textproc/p5-PDF-Table : Update to 0.11.0
Hi, ports@: Here is a patch to update textproc/p5-PDF-Table to 0.11.0. It build well and passed all tests on my amd64 head system. No other ports depends on it. Comments? OK ? Regards, wen Index: Makefile === RCS file: /cvs/ports/textproc/p5-PDF-Table/Makefile,v retrieving revision 1.13 diff -u -p -r1.13 Makefile --- Makefile21 Feb 2019 18:56:05 - 1.13 +++ Makefile1 Jul 2019 06:14:01 - @@ -2,12 +2,12 @@ COMMENT = create PDF tables with perl -DISTNAME = PDF-Table-0.10.1 +DISTNAME = PDF-Table-0.11.0 CATEGORIES = textproc # Perl -PERMIT_PACKAGE_CDROM = Yes +PERMIT_PACKAGE = Yes MODULES = cpan PKG_ARCH = * Index: distinfo === RCS file: /cvs/ports/textproc/p5-PDF-Table/distinfo,v retrieving revision 1.7 diff -u -p -r1.7 distinfo --- distinfo21 Feb 2019 18:56:05 - 1.7 +++ distinfo1 Jul 2019 06:14:01 - @@ -1,2 +1,2 @@ -SHA256 (PDF-Table-0.10.1.tar.gz) = +y02nQl5funPeYEf+FBhu6jlbGHmXBOspnS1MhBaJxM= -SIZE (PDF-Table-0.10.1.tar.gz) = 28459 +SHA256 (PDF-Table-0.11.0.tar.gz) = kUtcRnHaQArOFa7DIkSgvRWW4is05LNBp/KLJey0vZg= +SIZE (PDF-Table-0.11.0.tar.gz) = 29930
textproc/p5-PDF-API2: Update to 2.034
Hi, Here is a patch to update textproc/p5-PDF-API2 to 2.034. It build well and passed all test on my amd64 system. Two ports depends on it, textproc/p5-PDF-API2-Simple and textproc/p5-PDF-Table. Both build well and passed all tests. Comments? OK? Regards, wen Index: Makefile === RCS file: /cvs/ports/textproc/p5-PDF-API2/Makefile,v retrieving revision 1.23 diff -u -p -r1.23 Makefile --- Makefile3 Jun 2019 16:06:57 - 1.23 +++ Makefile1 Jul 2019 03:25:33 - @@ -5,7 +5,7 @@ COMMENT = create PDF documents with perl MODULES = cpan PKG_ARCH = * -DISTNAME = PDF-API2-2.033 +DISTNAME = PDF-API2-2.034 CATEGORIES = textproc MAINTAINER = Stuart Henderson Index: distinfo === RCS file: /cvs/ports/textproc/p5-PDF-API2/distinfo,v retrieving revision 1.12 diff -u -p -r1.12 distinfo --- distinfo23 Aug 2017 13:38:25 - 1.12 +++ distinfo1 Jul 2019 03:25:33 - @@ -1,2 +1,2 @@ -SHA256 (PDF-API2-2.033.tar.gz) = nAhm7BowU/c6+spfXNvmklkDtM5gb0v0rDF3MaadJ6A= -SIZE (PDF-API2-2.033.tar.gz) = 3533753 +SHA256 (PDF-API2-2.034.tar.gz) = iqmIGPtuS+vW+QluIimJ3N1f1MX6KtHH8BSQU/xo8cw= +SIZE (PDF-API2-2.034.tar.gz) = 3510253
Update: deve/py-freezegun 0.3.9 -> 0.3.12
This updates py-freezegun to 0.3.12. Our current version isn't compatible with Python 3.7.x so anything python3 that uses py-freezegun for tests is currently broken. I tested all consumers of py-freezegun. The only change that isn't 100% positive is the python 2 flavor of py-cached-property fails one test now: == FAIL: test_threads_ttl_expiry (tests.test_cached_property.TestCachedPropertyWithTTL) -- Traceback (most recent call last): File "/home/ports/pobj/py-cached-property-1.5.1/cached-property-1.5.1/tests/test_cached_property.py", line 207, in test_threads_ttl_expiry self.assert_cached(check, 2 * num_threads) File "/home/ports/pobj/py-cached-property-1.5.1/cached-property-1.5.1/tests/test_cached_property.py", line 69, in assert_cached self.assertEqual(check.add_cached, expected) AssertionError: 6 != 10 All others are the smae