE_FLAGS down to 80chars.. :-| > >> > +_i =-include > >> > +_incs = ${_i} src/GLibFacade.h ${_i} src/version.h ${_i} > >> > src/parser.h > >> > +MAKE_FLAGS =CFLAGS="${CFLAGS} ${_incs}" > >> > >> Why golf? Can't you just use backslashes as necessary? I would get rid > >> of _i and _incs. > >> > > > > I think this approach is dangerous, better to keep those bits which are > > normally part of upstream's Makefile in their Makefile and change the > > way that you pass in CFLAGS. (e.g. maybe pass in COPTFLAGS instead and > > change their Makefile to do COPTFLAGS?=-O3 and "CFLAGS?=${COPTFLAGS} ... > > -include ...") > > > > But then again, we have cmake, why not just use that? It's upstream's > > preferred build infrastructure, whereas for the make-based one they say > > "I don't recommend this approach, but it should work in a pinch".. > > Agree, this was sleazy and poorly considered. I'm sorry. Anytime I > find myself golfing I should stop. > > Attached is new attempt that uses cmake instead. Works for me on i386. > > N.B.: I set NO_TEST to Yes now because the tests are buried in a git > submodule; furthermore, the repo on which the submodule is based has > no tags and thus no releases, making it impossible for me to reach out > for it as a distfile (unless I'm missing something). I'm going to try > to convince the upstream to start tagging that repo, but if I fail > then I'll self-host a tarball for tests on the next update. > > Thanks as always for the feedback. > > Pax, -A > -- > http://haqistan.net/~attila | att...@stalphonsos.com | 0x62A729CF Atilla's patch works well for me on amd64. I have used it heavily for several days. However, upstream has a new release, so here's a new patch 99% based on atilla's. Index: Makefile === RCS file: /cvs/ports/textproc/multimarkdown/Makefile,v retrieving revision 1.2 diff -u -p -r1.2 Makefile --- Makefile5 Apr 2015 13:31:15 - 1.2 +++ Makefile22 Mar 2016 18:48:31 - @@ -1,13 +1,13 @@ # $OpenBSD: Makefile,v 1.2 2015/04/05 13:31:15 sthen Exp $ -V =4.7.1 +V =5.2.0 COMMENT = marked-up plain text to formatted document converter DISTNAME = ${GH_PROJECT}-${V} PKGNAME = multimarkdown-${V} CATEGORIES = textproc GH_ACCOUNT = fletcher -GH_PROJECT = MultiMarkdown-4 +GH_PROJECT = MultiMarkdown-5 GH_TAGNAME = ${V} HOMEPAGE = http://fletcherpenney.net/multimarkdown/ @@ -20,15 +20,19 @@ WANTLIB += c BUILD_DEPENDS =devel/greg -USE_GMAKE =Yes -ALL_TARGET = ALL -MAKE_FLAGS = CFLAGS="${CFLAGS} -include GLibFacade.h -DHAVE_ARC4RANDOM" \ - GREG=${LOCALBASE}/bin/greg +MODULES = devel/cmake + +CFLAGS += -DHAVE_SRAND_DETERMINISTIC + +CONFIGURE_ARGS += -DGREG=${LOCALBASE}/bin/greg # Test files aren't included in distfile. NO_TEST = Yes +pre-configure: + touch ${WRKBUILD}/README.html + do-install: - ${INSTALL_PROGRAM} ${WRKSRC}/multimarkdown ${PREFIX}/bin/ + ${INSTALL_PROGRAM} ${WRKBUILD}/multimarkdown ${PREFIX}/bin/ .include Index: distinfo === RCS file: /cvs/ports/textproc/multimarkdown/distinfo,v retrieving revision 1.1.1.1 diff -u -p -r1.1.1.1 distinfo --- distinfo24 Mar 2015 19:16:13 - 1.1.1.1 +++ distinfo22 Mar 2016 18:48:31 - @@ -1,2 +1,2 @@ -SHA256 (MultiMarkdown-4-4.7.1.tar.gz) = gy5dzm+hv/TWfmSsLMJPCSCujZEoDlPqh5I3odQ/SGU= -SIZE (MultiMarkdown-4-4.7.1.tar.gz) = 120896 +SHA256 (MultiMarkdown-5-5.2.0.tar.gz) = 7tjLjDmJa1e41uAJOn2WhDMmp/cxA2xLp5CsBPhOEsM= +SIZE (MultiMarkdown-5-5.2.0.tar.gz) = 135547 Index: patches/patch-CMakeLists_txt === RCS file: patches/patch-CMakeLists_txt diff -N patches/patch-CMakeLists_txt --- /dev/null 1 Jan 1970 00:00:00 - +++ patches/patch-CMakeLists_txt22 Mar 2016 18:48:31 - @@ -0,0 +1,14 @@ +$OpenBSD$ + +Use installed greg +--- CMakeLists.txt.origMon Feb 22 20:05:02 2016 CMakeLists.txt Tue Mar 22 12:45:41 2016 +@@ -149,7 +149,7 @@ endif () + # Need to build parser.c via greg + add_custom_command ( + OUTPUT ${CMAKE_CURRENT_BINARY_DIR}/parser.c +- COMMAND ${PROJECT_SOURCE_DIR}/submodules/greg/greg -o ${CMAKE_CURRENT_BINARY_DIR}/parser.c ${PROJECT_SOURCE_DIR}/src/parser.leg ++ COMMAND ${GREG} -o ${CMAKE_CURRENT_BINARY_DIR}/parser.c ${PROJECT_SOURCE_DIR}/src/parser.leg + ) + + # src_files are the primary files, and will be included in doxygen documentation Index: patches/patch-Makefile === RCS
Re: update security/py-rsa to 3.4.1
On Sun, Mar 27, 2016 at 08:48:25PM +0200, Daniel Jakots wrote: > Hi, > > Best news about it is that it gets rid of devel/py-unittest2. > > I did make test on {amd64,i386} with {python2.7,python3.4}: all's good. Did you try building sysutils/awscli with it? I doubt it does. > Index: Makefile > === > RCS file: /cvs/ports/security/py-rsa/Makefile,v > retrieving revision 1.3 > diff -u -p -r1.3 Makefile > --- Makefile 8 Jan 2016 09:23:00 - 1.3 > +++ Makefile 27 Mar 2016 18:35:22 - > @@ -2,10 +2,9 @@ > > COMMENT= Python RSA implementation > > -MODPY_EGG_VERSION= 3.2.3 > +MODPY_EGG_VERSION= 3.4.1 > DISTNAME=rsa-${MODPY_EGG_VERSION} > PKGNAME= py-${DISTNAME} > -REVISION=0 > > CATEGORIES= security > > @@ -23,13 +22,7 @@ RUN_DEPENDS= devel/py-asn1${MODPY_FLAVO > FLAVORS= python3 > FLAVOR ?= > > -.if ${FLAVOR:Mpython3} > -# needs devel/py-unittest2,python3 > -#NO_TEST=Yes > -.else > -TEST_DEPENDS=${RUN_DEPENDS} \ > - devel/py-unittest2 > -.endif > +TEST_DEPENDS=${RUN_DEPENDS} > > .if ${FLAVOR:Mpython3} > post-install: > @@ -39,6 +32,6 @@ post-install: > .endif > > do-test: > - cd ${WRKSRC} && ${MODPY_BIN} ./run_tests.py > + cd ${WRKSRC} && ${MODPY_BIN} -m pytest > > .include > Index: distinfo > === > RCS file: /cvs/ports/security/py-rsa/distinfo,v > retrieving revision 1.1.1.1 > diff -u -p -r1.1.1.1 distinfo > --- distinfo 6 Jan 2016 15:45:14 - 1.1.1.1 > +++ distinfo 27 Mar 2016 18:35:22 - > @@ -1,2 +1,2 @@ > -SHA256 (rsa-3.2.3.tar.gz) = FNsojMQNYzne32DXpHBTqwBLSol2pcWUAqIR2Pxb8h8= > -SIZE (rsa-3.2.3.tar.gz) = 35628 > +SHA256 (rsa-3.4.1.tar.gz) = b7dNfX1uy63TfbrTKbMdhI+6X+i0CtbRLPfi5aoV5GQ= > +SIZE (rsa-3.4.1.tar.gz) = 40938 > Index: patches/patch-rsa_pkcs1_py > === > RCS file: patches/patch-rsa_pkcs1_py > diff -N patches/patch-rsa_pkcs1_py > --- patches/patch-rsa_pkcs1_py8 Jan 2016 09:23:00 - 1.1 > +++ /dev/null 1 Jan 1970 00:00:00 - > @@ -1,100 +0,0 @@ > -$OpenBSD: patch-rsa_pkcs1_py,v 1.1 2016/01/08 09:23:00 ajacoutot Exp $ > - > -https://bitbucket.org/sybren/python-rsa/commits/0cbcc529926afd61c6df4f50cfc29971beafd2c2/raw/ > - > rsa/pkcs1.py.origThu Nov 5 21:23:16 2015 > -+++ rsa/pkcs1.py Fri Jan 8 10:20:09 2016 > -@@ -22,10 +22,10 @@ very clear example, read http://www.di-mgt.com.au/rsa_ > - At least 8 bytes of random padding is used when encrypting a message. This > makes > - these methods much more secure than the ones in the ``rsa`` module. > - > --WARNING: this module leaks information when decryption or verification > fails. > --The exceptions that are raised contain the Python traceback information, > which > --can be used to deduce where in the process the failure occurred. DO NOT PASS > --SUCH INFORMATION to your users. > -+WARNING: this module leaks information when decryption fails. The exceptions > -+that are raised contain the Python traceback information, which can be used > to > -+deduce where in the process the failure occurred. DO NOT PASS SUCH > INFORMATION > -+to your users. > - ''' > - > - import hashlib > -@@ -288,37 +288,23 @@ def verify(message, signature, pub_key): > - :param pub_key: the :py:class:`rsa.PublicKey` of the person signing the > message. > - :raise VerificationError: when the signature doesn't match the message. > - > --.. warning:: > -- > --Never display the stack trace of a > --:py:class:`rsa.pkcs1.VerificationError` exception. It shows where in > --the code the exception occurred, and thus leaks information about > the > --key. It's only a tiny bit of information, but every bit makes > cracking > --the keys easier. > -- > - ''' > - > --blocksize = common.byte_size(pub_key.n) > -+keylength = common.byte_size(pub_key.n) > - encrypted = transform.bytes2int(signature) > - decrypted = core.decrypt_int(encrypted, pub_key.e, pub_key.n) > --clearsig = transform.int2bytes(decrypted, blocksize) > -- > --# If we can't find the signature marker, verification failed. > --if clearsig[0:2] != b('\x00\x01'): > --raise VerificationError('Verification failed') > -+clearsig = transform.int2bytes(decrypted, keylength) > - > --# Find the 00 separator between the padding and the payload > --try: > --sep_idx = clearsig.index(b('\x00'), 2) > --except ValueError: > --raise VerificationError('Verification failed') > -- > --# Get the hash and the hash method > --(method_name, signature_hash) = _find_method_hash(clearsig[sep_idx+1:]) > -+# Get the hash method > -+method_name =
Re: update security/py-rsa to 3.4.1
Daniel Jakotswrites: > Hi, > > Best news about it is that it gets rid of devel/py-unittest2. > > I did make test on {amd64,i386} with {python2.7,python3.4}: all's good. As discussed, better get tests for sysutils/awscli. Note that you're missing a test-dep on devel/py-test${MODPY_FLAVOR}. TEST_DEPENDS can be moved up, below RUN_DEPENDS. With those items addressed, ok jca@ -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
Re: update www/py-werkzeug to 0.11.5
Daniel Jakotswrites: > Hi, > > Here's an update to the latest py-werkzeug. make test is the same on > amd64/i386. I also tested py-flask on both arch and py-acme on amd64 > only. ok jca@ -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: j...@cvs.openbsd.org2016/03/28 00:51:19 Modified files: multimedia/gstreamer-0.10/plugins-base: Makefile Log message: this builds on arm again after the recent ld.so/crt changes