or better. Especially the python3 flavors, since none of them worked before this update. py-test-benchmark tests don't run before or after this update. I'm looking into why that is the case. cc maintainer --Kurt Index: Makefile === RCS file: /cvs/ports/devel/py-freezegun/Makefile,v retrieving revision 1.4 diff -u -p -r1.4 Makefile --- Makefile15 May 2019 12:04:36 - 1.4 +++ Makefile28 Jun 2019 06:25:59 - @@ -2,22 +2,22 @@ COMMENT = let your Python tests travel through time -MODPY_EGG_VERSION =0.3.9 +MODPY_EGG_VERSION =0.3.12 DISTNAME = freezegun-${MODPY_EGG_VERSION} PKGNAME = py-freezegun-${MODPY_EGG_VERSION} -REVISION = 0 CATEGORIES = devel MAINTAINER = Joerg Jung # Apache 2.0 -PERMIT_PACKAGE_CDROM = Yes +PERMIT_PACKAGE = Yes MODULES = lang/python MODPY_PI = Yes MODPY_SETUPTOOLS = Yes +MODPY_PYTEST = Yes RUN_DEPENDS = devel/py-dateutil${MODPY_FLAVOR} \ devel/py-six${MODPY_FLAVOR} @@ -28,8 +28,9 @@ TEST_DEPENDS =devel/py-nose${MODPY_FLA FLAVORS = python3 FLAVOR ?= -do-test: - cd ${WRKSRC} && \ - ${LOCALBASE}/bin/nosetests${MODPY_BIN_SUFFIX} tests/*.py +.if !${FLAVOR:Mpython3} +post-extract: + rm ${WRKSRC}/freezegun/_async.py +.endif .include Index: distinfo === RCS file: /cvs/ports/devel/py-freezegun/distinfo,v retrieving revision 1.2 diff -u -p -r1.2 distinfo --- distinfo27 May 2017 19:26:52 - 1.2 +++ distinfo28 Jun 2019 06:25:59 - @@ -1,2 +1,2 @@ -SHA256 (freezegun-0.3.9.tar.gz) = eDzMzX9glov+Sa2eEUwY6itjgx+qr2HB8fcd394cDu4= -SIZE (freezegun-0.3.9.tar.gz) = 18118 +SHA256 (freezegun-0.3.12.tar.gz) = Kk2cjNPASiAeIMMTyvi2M48c+kzaQ/RqlMxKn9E+pec= +SIZE (freezegun-0.3.12.tar.gz) = 24346 Index: pkg/PLIST === RCS file: /cvs/ports/devel/py-freezegun/pkg/PLIST,v retrieving revision 1.1.1.1 diff -u -p -r1.1.1.1 PLIST --- pkg/PLIST 26 May 2017 19:26:15 - 1.1.1.1 +++ pkg/PLIST 28 Jun 2019 06:25:59 - @@ -9,5 +9,7 @@ lib/python${MODPY_VERSION}/site-packages lib/python${MODPY_VERSION}/site-packages/freezegun/__init__.py ${MODPY_COMMENT}lib/python${MODPY_VERSION}/site-packages/freezegun/${MODPY_PYCACHE}/ lib/python${MODPY_VERSION}/site-packages/freezegun/${MODPY_PYCACHE}__init__.${MODPY_PYC_MAGIC_TAG}pyc +${MODPY_COMMENT}lib/python${MODPY_VERSION}/site-packages/freezegun/${MODPY_PYCACHE}_async.${MODPY_PYC_MAGIC_TAG}pyc lib/python${MODPY_VERSION}/site-packages/freezegun/${MODPY_PYCACHE}api.${MODPY_PYC_MAGIC_TAG}pyc +${MODPY_COMMENT}lib/python${MODPY_VERSION}/site-packages/freezegun/_async.py lib/python${MODPY_VERSION}/site-packages/freezegun/api.py
Re: NEW: Tacacs+ port - shrubbery.net version
Hi sthen, > Slightly tweaked version attached, this one's ok with me: > > - https homepage > - PERMIT_*_CDROM is not used for new ports > - whitespace nit in Makefile > - tweak comment in patch > - place @extraunexec above the @sample line, that way pkg_delete -c doesn't > complain about a missing dir. (pkg_delete without -c will complain about > not being able to remove the dir, that is no problem). > - regen plist to include pkg-readme > - adjust pkg-readme to set uid/gid on the files > - change group ownership of log dir to wheel, easier for admins thanks for the review, gonna commit this one.
sparc64 bulk build report
bulk build on sparc64-1.ports.openbsd.org started on Sun Jun 9 12:17:30 MDT 2019 finished at Sun Jun 30 14:06:26 MDT 2019 lasted 21D18h48m done with kern.version=OpenBSD 6.5-current (GENERIC) #196: Sun Jun 9 06:01:53 MDT 2019 built packages:9664 Jun 9:303 Jun 10:1752 Jun 11:144 Jun 12:90 Jun 13:105 Jun 14:115 Jun 15:35 Jun 16:52 Jun 17:86 Jun 18:54 Jun 19:40 Jun 20:94 Jun 21:108 Jun 22:87 Jun 23:93 Jun 24:95 Jun 25:172 Jun 26:286 Jun 27:341 Jun 28:503 Jun 29:1588 Jun 30:3520 critical path missing pkgs: http://build-failures.rhaalovely.net//sparc64/2019-06-09/summary.log build failures: 79 http://build-failures.rhaalovely.net//sparc64/2019-06-09/astro/celestia.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/audio/clementine.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/audio/gradio.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/cad/gnucap.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/cad/magic.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/cad/netgen.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/cad/qucs.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/chinese/libpinyin.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/devel/acpica.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/devel/kdevelop.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/devel/openmpi.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/devel/proj.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/devel/py-unicorn.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/devel/pycdc.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/emulators/citra.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/emulators/desmume.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/emulators/fs-uae.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/emulators/gambatte,-main.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/emulators/nestopia,-libretro.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/emulators/ppsspp.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/emulators/vbam.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/emulators/xnp2.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/games/dxx-rebirth.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/games/godot.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/games/love.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/games/maelstrom.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/games/mvdsv.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/games/pokerth.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/games/xevil.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/games/xmoto.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/graphics/dibuja.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/graphics/makehuman.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/graphics/openimageio.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/inputmethods/scim-fcitx.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/inputmethods/scim-hangul.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/inputmethods/scim-pinyin.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/inputmethods/scim-tables.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/lang/apl.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/lang/erlang/16.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/lang/erlang/17,-main.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/lang/erlang/18,-main.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/lang/erlang/19.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/lang/erlang/21.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/lang/janet.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/mail/kopano/core.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/math/gbc.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/math/kst.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/math/plplot,-c++.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/multimedia/mkvtoolnix.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/multimedia/qtav.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/multimedia/synfig.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/net/dleyna/renderer.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/net/dleyna/server.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/net/mac-telnet.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/net/mutella.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/net/pmacct.log http://build-failures.rhaalovely.net//sparc64/2019-06-09/net/telegram-purple.log http:/
NEW: multimedia/gerbera [WIP]
Here's a working port for gerbera, but I never really used it since my use case for it disappeared soon after getting the port done. The server starts, full offline documentation is available, one (small?) nit with inotify support remains unsolved, otherwise the port is in good shape. You'll find comments in the Makefile, active upstream work towards C++17 and an almost ready rc.d script where proper daemon user integration shouuld be worked out. So in case anyone is interested in this, feel free to pick it up, improve things and perhaps import it. I'll gladly review things and would be even happier so see someone maintain it, but I personally put any more effort into it for above mentioned reasons. Information for inst:gerbera-1.3.1 Comment: UPnP Media Server Description: gerbera is a UPnP media server which allows you to stream your digital media through your home network and consume it on a variety of UPnP compatible devices. Features include: * Browse and playback your media via your network on all kinds of devices. * Metadata extraction from MP3, OGG, AAC, M4A, FLAC, JPG (and many more!) files. * Media thumbnail support * Web UI with a tree view of the database and the file system, allowing to add/remove/edit/browse your media * Highly flexible media format transcoding via plugins / scripts * Automatic directory rescans (timed, inotify) * User defined server layout based on extracted metadata * On the fly video thumbnail generation with libffmpegthumbnailer * Support for external URLs (create links to internet content and serve them via UPnP to your renderer) Maintainer: The OpenBSD ports mailing-list WWW: https://github.com/gerbera/gerbera gerbera.tgz Description: Binary data
Re: [maintainer update] www/p5-Mojo 8.18
Hi, On Sun, 30 Jun 2019 20:11:01 +0300 Manolis Tzanidakis wrote: > Hello, > this diff updates www/p5-Mojo to 8.18. All tests pass. Thanks! I've found no issues while testing direct consumers as well. We should move to PERMIT_PACKAGE though. OK cwen@ (i'll commit and make the PERMIT_PACKAGE change with another OK). > Changes: https://metacpan.org/release/SRI/Mojolicious-8.18 > > Index: Makefile > === > RCS file: /cvs/ports/www/p5-Mojo/Makefile,v > retrieving revision 1.34 > diff -u -p -u -r1.34 Makefile > --- Makefile 30 Apr 2019 02:11:35 - 1.34 > +++ Makefile 30 Jun 2019 17:06:43 - > @@ -4,7 +4,7 @@ COMMENT = next generation web framework > > MODULES =cpan > PKG_ARCH = * > -DISTNAME = Mojolicious-8.15 > +DISTNAME = Mojolicious-8.18 > CATEGORIES = www > > MAINTAINER = Manolis Tzanidakis > Index: distinfo > === > RCS file: /cvs/ports/www/p5-Mojo/distinfo,v > retrieving revision 1.26 > diff -u -p -u -r1.26 distinfo > --- distinfo 30 Apr 2019 02:11:35 - 1.26 > +++ distinfo 30 Jun 2019 17:06:43 - > @@ -1,2 +1,2 @@ > -SHA256 (Mojolicious-8.15.tar.gz) = > yFArI+TyE1RIYfIECjTf6nLj0QkzVcJMcPEtZdB37Ns= -SIZE > (Mojolicious-8.15.tar.gz) = 755156 +SHA256 (Mojolicious-8.18.tar.gz) > = /naCG/ir5sZKar4uMnIQ9WSgycsMX03DLC26DUO3AG4= +SIZE > (Mojolicious-8.18.tar.gz) = 761103 >
Re: llvm-7.0.1 fallout: lang/crystal
On Sun, Jun 30, 2019 at 06:18:38PM +0200, Jeremie Courreges-Anglas wrote: > > Hi Wesley, > > On Wed, Feb 13 2019, Wesley Moxam wrote: > >> On Feb 13, 2019, at 7:47 AM, Jeremie Courreges-Anglas > >> wrote: > >> > >>> On Thu, Feb 07 2019, Stuart Henderson wrote: > On 2019/02/07 18:02, Jeremie Courreges-Anglas wrote: > > Hi, > > I'm not sure you were notified so here's a heads-up: lang/crystal > doesn't build any more after the update to clang 7 in both base and > ports. > >>> > >>> Not handled upstream yet. > >>> https://github.com/crystal-lang/crystal/issues/6754 > >> > >> Wesley, I see that you have commented on this issue. Do you mind if > >> we mark lang/crystal as temporarily BROKEN? > > > > That’s fine with me. It may be a while before the issue is fixed > > Any news regarding lang/crystal? The issue is still open and "@bcardiff > removed this from the 0.28.0 milestone on Mar 27". > > If upstream doesn't plan to support newer llvm releases and no > alternative fix is found, then the port should be removed. I don't know much about it, but can't it be built with ports gcc? -- Antoine
[maintainer update] www/p5-Mojo 8.18
Hello, this diff updates www/p5-Mojo to 8.18. All tests pass. Changes: https://metacpan.org/release/SRI/Mojolicious-8.18 Index: Makefile === RCS file: /cvs/ports/www/p5-Mojo/Makefile,v retrieving revision 1.34 diff -u -p -u -r1.34 Makefile --- Makefile30 Apr 2019 02:11:35 - 1.34 +++ Makefile30 Jun 2019 17:06:43 - @@ -4,7 +4,7 @@ COMMENT = next generation web framework MODULES = cpan PKG_ARCH = * -DISTNAME = Mojolicious-8.15 +DISTNAME = Mojolicious-8.18 CATEGORIES = www MAINTAINER = Manolis Tzanidakis Index: distinfo === RCS file: /cvs/ports/www/p5-Mojo/distinfo,v retrieving revision 1.26 diff -u -p -u -r1.26 distinfo --- distinfo30 Apr 2019 02:11:35 - 1.26 +++ distinfo30 Jun 2019 17:06:43 - @@ -1,2 +1,2 @@ -SHA256 (Mojolicious-8.15.tar.gz) = yFArI+TyE1RIYfIECjTf6nLj0QkzVcJMcPEtZdB37Ns= -SIZE (Mojolicious-8.15.tar.gz) = 755156 +SHA256 (Mojolicious-8.18.tar.gz) = /naCG/ir5sZKar4uMnIQ9WSgycsMX03DLC26DUO3AG4= +SIZE (Mojolicious-8.18.tar.gz) = 761103
Re: llvm-7.0.1 fallout: lang/crystal
Hi Wesley, On Wed, Feb 13 2019, Wesley Moxam wrote: >> On Feb 13, 2019, at 7:47 AM, Jeremie Courreges-Anglas >> wrote: >> >>> On Thu, Feb 07 2019, Stuart Henderson wrote: On 2019/02/07 18:02, Jeremie Courreges-Anglas wrote: Hi, I'm not sure you were notified so here's a heads-up: lang/crystal doesn't build any more after the update to clang 7 in both base and ports. >>> >>> Not handled upstream yet. >>> https://github.com/crystal-lang/crystal/issues/6754 >> >> Wesley, I see that you have commented on this issue. Do you mind if >> we mark lang/crystal as temporarily BROKEN? > > That’s fine with me. It may be a while before the issue is fixed Any news regarding lang/crystal? The issue is still open and "@bcardiff removed this from the 0.28.0 milestone on Mar 27". If upstream doesn't plan to support newer llvm releases and no alternative fix is found, then the port should be removed. -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
Re: dedicated user for sysutils/monit
On Fri, Jun 28, 2019 at 02:49:26PM +0200, Joel Carnat wrote: > BTW, following stu@'s "(...) I think it really needs more support (...)" > remark, I searched for things that would break if Monit would not run as > root. I found that the "network ping test" requires root access to run. > I don't use it myself so I didn't notice it when running as _monit. > Documentation says: "Monit must also run as the root user in order to be > able to perform the ping test (because the ping test must use raw > sockets which usually only the super user is allowed to)." For the reason that you mentioned above, I don't think it is a good idea to make monit run as a non-root user by default. As you noticed, monit doesn't appear to be designed to be run as a non-root user. Best regards, Caspar Schutijser
Re: NEW: Tacacs+ port - shrubbery.net version
On 2019/05/23 20:09, Jan Vlach wrote: > Hi Gleydson, Stuart, ports, > > I'm running tac_plus with 200+ boxes with IOS, IOS-XE and IOS-XR. > > please see attached tgz for updated port. > > - I've taken Gleydson's latest work from openbsd-wip (I don't see the > unexec and/or doc/shared implemented in PLIST) * > - provided simplified tac_plus.conf.sample of stuff I have tested - > logging in as full admins with level 15 and limited show users that I > use for scripting/metrics. I can't really vouch for the functionality of > dialup users etc. The full-blown config file example is still in the > manpage > - fixed typo in manpage for accounting to syslog - using `accounting > syslog;` (including semicolon) does not work, but parser does not > complain. If I remove the semicolon, accounting info gets logged to > syslog as daemon.info (this was nasty :) ) > - fixed paths for tac.acct, tac.log and tac.who - all of them go to > /var/log/tac_plus directory that's owned by _tacacs:_tacacs > - ^ This fixes the case where you don't want to log into accounting file > and want syslog accounting only (disabling accounting file directive > leads to tacacs complaining of permission denied with with default path > of /var/log/tac.acct) Changing the default path to > /var/log/tac_plus/tac.acct and removing `accounting file = ...' > directive properly disables logging to this file. Go figure :) > - Updated paths in manpage (tac_plus.conf.5.in) as one is automatically > substituted from configure variables, while the other is hardcoded. > - Added README file to remind administrator to rotate his/her files. > > * I've tried to add the @extraunexec rm -rf /var/log/tac_plus/*, but I'm > not sure it works: > > On package deletion pkg_delete complains that directory is not empty: > [20:07][root@samsara:/var/log]# pkg_delete tacacs+ > tacacs+-4.0.4.28v0: ok > Read shared items: ok > --- -tacacs+-4.0.4.28v0 --- > You should also remove /etc/tac_plus.conf (which was modified) > You should also run rm -f /var/log/tac_plus/* > Error deleting directory /var/log/tac_plus: Directory not empty > You should also run /usr/sbin/userdel _tacacs > You should also run /usr/sbin/groupdel _tacacs > > I'm sorry, I've wrestled, but I don't understand how the doc/examples > directories work - > what needs to be done in pkg configure phase and what is done in PLIST? > > Cluestick please? > > I've tested the accounting part with py-tacacs_plus on -current, don't have a > real > network box around at this time. (Gonna dogfood this tomorrow or next > week) > > Could you please have a look if this is okay? > > jvl > > On Thu, May 23, 2019 at 11:34:23AM -0300, Gleydson Soares wrote: > > > Can you use the standard locations for doc/examples please rather > > > than /usr/local/share/tacacs? > > > > Yep. > > > > > Needs @extraunexec rm -f /var/log/tac_plus/* for pkg_delete -c. > > > > Done. > > Thanks for the feedback, i'm pushing it to openbsd-wip. > > > > PS.: I'm running it and works just fine It has a dozen of Cisco Nexus > > switches already connected. > > privdrop (_tacacs) fine. > > > > I will add some changes to example files provided by Jan Vlach, for > > pointing out how to use tac_plus on the fly on OpenBSD.(like features > > available with and without privdrop / etc). > > > > Also should be nice sent patches upstream. Jan Vlach, what do you think > > about? > > > > Cheers, > > Slightly tweaked version attached, this one's ok with me: - https homepage - PERMIT_*_CDROM is not used for new ports - whitespace nit in Makefile - tweak comment in patch - place @extraunexec above the @sample line, that way pkg_delete -c doesn't complain about a missing dir. (pkg_delete without -c will complain about not being able to remove the dir, that is no problem). - regen plist to include pkg-readme - adjust pkg-readme to set uid/gid on the files - change group ownership of log dir to wheel, easier for admins tacacs+,3.tgz Description: application/tar-gz
Re: suricata, librsvg : set CARGO_HOME for upcoming lang/rust-1.36
On Sun, Jun 30, 2019 at 02:02:47PM +0200, Antoine Jacoutot wrote: > > Why not setting PORTHOME=${WRKDIST} by default for cargo ports? Setting CARGO_HOME is enough for this purpose and should affect only cargo. At opposite setting PORTHOME could have unwanted side-effects, and we will not catch write attempt in HOME any more for such ports. It is why I prefer using CARGO_HOME in environment when possible. Thanks. -- Sebastien Marie
Re: suricata, librsvg : set CARGO_HOME for upcoming lang/rust-1.36
On Sun, Jun 30, 2019 at 12:41:02PM +0200, Sebastien Marie wrote: > Hi, > > I am working on updating lang/rust to 1.36.0 > > The release ships a new version of cargo which will want to write a file > in $HOME (a lock file to manage concurrent access on package cache). > > To avoid an error while acquiring package cache lock, CARGO_HOME should > be defined to a directory writable by the build. The default value of > CARGO_HOME is "$HOME/.cargo", so alternatively PORTHOME in the port > could be set too. > > As i think it is less invasive to set CARGO_HOME, I followed this way. > > The following diff takes care of: > - security/suricata : configure look for CARGO_HOME env var > - x11/gnome/librsvg : set CARGO_HOME via MAKE_ENV > > Some others ports not using devel/cargo sets PORTHOME (hey firefox o/) > so the problem doesn't occurs. And ports using devel/cargo modules > already defines CARGO_HOME by default. > > One possible leftover is devel/meson where lang/rust is a TEST_DEPS. But > it seems it is directly using rustc compiler and not cargo. So I assume > it is fine. > > No package bump as it is a build setting that don't affect anything with > rust-1.35 we have in port. > > Comments or OK ? > -- > Sebastien Marie Why not setting PORTHOME=${WRKDIST} by default for cargo ports? > > Index: security/suricata/Makefile > === > RCS file: /cvs/ports/security/suricata/Makefile,v > retrieving revision 1.20 > diff -u -p -r1.20 Makefile > --- security/suricata/Makefile3 May 2019 06:22:34 - 1.20 > +++ security/suricata/Makefile30 Jun 2019 10:05:52 - > @@ -49,7 +49,8 @@ CONFIGURE_STYLE = autoconf > AUTOCONF_VERSION = 2.69 > > CONFIGURE_ENV = ac_cv_path_HAVE_PDFLATEX= \ > - ac_cv_path_HAVE_GIT_CMD= > + ac_cv_path_HAVE_GIT_CMD= \ > + CARGO_HOME=${WRKBUILD}/cargo-home > > CONFIGURE_ARGS = --disable-gccmarch-native \ > --enable-ipfw > Index: x11/gnome/librsvg/Makefile > === > RCS file: /cvs/ports/x11/gnome/librsvg/Makefile,v > retrieving revision 1.148 > diff -u -p -r1.148 Makefile > --- x11/gnome/librsvg/Makefile13 May 2019 22:47:45 - 1.148 > +++ x11/gnome/librsvg/Makefile30 Jun 2019 07:25:33 - > @@ -19,7 +19,8 @@ SHARED_LIBS += rsvg-2 39. > GNOME_VERSION= ${STABLE_VERSION} > BUILD_DEPENDS= lang/rust > PKG_ARGS=-Dold=0 -Dstable=1 > -MAKE_ENV+= CARGO_BUILD_JOBS=${MAKE_JOBS} > +MAKE_ENV+= CARGO_BUILD_JOBS=${MAKE_JOBS} \ > + CARGO_HOME=${WRKBUILD}/cargo-home > .else > ### old > REVISION=3 > -- Antoine
Re: [NEW] textproc/py-recommonmark
On 2019/06/29 22:31, Daniel Jakots wrote: > On Sat, 29 Jun 2019 21:41:50 -0400, Kurt Mosiejczuk > wrote: > > > Also, do you wish to have the package named recommonmark-0.5.0 or > > py-recommonmark-0.5.0? Right now it is the former since there isn't a > > PKGNAME line. > > Since Jérémie chose to have a flavor, it needs to be py- prefixed. > > > Also, setting MODPY_EGG_VERSION is useless or I'm missing something? It avoids unneeded churn in PLISTs for simple version updates
Re: infrastructure patch: improve test handling
On Sat, Jun 29 2019, Marc Espie wrote: > Documentation AND infra patch. > > - add test to the list of things that can be rebuilt/cleaned This part needs a fix for PORTS_PRIVSEP=Yes, please see below. > - recognize PORTS_PRIVSEP as a special case for automated testing, > specifically, set TEST_IS_INTERACTIVE=network for testsuites that require > network access. Not reviewed / tested yet. > > > Index: bsd.port.mk.5 > === > RCS file: /cvs/src/share/man/man5/bsd.port.mk.5,v > retrieving revision 1.511 > diff -u -p -r1.511 bsd.port.mk.5 > --- bsd.port.mk.5 31 May 2019 21:27:48 - 1.511 > +++ bsd.port.mk.5 29 Jun 2019 12:39:47 - > @@ -169,7 +169,7 @@ Clean ports contents. > By default, it will clean the work directory. > It can be invoked as > make clean='[depends build bulk work fake flavors dist install sub package > -packages plist]'. > +packages plist test]'. > .Bl -tag -width packages > .It Va work > Clean work directory. > @@ -195,6 +195,8 @@ Uninstall package. > Remove all copies of package file. > .It Va plist > Remove registered packing lists of all subpackages. > +.It Va test > +Clean test cookie. > .It Va sub > With > .Va install > @@ -702,8 +704,11 @@ see > .It Cm reprepare > Force running the > .Ar prepare > -target > -again. > +target again. > +.It Cm retest > +Force running the > +.Ar test > +target again. > .It Cm show > Invoked as make show=name, show the contents of ${name}. > Invoked as make show="name1 name2 ...", > @@ -3015,9 +3020,18 @@ Empty by default. > .It Ev TEST_IS_INTERACTIVE > Set to > .Sq Yes > -if port needs human interaction to run its tests, or set to > +if port needs human interaction to run its tests, set to > .Sq X11 > -if the tests need an active X11 display to work. > +if the tests need an active X11 display to work, > +set to > +.Sq network > +if the tests need access network > +(setting > +.Ev PORTS_PRIVSEP > +disables network access for the > +.Sq _pbuild > +user, so these tests become, in effect, > +interactive). > .It Ev TEST_LOG > Command used to log the results of regression tests to TEST_LOGFILE. > Read-only. > > > > > > Index: bsd.port.mk > === > RCS file: /cvs/ports/infrastructure/mk/bsd.port.mk,v > retrieving revision 1.1471 > diff -u -p -r1.1471 bsd.port.mk > --- bsd.port.mk 16 Jun 2019 14:27:42 - 1.1471 > +++ bsd.port.mk 29 Jun 2019 12:39:53 - > @@ -253,7 +253,7 @@ _clean += -f > .endif > > _okay_words = depends work fake -f flavors dist install sub packages package > \ > - bulk force plist build all > + bulk force plist build all test > .for _w in ${_clean} > . if !${_okay_words:M${_w}} > ERRORS += "Fatal: unknown clean command: ${_w}\n(not in ${_okay_words})" > @@ -928,8 +928,13 @@ TEST_LOGFILE ?= ${WRKDIR}/test.log > TEST_LOG = | ${_PBUILD} tee ${TEST_LOGFILE} > IS_INTERACTIVE ?= No > TEST_IS_INTERACTIVE ?= No > +.if ${TEST_IS_INTERACTIVE:L} == "network" && ${PORTS_PRIVSEP:L} != "yes" > +_TEST_IS_INTERACTIVE = No > +.else > +_TEST_IS_INTERACTIVE = ${TEST_IS_INTERACTIVE} > +.endif > > -.if ${TEST_IS_INTERACTIVE:L} == "x11" > +.if ${_TEST_IS_INTERACTIVE:L} == "x11" > TEST_ENV += DISPLAY=${DISPLAY} XAUTHORITY=${XAUTHORITY} > XAUTHORITY ?= ${HOME}/.Xauthority > .endif > @@ -1422,9 +1427,9 @@ MISSING_FILES += ${_F} > > TRY_BROKEN ?= No > _IGNORE_TEST ?= > -.if ${TEST_IS_INTERACTIVE:L} != "no" && defined(BATCH) > +.if ${_TEST_IS_INTERACTIVE:L} != "no" && defined(BATCH) > _IGNORE_TEST += "has interactive tests" > -.elif ${TEST_IS_INTERACTIVE:L} == "no" && defined(INTERACTIVE) > +.elif ${_TEST_IS_INTERACTIVE:L} == "no" && defined(INTERACTIVE) > _IGNORE_TEST += "does not have interactive tests" > .endif > > @@ -2826,7 +2831,7 @@ ${_TEST_COOKIE}: ${_BUILD_COOKIE} > .if ${NO_TEST:L} == "no" > @${ECHO_MSG} "===> Regression tests for ${FULLPKGNAME}${_MASTER}" > # When interactive tests need X11 > -. if ${TEST_IS_INTERACTIVE:L} == "x11" > +. if ${_TEST_IS_INTERACTIVE:L} == "x11" > .if !defined(DISPLAY) || !exists(${XAUTHORITY}) > @echo 1>&2 "The regression tests require a running instance of X." > @echo 1>&2 "You will also need to set the environment variable DISPLAY" > @@ -3151,6 +3156,9 @@ _internal-clean: > .endfor > . endif > .endif > +.if ${_clean:Mtest} > + rm -f ${_TEST_COOKIE} Should be > + ${_PBUILD} rm -f ${_TEST_COOKIE} else I get an error with make clean=test: russell /usr/ports/x11/ratpoison$ make clean=test ===> Cleaning for ratpoison-1.4.9 rm -f /usr/ports/pobj/ratpoison-1.4.9/build-amd64/.test_done rm: /usr/ports/pobj/ratpoison-1.4.9/build-amd64/.test_done: Permission denied *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:3160 '_internal-clean') *** Error 1 in /usr/ports/x11/ratpoison (/usr/ports/in
Re: bsd.port.mk: make update-patches depend on patch
On Mon, Jun 24, 2019 at 11:29:37PM +0100, Stuart Henderson wrote: > On 2019/06/25 00:06, Klemens Nanni wrote: > > Logically obvious to me; any reasons not to do it? > > Yes - "make patch" may have failed but you still want to update files > for whichever of the patches *did* manage to get applied. Fair enough,thanks. I probably never ran into this because I either try fixing patches before running `update-patches' or move them aside to clean, apply and fix them one by one. As for the alias, that's what I wanted to avoid. I'd prefer our framework to provide comprehensive tooling and targets so developers do not have to augment their environment that much, although such an alias really isn't much either.
suricata, librsvg : set CARGO_HOME for upcoming lang/rust-1.36
Hi, I am working on updating lang/rust to 1.36.0 The release ships a new version of cargo which will want to write a file in $HOME (a lock file to manage concurrent access on package cache). To avoid an error while acquiring package cache lock, CARGO_HOME should be defined to a directory writable by the build. The default value of CARGO_HOME is "$HOME/.cargo", so alternatively PORTHOME in the port could be set too. As i think it is less invasive to set CARGO_HOME, I followed this way. The following diff takes care of: - security/suricata : configure look for CARGO_HOME env var - x11/gnome/librsvg : set CARGO_HOME via MAKE_ENV Some others ports not using devel/cargo sets PORTHOME (hey firefox o/) so the problem doesn't occurs. And ports using devel/cargo modules already defines CARGO_HOME by default. One possible leftover is devel/meson where lang/rust is a TEST_DEPS. But it seems it is directly using rustc compiler and not cargo. So I assume it is fine. No package bump as it is a build setting that don't affect anything with rust-1.35 we have in port. Comments or OK ? -- Sebastien Marie Index: security/suricata/Makefile === RCS file: /cvs/ports/security/suricata/Makefile,v retrieving revision 1.20 diff -u -p -r1.20 Makefile --- security/suricata/Makefile 3 May 2019 06:22:34 - 1.20 +++ security/suricata/Makefile 30 Jun 2019 10:05:52 - @@ -49,7 +49,8 @@ CONFIGURE_STYLE = autoconf AUTOCONF_VERSION = 2.69 CONFIGURE_ENV =ac_cv_path_HAVE_PDFLATEX= \ - ac_cv_path_HAVE_GIT_CMD= + ac_cv_path_HAVE_GIT_CMD= \ + CARGO_HOME=${WRKBUILD}/cargo-home CONFIGURE_ARGS = --disable-gccmarch-native \ --enable-ipfw Index: x11/gnome/librsvg/Makefile === RCS file: /cvs/ports/x11/gnome/librsvg/Makefile,v retrieving revision 1.148 diff -u -p -r1.148 Makefile --- x11/gnome/librsvg/Makefile 13 May 2019 22:47:45 - 1.148 +++ x11/gnome/librsvg/Makefile 30 Jun 2019 07:25:33 - @@ -19,7 +19,8 @@ SHARED_LIBS += rsvg-2 39. GNOME_VERSION= ${STABLE_VERSION} BUILD_DEPENDS= lang/rust PKG_ARGS= -Dold=0 -Dstable=1 -MAKE_ENV+= CARGO_BUILD_JOBS=${MAKE_JOBS} +MAKE_ENV+= CARGO_BUILD_JOBS=${MAKE_JOBS} \ + CARGO_HOME=${WRKBUILD}/cargo-home .else ### old REVISION= 3
Re: [NEW] textproc/py-recommonmark
On Sat, Jun 29 2019, Kurt Mosiejczuk wrote: > On Sat, Jun 29, 2019 at 10:31:52PM -0400, Daniel Jakots wrote: >> On Sat, 29 Jun 2019 21:41:50 -0400, Kurt Mosiejczuk >> wrote: > >> > Also, do you wish to have the package named recommonmark-0.5.0 or >> > py-recommonmark-0.5.0? Right now it is the former since there isn't a >> > PKGNAME line. > >> Since Jeremie chose to have a flavor, it needs to be py- prefixed. > > Good point. Duh, indeed. >> Also, setting MODPY_EGG_VERSION is useless or I'm missing something? > > It's not necessary if using the GH_* tags, but if there is intention to > move it to PyPI in the future, it makes sense. The package started as a MODPY_PI port, but I indeed decided to go for the github repo to get more tests to run. It's not perfect yet, as I'm not sure I understand the cause for the "KeyError: 'refdoc'" tracebacks. Updated tarball attached, diff below. Thanks for your help. --8<-- diff -pruN py-recommonmark.old/Makefile py-recommonmark/Makefile --- py-recommonmark.old/MakefileSat Jun 29 07:24:49 2019 +++ py-recommonmark/MakefileSun Jun 30 12:21:24 2019 @@ -2,10 +2,13 @@ COMMENT = markdown parser for docutils -MODPY_EGG_VERSION =0.5.0 +MODPY_EGG_VERSION =0.5.0 +# Using github autogenerated tarball because pypi tarballs don't ship +# files needed by tests. GH_ACCOUNT = readthedocs GH_PROJECT = recommonmark GH_TAGNAME = ${MODPY_EGG_VERSION} +PKGNAME = py-recommonmark-${MODPY_EGG_VERSION} CATEGORIES = textproc devel @@ -20,7 +23,6 @@ MODPY_SETUPTOOLS =Yes RUN_DEPENDS = textproc/py-commonmark${MODPY_FLAVOR} \ textproc/py-docutils${MODPY_FLAVOR} \ textproc/py-sphinx${MODPY_FLAVOR} -TEST_DEPENDS = ${RUN_DEPENDS} FLAVORS = python3 FLAVOR ?= -->8-- py-recommonmark.2.tgz Description: Binary data -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE