Re: [patch] Test target PATH
On Thu, Aug 27, 2015 at 11:04 PM, Matthew Martin wrote: > On Thu, Aug 27, 2015 at 11:47:19PM +0200, Juan Francisco Cantero Hurtado > wrote: >> >---rm -f ${WRKDIR}/bin/chgrp > > While this would fix the problem, since the problem is not in zsh > itself, I don't think that the zsh port is the right place to fix this. Since the problem is not in zsh itself, that means that zsh itself is not the right place to fix this. Since the bug only occurs in the port of zsh, that means that the zsh port is the right place to fix this. Or am I missing some important issue? (If so, could you explain what breaks?) Thanks, -- Raul
Re: [patch] Test target PATH
On Thu, Aug 27, 2015 at 11:47:19PM +0200, Juan Francisco Cantero Hurtado wrote: > I see the problem now. You can remove the symlinks from ${WRKDIR}/bin with: > > pre-test: > >---rm -f ${WRKDIR}/bin/chgrp While this would fix the problem, since the problem is not in zsh itself, I don't think that the zsh port is the right place to fix this. > The problem with your patch is that you're modifying the environment > for thousands of ports and we can't test manually every port to see if > your changes are breaking something. However, overriding chown and chgrp during test is a bug. They clearly don't need to be overridden during test since the symlinks only exist if test is called after fake and don't exist for make clean test. ${WRKDIR}/bin is only in PATH during extract, configure, build, fake, and test. Since fake creates the chown and chgrp symlinks, extract, configure, build, and test must not depend on those symlinks as they don't depend on fake. Would it be acceptable to clean up the symlinks after fake completes or rm -f them in test in case test is run after fake? Both patches below fix the issue. However, I still believe that the first patch is correct as any ports it breaks has tests that depend on non-standard install(1) behavior caused by the perl wrapper so including ${WRKDIR}/bin in PATH is just concealing the breakage. - Matthew Martin Index: infrastructure/mk/bsd.port.mk === RCS file: /cvs/ports/infrastructure/mk/bsd.port.mk,v retrieving revision 1.1298 diff -u -p -r1.1298 bsd.port.mk --- infrastructure/mk/bsd.port.mk 19 Jul 2015 17:31:44 - 1.1298 +++ infrastructure/mk/bsd.port.mk 28 Aug 2015 01:58:04 - @@ -2886,6 +2886,8 @@ ${_FAKE_COOKIE}: ${_BUILD_COOKIE} ${_FAK fi; \ done + @rm -f ${WRKDIR}/bin/chown ${WRKDIR}/bin/chgrp + @${_FAKESUDO} ${_MAKE_COOKIE} $@ print-plist: Index: infrastructure/mk/bsd.port.mk === RCS file: /cvs/ports/infrastructure/mk/bsd.port.mk,v retrieving revision 1.1298 diff -u -p -r1.1298 bsd.port.mk --- infrastructure/mk/bsd.port.mk 19 Jul 2015 17:31:44 - 1.1298 +++ infrastructure/mk/bsd.port.mk 28 Aug 2015 02:06:13 - @@ -2786,6 +2786,7 @@ ${_TEST_COOKIE}: ${_BUILD_COOKIE} @exit 1 .endif . endif + @rm -f ${WRKDIR}/bin/chown ${WRKDIR}/bin/chgrp . if target(pre-test) @${_MAKE} pre-test . endif
Re: [UPDATE] audio/picard to 1.3.2
Hi ports@ Attached and below are the diff's to update audio/picard to 1.3.2 This update removes the need for the patches and has been tested on amd64 and i386 - I'm just bringing my macppc up to current to test there as well. Cheers Fred Index: Makefile === RCS file: /cvs/ports/audio/picard/Makefile,v retrieving revision 1.18 diff -u -p -u -p -r1.18 Makefile --- Makefile30 Jul 2015 15:09:27 - 1.18 +++ Makefile27 Aug 2015 15:20:02 - @@ -2,9 +2,8 @@ COMMENT = automatic audio files tagger using the MusicBrainz database -MODPY_EGG_VERSION =0.16 +MODPY_EGG_VERSION =1.3.2 DISTNAME = picard-${MODPY_EGG_VERSION} -REVISION = 1 CATEGORIES = audio @@ -18,8 +17,9 @@ PERMIT_PACKAGE_CDROM =Yes WANTLIB += ${MODPY_WANTLIB} WANTLIB += avcodec avformat m ofa pthread stdc++ -MASTER_SITES = http://ftp.musicbrainz.org/pub/musicbrainz/picard/ \ - ftp://ftp.musicbrainz.org/pub/musicbrainz/picard/ +GH_ACCOUNT = musicbrainz +GH_PROJECT = picard +GH_TAGNAME = ${MODPY_EGG_VERSION} MODULES = lang/python @@ -39,11 +39,14 @@ RUN_DEPENDS = x11/py-qt4>=4.1 \ TEST_DEPENDS = ${BASE_PKGPATH} -PLUGINS_DIRS = lastfm \ - lastfmplus \ - replaygain -PLUGINS = classicdiscnumber.py \ - coverart.py \ +PLUGINS_DIRS = lastfm \ + lastfmplus \ + replaygain \ + viewvariables + +PLUGINS = addrelease.py \ + albumartist_website.py \ + classicdiscnumber.py \ cuesheet.py \ discnumber.py \ featartist.py \ @@ -54,14 +57,17 @@ PLUGINS = classicdiscnumber.py \ lastfmplus/__init__.py \ lastfmplus/ui_options_lastfm.py \ no_release.py \ - open_in_gui.py \ release_type.py \ replaygain/__init__.py \ replaygain/options_replaygain.ui \ replaygain/ui_options_replaygain.py \ - swapprefix.py \ + sort_multivalue_tags.py \ + standardise_performers.py \ titlecase.py \ - tracks2clipboard.py + tracks2clipboard.py \ + viewvariables/__init__.py \ + viewvariables/ui_variables_dialog.py \ + viewvariables/variables_dialog.ui pre-configure: ${SUBST_CMD} ${WRKSRC}/setup.py Index: distinfo === RCS file: /cvs/ports/audio/picard/distinfo,v retrieving revision 1.4 diff -u -p -u -p -r1.4 distinfo --- distinfo18 Jan 2015 03:12:47 - 1.4 +++ distinfo27 Aug 2015 15:20:02 - @@ -1,2 +1,2 @@ -SHA256 (picard-0.16.tar.gz) = ViMhb+3fYAAYfOrTdGs66+X2A4N+MTx9/Jl4NXb1Sxg= -SIZE (picard-0.16.tar.gz) = 1828595 +SHA256 (picard-1.3.2.tar.gz) = y2owR+dW56UNOIv/F5SvYnTnvZIA2V0nK7CPZVi9mWs= +SIZE (picard-1.3.2.tar.gz) = 2047602 Index: patches/patch-picard_musicdns_avcodec_c === RCS file: patches/patch-picard_musicdns_avcodec_c diff -N patches/patch-picard_musicdns_avcodec_c --- patches/patch-picard_musicdns_avcodec_c 5 May 2014 07:55:50 - 1.1 +++ /dev/null 1 Jan 1970 00:00:00 - @@ -1,38 +0,0 @@ -$OpenBSD: patch-picard_musicdns_avcodec_c,v 1.1 2014/05/05 07:55:50 brad Exp $ - -Deal with newer FFmpeg API. - picard/musicdns/avcodec.c.orig Sun Oct 23 13:15:29 2011 -+++ picard/musicdns/avcodec.c Thu May 1 04:58:17 2014 -@@ -39,6 +39,10 @@ - #define AVMEDIA_TYPE_AUDIO CODEC_TYPE_AUDIO - #endif - -+#ifndef AVCODEC_MAX_AUDIO_FRAME_SIZE -+#define AVCODEC_MAX_AUDIO_FRAME_SIZE 192000 // 1 second of 48khz 32bit audio -+#endif -+ - static PyObject * - init(PyObject *self, PyObject *args) - { -@@ -55,7 +59,7 @@ done(PyObject *self, PyObject *args) - static PyObject * - decode(PyObject *self, PyObject *args) - { --AVFormatContext *format_context; -+AVFormatContext *format_context = NULL; - AVCodecContext *codec_context; - AVCodec *codec; - PyObject *filename; -@@ -69,7 +73,11 @@ decode(PyObject *self, PyObject *args) - return NULL; - - Py_UNBLOCK_THREADS -+#if LIBAVFORMAT_VERSION_INT < AV_VERSION_INT(53, 2, 0) - if (av_open_input_file(&format_context, PyString_AS_STRING(filename), NULL, 0, NULL) != 0) { -+#else -+if (avformat_open_input(&format_context, PyString_AS_STRING(filename), NULL, NULL) != 0) { -+#endif - Py_BLOCK_THREADS - PyErr_SetString(
Re: [patch] Test target PATH
I see the problem now. You can remove the symlinks from ${WRKDIR}/bin with: pre-test: >---rm -f ${WRKDIR}/bin/chgrp The problem with your patch is that you're modifying the environment for thousands of ports and we can't test manually every port to see if your changes are breaking something. On Thu, Aug 27, 2015 at 12:47:24PM -0500, Matthew Martin wrote: > On Thu, Aug 27, 2015 at 11:50:19AM +0200, Juan Francisco Cantero Hurtado > wrote: > > Can you try this in your port? > > > > TEST_ENV="PATH=${PATH}:YOURPATH" or TEST_ENV="PATH=YOURPATH:${PATH}" > > I don't think you're following the problem. It's not that anything needs > to be added to PATH; it's that the first element needs to be removed. > The PATH that ports uses is set by PORTPATH to be > > > ${WRKDIR}/bin:/usr/bin:/bin:/usr/sbin:/sbin:${DEPBASE}/bin:${LOCALBASE}/bin:${X11BASE}/bin > > With make test, ${WRKDIR}/bin only has install in it, so chgrp runs > /bin/chgrp. When make fake is run, a symlink from ${WRKDIR}/bin/chgrp to > /bin/echo is added. This resuls in make fake test using the chgrp in > ${WRKDIR}/bin and the test failing since that chgrp doesn't change the > group. I am proposing to remove ${WRKDIR}/bin from PATH for tests so > that all commands work as expected. > > -Matthew Martin > > > On Thu, Aug 27, 2015 at 12:37:22AM -0500, Matthew Martin wrote: > > > The zsh port fails an additional test when invoked with make clean fake > > > test instead of make clean test. This is because in its tests it > > > > > > touch zerolength > > > chgrp $EGID zerolength > > > > > > and then tests that zerolength has a group of EGID. Since make fake puts > > > a symlink in ${WRKDIR}/bin from chgrp to /bin/echo and ${WRKDIR}/bin is > > > first in the environment's PATH, the chgrp does nothing, and test > > > C02cond fails. With the below patch it completes successfully. > > > > > > This patch does slightly change behavior as PATH can no longer be set in > > > TEST_ENV (which no ports seem to do anyway) and install is no longer the > > > perl wrapper in ${WRKDIR}/bin; however, during tests we probably > > > shouldn't be messing with install either. > > > > > > - Matthew Martin > > > > > > > > > Index: infrastructure/mk/bsd.port.mk > > > === > > > RCS file: /cvs/ports/infrastructure/mk/bsd.port.mk,v > > > retrieving revision 1.1298 > > > diff -u -p -r1.1298 bsd.port.mk > > > --- infrastructure/mk/bsd.port.mk 19 Jul 2015 17:31:44 - 1.1298 > > > +++ infrastructure/mk/bsd.port.mk 27 Aug 2015 03:53:55 - > > > @@ -809,7 +809,7 @@ FAKE_TARGET ?= ${INSTALL_TARGET} > > > > > > TEST_TARGET ?= test > > > TEST_FLAGS ?= > > > -TEST_ENV ?= > > > +TEST_ENV += > > > PATH=/usr/bin:/bin:/usr/sbin:/sbin:${DEPBASE}/bin:${LOCALBASE}/bin:${X11BASE}/bin > > > ALL_TEST_FLAGS = ${MAKE_FLAGS} ${TEST_FLAGS} > > > ALL_TEST_ENV = ${MAKE_ENV} ${TEST_ENV} > > > TEST_LOGFILE ?= ${WRKDIR}/test.log > > > -- Juan Francisco Cantero Hurtado http://juanfra.info
Building ports with the Clang static analyzer
I've been having trouble with this. It's probably because the port I'm testing (my WIP of Pidgin 3) uses autogen.sh for both configuration and compilation. I've tried the obvious way ('scan-build make build'). In that case I only get a report for the autogen.sh test programs, although the main compilation also uses the static analyzer. I've also tried using a couple different env vars to prepend scan-build, but they caused errors. Any tips?
NEW: fonts/junicode
(Re-) Submitting the first (of many) ports for fonts/ Junicode is a Unicode font for medievalists. Created by Peter S. Baker of the University of Virginia, Junicode is based on a 17th century typeface used in Oxford, England. Junicode contains many special characters and ligatures for medievalists, along with numerous other Unicode glyphs. g junicode.tgz Description: Binary data
Re: NEW: security/lastpass-cli
On Thu 27/08/2015 17:51, Stuart Henderson wrote: > On 2015/08/27 17:26, Björn Ketelaars wrote: > > On Thu 27/08/2015 12:35, Stuart Henderson wrote: > > > On 2015/03/17 07:11, Björn Ketelaars wrote: > > > > Please find enclosed a port for security/lastpass-cli [0]. lastpass-cli > > > > is an > > > > open-source command-line interface to the popular LastPass password > > > > management > > > > service. > > > > > > > > Comments/OK? > > > > > > > > [0] https://github.com/lastpass/lastpass-cli > > > > > > > > -- > > > > Björn Ketelaars > > > > GPG key: 0x4F0E5F21 > > > > > > [replying to old mail] > > > > > > 0.5.0 doesn't work for me, updating this to 0.6.0 does though. Ports-wise, > > > it overrides CFLAGS with -O3 -march=native -fomit-frame-pointer which > > > isn't suitable for ports builds. Also needs a WANTLIB regen (+pthread). > > > > > > > Thank you for your feedback! Included an update to 0.6.0.. > > > > OK? > > Hmm, that still doesn't pass in CFLAGS from the environment (which > is wanted so you can do things like 'make package DEBUG=-g'). The easiest > way is probably to use ALL_TARGET to set the desired make targets, which > passes in CFLAGS etc automatically. Then that part of the patch is > also not needed. How does this look to you? Bonus 64-bit time_t patch > in there too. > > OK with you Björn? Any OKs to import? > OK for your changes...
Re: [patch] Test target PATH
On Thu, Aug 27, 2015 at 11:50:19AM +0200, Juan Francisco Cantero Hurtado wrote: > Can you try this in your port? > > TEST_ENV="PATH=${PATH}:YOURPATH" or TEST_ENV="PATH=YOURPATH:${PATH}" I don't think you're following the problem. It's not that anything needs to be added to PATH; it's that the first element needs to be removed. The PATH that ports uses is set by PORTPATH to be ${WRKDIR}/bin:/usr/bin:/bin:/usr/sbin:/sbin:${DEPBASE}/bin:${LOCALBASE}/bin:${X11BASE}/bin With make test, ${WRKDIR}/bin only has install in it, so chgrp runs /bin/chgrp. When make fake is run, a symlink from ${WRKDIR}/bin/chgrp to /bin/echo is added. This resuls in make fake test using the chgrp in ${WRKDIR}/bin and the test failing since that chgrp doesn't change the group. I am proposing to remove ${WRKDIR}/bin from PATH for tests so that all commands work as expected. -Matthew Martin > On Thu, Aug 27, 2015 at 12:37:22AM -0500, Matthew Martin wrote: > > The zsh port fails an additional test when invoked with make clean fake > > test instead of make clean test. This is because in its tests it > > > > touch zerolength > > chgrp $EGID zerolength > > > > and then tests that zerolength has a group of EGID. Since make fake puts > > a symlink in ${WRKDIR}/bin from chgrp to /bin/echo and ${WRKDIR}/bin is > > first in the environment's PATH, the chgrp does nothing, and test > > C02cond fails. With the below patch it completes successfully. > > > > This patch does slightly change behavior as PATH can no longer be set in > > TEST_ENV (which no ports seem to do anyway) and install is no longer the > > perl wrapper in ${WRKDIR}/bin; however, during tests we probably > > shouldn't be messing with install either. > > > > - Matthew Martin > > > > > > Index: infrastructure/mk/bsd.port.mk > > === > > RCS file: /cvs/ports/infrastructure/mk/bsd.port.mk,v > > retrieving revision 1.1298 > > diff -u -p -r1.1298 bsd.port.mk > > --- infrastructure/mk/bsd.port.mk 19 Jul 2015 17:31:44 - 1.1298 > > +++ infrastructure/mk/bsd.port.mk 27 Aug 2015 03:53:55 - > > @@ -809,7 +809,7 @@ FAKE_TARGET ?= ${INSTALL_TARGET} > > > > TEST_TARGET ?= test > > TEST_FLAGS ?= > > -TEST_ENV ?= > > +TEST_ENV += > > PATH=/usr/bin:/bin:/usr/sbin:/sbin:${DEPBASE}/bin:${LOCALBASE}/bin:${X11BASE}/bin > > ALL_TEST_FLAGS = ${MAKE_FLAGS} ${TEST_FLAGS} > > ALL_TEST_ENV = ${MAKE_ENV} ${TEST_ENV} > > TEST_LOGFILE ?= ${WRKDIR}/test.log > > > > -- > Juan Francisco Cantero Hurtado http://juanfra.info >
Re: softhsm: add portroach limit
On Thu, Aug 27, 2015 at 05:35:54PM +0100, Stuart Henderson wrote: > On 2015/08/27 18:17, Antoine Jacoutot wrote: > > On Thu, Aug 27, 2015 at 06:06:44PM +0200, Patrik Lundin wrote: > > > Hello, > > > > > > portroach currently thinks security/softhsm is outdated. The reason for > > > this is that while the version in ports is 1.3.7, there is a development > > > release using a 2.0.0 version number. > > > > > > The diff below should limit the version check to the 1.x.x versions. > > > > Well the problem here is that once uptream release a 2.X version as stable, > > then portroach will never know. > > I'd rather have false positive in this case, rather than missing updates. > > I agree - I imagine that you could skipv:2.0.0 for now though. > I am pretty sure the 2.0.0 version would be a different port (security/softhsm2). Looking at the example at https://wiki.opendnssec.org/display/SoftHSMDOCS/SoftHSM+Documentation+v2.0 it can be seen that a different command line utility is used (specifically mentioning "2" in the name): === softhsm2-util --init-token --slot 0 --label "My token 1" === -- Patrik Lundin
Re: desktop environment mime setup
On 2015/08/27 17:32, Stuart Henderson wrote: > I need to associate audio/gsm (.gsm) files with a player. I've worked > out half of it (gvfs-mime --set audio/gsm vlc.desktop) but not sure how > to register .gsm files as having type audio/gsm. Can anyone throw me a > clue please? TIA. Oh, please ignore. There's a reg for audio/x-gsm (even though rfc3555 defines audio/gsm) in freedesktop.org.xml that I can use.
Re: NEW: security/lastpass-cli
On 2015/08/27 17:26, Björn Ketelaars wrote: > On Thu 27/08/2015 12:35, Stuart Henderson wrote: > > On 2015/03/17 07:11, Björn Ketelaars wrote: > > > Please find enclosed a port for security/lastpass-cli [0]. lastpass-cli > > > is an > > > open-source command-line interface to the popular LastPass password > > > management > > > service. > > > > > > Comments/OK? > > > > > > [0] https://github.com/lastpass/lastpass-cli > > > > > > -- > > > Björn Ketelaars > > > GPG key: 0x4F0E5F21 > > > > [replying to old mail] > > > > 0.5.0 doesn't work for me, updating this to 0.6.0 does though. Ports-wise, > > it overrides CFLAGS with -O3 -march=native -fomit-frame-pointer which > > isn't suitable for ports builds. Also needs a WANTLIB regen (+pthread). > > > > Thank you for your feedback! Included an update to 0.6.0.. > > OK? Hmm, that still doesn't pass in CFLAGS from the environment (which is wanted so you can do things like 'make package DEBUG=-g'). The easiest way is probably to use ALL_TARGET to set the desired make targets, which passes in CFLAGS etc automatically. Then that part of the patch is also not needed. How does this look to you? Bonus 64-bit time_t patch in there too. OK with you Björn? Any OKs to import? lastpass-cli.2.tgz Description: application/tar-gz
softhsm: add portroach limit
Hello, portroach currently thinks security/softhsm is outdated. The reason for this is that while the version in ports is 1.3.7, there is a development release using a 2.0.0 version number. The diff below should limit the version check to the 1.x.x versions. -- Patrik Lundin Index: Makefile === RCS file: /cvs/ports/security/softhsm/Makefile,v retrieving revision 1.1.1.1 diff -u -p -u -r1.1.1.1 Makefile --- Makefile23 Jun 2015 11:26:05 - 1.1.1.1 +++ Makefile27 Aug 2015 16:00:25 - @@ -1,5 +1,7 @@ # $OpenBSD: Makefile,v 1.1.1.1 2015/06/23 11:26:05 jca Exp $ +PORTROACH= limit:^1\. + SHARED_ONLY= Yes COMMENT= software PKCS\#11 cryptographic token
Re: softhsm: add portroach limit
On 2015/08/27 18:17, Antoine Jacoutot wrote: > On Thu, Aug 27, 2015 at 06:06:44PM +0200, Patrik Lundin wrote: > > Hello, > > > > portroach currently thinks security/softhsm is outdated. The reason for > > this is that while the version in ports is 1.3.7, there is a development > > release using a 2.0.0 version number. > > > > The diff below should limit the version check to the 1.x.x versions. > > Well the problem here is that once uptream release a 2.X version as stable, > then portroach will never know. > I'd rather have false positive in this case, rather than missing updates. I agree - I imagine that you could skipv:2.0.0 for now though.
Re: softhsm: add portroach limit
On Thu, Aug 27, 2015 at 06:06:44PM +0200, Patrik Lundin wrote: > Hello, > > portroach currently thinks security/softhsm is outdated. The reason for > this is that while the version in ports is 1.3.7, there is a development > release using a 2.0.0 version number. > > The diff below should limit the version check to the 1.x.x versions. Well the problem here is that once uptream release a 2.X version as stable, then portroach will never know. I'd rather have false positive in this case, rather than missing updates. > -- > Patrik Lundin > > Index: Makefile > === > RCS file: /cvs/ports/security/softhsm/Makefile,v > retrieving revision 1.1.1.1 > diff -u -p -u -r1.1.1.1 Makefile > --- Makefile 23 Jun 2015 11:26:05 - 1.1.1.1 > +++ Makefile 27 Aug 2015 16:00:25 - > @@ -1,5 +1,7 @@ > # $OpenBSD: Makefile,v 1.1.1.1 2015/06/23 11:26:05 jca Exp $ > > +PORTROACH= limit:^1\. > + > SHARED_ONLY= Yes > > COMMENT= software PKCS\#11 cryptographic token > -- Antoine
desktop environment mime setup
I need to associate audio/gsm (.gsm) files with a player. I've worked out half of it (gvfs-mime --set audio/gsm vlc.desktop) but not sure how to register .gsm files as having type audio/gsm. Can anyone throw me a clue please? TIA.
Re: update chicken-4.10.0
Juan Francisco Cantero Hurtado writes: > On Thu, Aug 27, 2015 at 12:40:43PM +0300, Timo Myyrä wrote: > >> Hmm, >> >> I was remaking my changes but I can't seem to get the core package to work. >> I've made changes to the bootstrap package, built and installed it. >> When I tried to extract the chicken core I got following: >> >> $ cd /usr/ports/mystuff/lang/chicken/core >> $ make depends >> ===> chicken-4.10.0 depends on: chicken-bootstrap->=4.10.0 - default >> chicken-bootstrap-4.9.0.1 does not match >> *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2112 >> '/usr/ports/pobj/chicken-4.10.0/.dep-STEM-ge-4.10.0-lang-chicken-bootstrap') >> *** Error 1 in /usr/ports/mystuff/lang/chicken/core >> (/usr/ports/infrastructure/mk/bsd.port.mk:2488 'depends') >> $ pkg_info | grep chicken >> chicken-bootstrap-4.10.0 practical and portable Scheme system >> $ >> >> I have the new bootstrap installed but 'make depends' doesn't seem to find >> it. >> >> I then tested it a bit by changing the core/Makefile BUILD_DEPENDS like >> follows: >> -BUILD_DEPENDS= lang/chicken/bootstrap>=${V} >> +BUILD_DEPENDS= lang/chicken/bootstrap>=4.9.0 >> >> The build proceeded a bit further: >> $ make depends >> ===> chicken-4.10.0 depends on: chicken-bootstrap->=4.9.0 -> >> chicken-bootstrap-4.10.0 >> ===> chicken-4.10.0 depends on: gmake-* -> gmake-4.1p0 >> ===> Verifying specs: c m pthread >> ===> found c.81.0 m.9.0 pthread.19.0 >> ===> Verifying specs: c m pthread >> ===> found c.81.0 m.9.0 pthread.19.0 >> ===> chicken-4.10.0 depends on: chicken-* - not found >> ===> Verifying install for chicken-* in lang/chicken/core >> ===> Checking files for chicken-4.9.0.1p1 >> `/usr/ports/distfiles/chicken-4.9.0.1.tar.gz' is up to date. >> >> (SHA256) chicken-4.9.0.1.tar.gz: OK >> ===> chicken-4.9.0.1p1 depends on: chicken-bootstrap->=4.9.0.1 -> >> chicken-bootstrap-4.10.0 >> ... >> >> The odd thing here is that now chicken-4.10.0 seems to depend on chicken-* >> and >> tries to build the old chicken version although it should only require the >> bootstrap to build. I'm a bit lost why the ports behaves like this. > > chicken-core uses chicken-bootstrap from lang/chicken instead of > mystuff/lang/chicken. The ports framework will pick always the > dependencies from ${PORTSDIR} which is usually /usr/ports. > > Move your patches to lang/chicken and the port will work > fine. Yesterday, I tested the update on sparc64 and it passed the > tests. Ah, thanks for pointing that out. Moving away from mystuff fixed it. Seems to me that I shouldn't make any updates to ports in there unless working on some new stuff. So, here's the updated chicken diff. Timo Index: Makefile.inc === RCS file: /cvs/ports/lang/chicken/Makefile.inc,v retrieving revision 1.4 diff -u -u -p -r1.4 Makefile.inc --- Makefile.inc4 Jul 2014 18:26:49 - 1.4 +++ Makefile.inc27 Aug 2015 16:27:03 - @@ -2,7 +2,7 @@ COMMENT= practical and portable Scheme system -V= 4.9.0.1 +V= 4.10.0 DISTNAME= chicken-${V} MAINTAINER=Timo Myyra @@ -30,6 +30,6 @@ FAKE_FLAGS= PREFIX=${TRUEPREFIX} # Based on param.h:MACHINE_STACK_GROWS_UP pre-build: .if ${MACHINE_ARCH} == "hppa" || ${MACHINE_ARCH} == "hppa64" - perl -pi -e 's/C_STACK_GROWS_DOWNWARD 1/C_STACK_GROWS_DOWNWARD 0/g' "${WRKSRC}/Makefile.bsd" + sed -i 's/C_STACK_GROWS_DOWNWARD 1/C_STACK_GROWS_DOWNWARD 0/g' "${WRKSRC}/Makefile.bsd" .endif Index: bootstrap/Makefile === RCS file: /cvs/ports/lang/chicken/bootstrap/Makefile,v retrieving revision 1.1 diff -u -u -p -r1.1 Makefile --- bootstrap/Makefile 9 Mar 2014 20:17:01 - 1.1 +++ bootstrap/Makefile 27 Aug 2015 16:27:03 - @@ -9,5 +9,4 @@ ALL_TARGET= boot-chicken do-install: ${INSTALL_PROGRAM} "${WRKSRC}/chicken-boot" "${PREFIX}/libexec/chicken-bootstrap" - .include Index: bootstrap/distinfo === RCS file: /cvs/ports/lang/chicken/bootstrap/distinfo,v retrieving revision 1.2 diff -u -u -p -r1.2 distinfo --- bootstrap/distinfo 4 Jul 2014 18:26:49 - 1.2 +++ bootstrap/distinfo 27 Aug 2015 16:27:03 - @@ -1,2 +1,2 @@ -SHA256 (chicken-4.9.0.1.tar.gz) = BN98Q5w2/BZEa9+hhuenAlj5EdLYJrUhao5rHLKqKBU= -SIZE (chicken-4.9.0.1.tar.gz) = 4023371 +SHA256 (chicken-4.10.0.tar.gz) = Dgf1q80RlhmGlQ2+qlpA20Ffihtl2v+cMA6bBbM0iZs= +SIZE (chicken-4.10.0.tar.gz) = 4020442 Index: bootstrap/patches/patch-Makefile_bsd === RCS file: bootstrap/patches/patch-Makefile_bsd diff -N bootstrap/patches/patch-Makefile_bsd --- bootstrap/patches/patch-Makefile_bsd4 Jul 2014 18:26:49 - 1.2 +++ /dev/null 1 Jan 1970 00:00:00 - @@ -1,15 +0,0 @@ -$OpenBSD: patch-Makefile_bsd,v 1.2 2014/07/04 18:26:49
Writing update for tint port - /etc/xdg and other questions
With much guidance, I'm updating x11/tint to stable 0.12.2. The exercise has generated 3 questions. The first question is the use of /etc/xdg. Referencing a Debian testing installation, default configuration files are starting to show up in /etc/xdg. Debian Examples: /etc/xdg/tint2/tint2rc and /etc/xdg/openbox/rc.xml. My understanding is that it also provides autostarting by placing applications *.desktop file in /etc/xdg/autostart/. The current tint port maintainer previously used post-install scripts to create the /usr/local/share/examples/tint2/ directory. The next line, @mv , copies rather than moves code from /etc/xdg to the newly created ${PREFIX}/share/examples/tint2/ directory. The new tint2 fake installs etc/xdg/tint2/tint2rc but gives the following error with "make plist" > make-plist: Bogus element outside of every prefix: /etc/xdg > make-plist: Bogus element outside of every prefix: /etc/xdg/tint2 > make-plist: Bogus element outside of every prefix: /etc/xdg/tint2 The default configuration needs to be installed and my sense is that this issue will be occurring with other ports in the future. Any suggestions as to what to do presently appreciated. The other questions are easier. Upstream now refers to their code base as tint2 and it is packaged as such in mainstream linux distributions. If the port update is acceptable, would it be appropriate to remove x11/tint and provide a newly named x11/tint2 port? Last question. Tint2 comes with a python based configuation wizard. It runs acceptably but writes unused, configuration examples in ~/.config/tint2/. These examples will already be in /usr/local/share/examples/tint2/. My bias is that OpenBSD is about correct, clean code and this is just plain messy. Is this enough reason to pass a port configuration option not to build the wizard? Thanks -- J. Scott Heppler
Re: NEW: security/lastpass-cli
On Thu 27/08/2015 12:35, Stuart Henderson wrote: > On 2015/03/17 07:11, Björn Ketelaars wrote: > > Please find enclosed a port for security/lastpass-cli [0]. lastpass-cli is > > an > > open-source command-line interface to the popular LastPass password > > management > > service. > > > > Comments/OK? > > > > [0] https://github.com/lastpass/lastpass-cli > > > > -- > > Björn Ketelaars > > GPG key: 0x4F0E5F21 > > [replying to old mail] > > 0.5.0 doesn't work for me, updating this to 0.6.0 does though. Ports-wise, > it overrides CFLAGS with -O3 -march=native -fomit-frame-pointer which > isn't suitable for ports builds. Also needs a WANTLIB regen (+pthread). > Thank you for your feedback! Included an update to 0.6.0.. OK? lastpass-cli.tgz Description: application/tar-gz
Re: [UPDATE] audio/picard to 1.3.2
On 08/27/15 13:00, Fred wrote: Hi, Attached is an update to picard to release version 1.3.2. So far with my testing on amd64 it appears to working well. Cheers Fred Sorry the diff was missing the following Makefile diff. Thanks to Tim for spotting this. I will fix it up and resubmit the whole thing shortly. Thanks Fred Index: Makefile === RCS file: /cvs/ports/audio/picard/Makefile,v retrieving revision 1.18 diff -u -p -u -p -r1.18 Makefile --- Makefile30 Jul 2015 15:09:27 - 1.18 +++ Makefile27 Aug 2015 15:10:14 - @@ -2,9 +2,8 @@ COMMENT = automatic audio files tagger using the MusicBrainz database -MODPY_EGG_VERSION =0.16 +MODPY_EGG_VERSION =1.3.2 DISTNAME = picard-${MODPY_EGG_VERSION} -REVISION = 1 CATEGORIES = audio @@ -18,8 +17,9 @@ PERMIT_PACKAGE_CDROM =Yes WANTLIB += ${MODPY_WANTLIB} WANTLIB += avcodec avformat m ofa pthread stdc++ -MASTER_SITES = http://ftp.musicbrainz.org/pub/musicbrainz/picard/ \ - ftp://ftp.musicbrainz.org/pub/musicbrainz/picard/ +GH_ACCOUNT = musicbrainz +GH_PROJECT = picard +GH_TAGNAME = ${MODPY_EGG_VERSION} MODULES = lang/python @@ -39,11 +39,14 @@ RUN_DEPENDS = x11/py-qt4>=4.1 \ TEST_DEPENDS = ${BASE_PKGPATH} -PLUGINS_DIRS = lastfm \ - lastfmplus \ - replaygain -PLUGINS = classicdiscnumber.py \ - coverart.py \ +PLUGINS_DIRS = lastfm \ + lastfmplus \ + replaygain \ + viewvariables + +PLUGINS = addrelease.py \ + albumartist_website.py \ + classicdiscnumber.py \ cuesheet.py \ discnumber.py \ featartist.py \ @@ -54,14 +57,17 @@ PLUGINS = classicdiscnumber.py \ lastfmplus/__init__.py \ lastfmplus/ui_options_lastfm.py \ no_release.py \ - open_in_gui.py \ release_type.py \ replaygain/__init__.py \ replaygain/options_replaygain.ui \ replaygain/ui_options_replaygain.py \ - swapprefix.py \ + sort_multivalue_tags.py \ + standardise_performers.py \ titlecase.py \ - tracks2clipboard.py + tracks2clipboard.py \ + viewvariables/__init__.py \ + viewvariables/ui_variables_dialog.py \ + viewvariables/variables_dialog.ui pre-configure: ${SUBST_CMD} ${WRKSRC}/setup.py
[UPDATE] audio/picard to 1.3.2
Hi, Attached is an update to picard to release version 1.3.2. So far with my testing on amd64 it appears to working well. Cheers Fred diff -uNr -x CVS /usr/ports/audio/picard/distinfo picard/distinfo --- /usr/ports/audio/picard/distinfo Sun Jan 18 03:12:47 2015 +++ picard/distinfo Thu Aug 27 03:14:50 2015 @@ -1,2 +1,2 @@ -SHA256 (picard-0.16.tar.gz) = ViMhb+3fYAAYfOrTdGs66+X2A4N+MTx9/Jl4NXb1Sxg= -SIZE (picard-0.16.tar.gz) = 1828595 +SHA256 (picard-1.3.2.tar.gz) = y2owR+dW56UNOIv/F5SvYnTnvZIA2V0nK7CPZVi9mWs= +SIZE (picard-1.3.2.tar.gz) = 2047602 diff -uNr -x CVS /usr/ports/audio/picard/patches/patch-picard_musicdns_avcodec_c picard/patches/patch-picard_musicdns_avcodec_c --- /usr/ports/audio/picard/patches/patch-picard_musicdns_avcodec_c Mon May 5 08:55:50 2014 +++ picard/patches/patch-picard_musicdns_avcodec_c Thu Jan 1 01:00:00 1970 @@ -1,38 +0,0 @@ -$OpenBSD: patch-picard_musicdns_avcodec_c,v 1.1 2014/05/05 07:55:50 brad Exp $ - -Deal with newer FFmpeg API. - picard/musicdns/avcodec.c.orig Sun Oct 23 13:15:29 2011 -+++ picard/musicdns/avcodec.c Thu May 1 04:58:17 2014 -@@ -39,6 +39,10 @@ - #define AVMEDIA_TYPE_AUDIO CODEC_TYPE_AUDIO - #endif - -+#ifndef AVCODEC_MAX_AUDIO_FRAME_SIZE -+#define AVCODEC_MAX_AUDIO_FRAME_SIZE 192000 // 1 second of 48khz 32bit audio -+#endif -+ - static PyObject * - init(PyObject *self, PyObject *args) - { -@@ -55,7 +59,7 @@ done(PyObject *self, PyObject *args) - static PyObject * - decode(PyObject *self, PyObject *args) - { --AVFormatContext *format_context; -+AVFormatContext *format_context = NULL; - AVCodecContext *codec_context; - AVCodec *codec; - PyObject *filename; -@@ -69,7 +73,11 @@ decode(PyObject *self, PyObject *args) - return NULL; - - Py_UNBLOCK_THREADS -+#if LIBAVFORMAT_VERSION_INT < AV_VERSION_INT(53, 2, 0) - if (av_open_input_file(&format_context, PyString_AS_STRING(filename), NULL, 0, NULL) != 0) { -+#else -+if (avformat_open_input(&format_context, PyString_AS_STRING(filename), NULL, NULL) != 0) { -+#endif - Py_BLOCK_THREADS - PyErr_SetString(PyExc_Exception, "Couldn't open the file."); - return NULL; diff -uNr -x CVS /usr/ports/audio/picard/patches/patch-setup_py picard/patches/patch-setup_py --- /usr/ports/audio/picard/patches/patch-setup_py Fri May 4 07:24:04 2012 +++ picard/patches/patch-setup_py Thu Jan 1 01:00:00 1970 @@ -1,59 +0,0 @@ -$OpenBSD: patch-setup_py,v 1.3 2012/05/04 06:24:04 ajacoutot Exp $ setup.py.orig Sun Oct 23 13:15:29 2011 -+++ setup.py Sat Apr 28 18:29:55 2012 -@@ -62,12 +62,12 @@ from distutils.dist import Distribution - defaults = { - 'build': { - 'with-directshow': 'False', --'with-avcodec': 'False', --'with-libofa': 'False', -+'with-avcodec': 'True', -+'with-libofa': 'True', - }, --'avcodec': {'cflags': '', 'libs': ''}, -+'avcodec': {'cflags': '-I${LOCALBASE}/include', 'libs': '-L${LOCALBASE}/lib -lavcodec -lavformat'}, - 'directshow': {'cflags': '', 'libs': ''}, --'libofa': {'cflags': '', 'libs': ''}, -+'libofa': {'cflags': '-I${LOCALBASE}/include', 'libs': '-L${LOCALBASE}/lib -lofa'}, - } - cfg = RawConfigParser() - for section, values in defaults.items(): -@@ -78,7 +78,7 @@ cfg.read(['build.cfg']) - - - ext_modules = [ --Extension('picard.util.astrcmp', sources=['picard/util/astrcmp.cpp']), -+Extension('picard.util.astrcmp', sources=['picard/util/astrcmp.cpp'], extra_link_args=['-fPIC']), - ] - - if cfg.getboolean('build', 'with-libofa'): -@@ -372,8 +372,8 @@ class picard_config(config): - if have_pkgconfig: - if self.pkgconfig_check_module('avcodec', 'libavcodec libavformat'): - include_dirs = cflags_to_include_dirs(cfg.get('avcodec', 'cflags')) --if self.try_compile('#include ', include_dirs=include_dirs): --cfg.set('avcodec', 'cflags', cfg.get('avcodec', 'cflags') + ' -DUSE_OLD_FFMPEG_LOCATIONS') -+if self.try_compile('#include ', include_dirs=include_dirs): -+cfg.set('avcodec', 'cflags', cfg.get('avcodec', 'cflags')) - else: - self.check_lib('avcodec', 'av_open_input_file', ['avcodec.h', 'avformat.h'], [['avcodec', 'avformat'], ['avcodec-51', 'avformat-51']]) - -@@ -550,12 +550,12 @@ if do_py2app: - - # FIXME: this should check for the actual command ('install' vs. 'bdist_nsis', 'py2app', ...), not installed libraries - if py2exe is None and do_py2app is False: --args['data_files'].append(('share/icons/hicolor/16x16/apps', ['resources/images/16x16/picard.png'])) --args['data_files'].append(('share/icons/hicolor/24x24/apps', ['resources/images/24x24/picard.png'])) --args['data_files'].append(('share/icons/hicolor/32x32/apps', ['resources/images/32x32/picard.png'])) --args['data_files'].append(('share/icons/hicolor/48x48/apps', ['resources/images/48x48/picard.png'])) --args['data_files'].append(('share/icons/hicolor/12
Re: UPDATE: mail/perdition
Giovanni Bechis writes: > Update to 2.1 and take maintainership. > Comments ? Ok ? Looks fine ports-wise, ok jca@ > Cheers >Giovanni > -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
Re: NEW: security/lastpass-cli
On 2015/03/17 07:11, Björn Ketelaars wrote: > Please find enclosed a port for security/lastpass-cli [0]. lastpass-cli is an > open-source command-line interface to the popular LastPass password management > service. > > Comments/OK? > > [0] https://github.com/lastpass/lastpass-cli > > -- > Björn Ketelaars > GPG key: 0x4F0E5F21 [replying to old mail] 0.5.0 doesn't work for me, updating this to 0.6.0 does though. Ports-wise, it overrides CFLAGS with -O3 -march=native -fomit-frame-pointer which isn't suitable for ports builds. Also needs a WANTLIB regen (+pthread).
UPDATE: mail/perdition
Update to 2.1 and take maintainership. Comments ? Ok ? Cheers Giovanni Index: Makefile === RCS file: /var/cvs/ports/mail/perdition/Makefile,v retrieving revision 1.31 diff -u -p -r1.31 Makefile --- Makefile 21 Aug 2015 22:51:10 - 1.31 +++ Makefile 27 Aug 2015 10:26:17 - @@ -2,21 +2,21 @@ COMMENT-main= fully featured POP3 and IMAP4 proxy server COMMENT-ldap= LDAP backend for perdition mail proxy -COMMENT-mysql= MySQL backend for perdition mail proxy +COMMENT-mysql= MySQL backend for perdition mail proxy COMMENT-pgsql= PostgreSQL backend for perdition mail proxy -DISTVER= 1.19-rc5 -REVISION= 6 -REVISION-mysql= 8 +DISTVER= 2.1 DISTNAME= perdition-${DISTVER} PKGNAME-main= perdition-${DISTVER:S/-//} PKGNAME-ldap= perdition-ldap-${DISTVER:S/-//} -PKGNAME-mysql= perdition-mysql-${DISTVER:S/-//} -PKGNAME-pgsql= perdition-pgsql-${DISTVER:S/-//} +PKGNAME-mysql= perdition-mysql-${DISTVER:S/-//} +PKGNAME-pgsql= perdition-pgsql-${DISTVER:S/-//} CATEGORIES= mail -SHARED_LIBS += jain 0.0 # .0.0 +HOMEPAGE= http://horms.net/projects/perdition/ +MAINTAINER= Giovanni Bechis + SHARED_LIBS += perditiondb_bdb 0.0 # .0.0 SHARED_LIBS += perditiondb_nis 0.0 # .0.0 SHARED_LIBS += perditiondb_posix_regex 0.0 # .0.0 @@ -26,8 +26,6 @@ SHARED_LIBS += perditiondb_ldap SHARED_LIBS += perditiondb_mysql0.0 # .0.0 SHARED_LIBS += perditiondb_postgresql 0.0 # .0.0 -HOMEPAGE= http://www.vergenet.net/linux/perdition/ - # GPLv2+ (included COPYING is v3, but all files with a # copyright claim specify version 2 or above). PERMIT_PACKAGE_CDROM= Yes @@ -60,18 +58,20 @@ MULTI_PACKAGES= -main -mysql -pgsql -ld .include -cWANTLIB= lib/db4/db>=4 vanessa_logger>=1.0 vanessa_socket>=1.0 \ - vanessa_adt>=1.0 -LIB_DEPENDS+= databases/db/v4 \ +cWANTLIB= lib/db4/db>=4 vanessa_logger>=1.0 +cLIB_DEPENDS+= databases/db/v4 \ + vanessa_logger->=0.0.10:devel/vanessa/logger + +WANTLIB-main= ${cWANTLIB} ${MODGETTEXT_WANTLIB} \ + c crypto ssl util popt idn \ + vanessa_adt>=1.0 vanessa_socket>=1.0 +LIB_DEPENDS-main= ${cLIB_DEPENDS} ${LIB_DEPENDS} \ + ${MODGETTEXT_LIB_DEPENDS} \ + devel/popt devel/libidn \ vanessa_adt->=0.0.9:devel/vanessa/adt \ - vanessa_logger->=0.0.10:devel/vanessa/logger \ vanessa_socket->=0.0.12:devel/vanessa/socket -WANTLIB-main= ${MODGETTEXT_WANTLIB} ${cWANTLIB} \ - c crypto ssl util popt idn -LIB_DEPENDS-main= ${LIB_DEPENDS} devel/popt devel/libidn - -LIB_DEPENDS-ldap= ${LIB_DEPENDS} security/cyrus-sasl2 \ +LIB_DEPENDS-ldap= ${cLIB_DEPENDS} security/cyrus-sasl2 \ openldap-client->=2,<3:databases/openldap WANTLIB-ldap= crypto lber-2.4 ldap-2.4 sasl2 ssl util ${cWANTLIB} @@ -83,9 +83,11 @@ CONFIGURE_ARGS+= --with-ldap-libraries=$ --disable-ldap-doc .endif -LIB_DEPENDS-mysql= ${LIB_DEPENDS} databases/mariadb -WANTLIB-mysql= crypto ssl util m z lib/mysql/mysqlclient \ - pthread stdc++ ${cWANTLIB} +LIB_DEPENDS-mysql= ${cLIB_DEPENDS} databases/mariadb \ + vanessa_adt->=0.0.9:devel/vanessa/adt +WANTLIB-mysql= crypto pthread ssl stdc++ util m z \ + lib/mysql/mysqlclient ${cWANTLIB} \ + vanessa_adt>=1.0 .if !${BUILD_PACKAGES:M-mysql} CONFIGURE_ARGS+= --disable-mysql @@ -94,8 +96,9 @@ CONFIGURE_ARGS+= --with-mysql-libraries= --with-mysql-includes=${LOCALBASE}/include/mysql .endif -LIB_DEPENDS-pgsql= ${LIB_DEPENDS} databases/postgresql -WANTLIB-pgsql= crypto pq ssl util ${cWANTLIB} +LIB_DEPENDS-pgsql= ${cLIB_DEPENDS} databases/postgresql \ + vanessa_adt->=0.0.9:devel/vanessa/adt +WANTLIB-pgsql= crypto pq ssl util ${cWANTLIB} vanessa_adt>=1.0 .if !${BUILD_PACKAGES:M-pgsql} CONFIGURE_ARGS+= --disable-pg Index: distinfo === RCS file: /var/cvs/ports/mail/perdition/distinfo,v retrieving revision 1.4 diff -u -p -r1.4 distinfo --- distinfo 15 Mar 2013 15:17:36 - 1.4 +++ distinfo 17 Jul 2015 10:22:21 - @@ -1,2 +1,2 @@ -SHA256 (perdition-1.19-rc5.tar.gz) = dgctaV0Y5VofNOPNOM54HS8Tj4WdCLeu37y1GlxIkrU= -SIZE (perdition-1.19-rc5.tar.gz) = 642566 +SHA256 (perdition-2.1.tar.gz) = ZF+7O7HQAfWuyV0oSLI+JFrqzHkRnws5K8V82JbXn+g= +SIZE (perdition-2.1.tar.gz) = 701062 Index: patches/patch-perdition_db_ldap_perditiondb_ldap_makedb === RCS file: patches/patch-perdition_db_ldap_perditiondb_ldap_makedb diff -N patches/patch-perdition_db_ldap_perditiondb_ldap_makedb --- /dev/null 1 Jan 1970 00:00:00 - +++ patches/patch-perdition_db_ldap_perditiondb_ldap_makedb 17 Jul 2015 16:26:47 - @@ -0,0 +1,9 @@ +$OpenBSD$ +--- perdition/db/ldap/perditiondb_ldap_makedb.orig Fri Jul 17 18:26:20 2015 perdition/db/ldap/perditiondb_ldap_makedb Fri Jul 17 18:26:30 2015 +@@ -1,4 +1,4 @@ +-#!/bin/bash ++#!/bin/sh + ###
Re: update chicken-4.10.0
On Thu, Aug 27, 2015 at 12:40:43PM +0300, Timo Myyrä wrote: > Hmm, > > I was remaking my changes but I can't seem to get the core package to work. > I've made changes to the bootstrap package, built and installed it. > When I tried to extract the chicken core I got following: > > $ cd /usr/ports/mystuff/lang/chicken/core > > > $ make depends > ===> chicken-4.10.0 depends on: chicken-bootstrap->=4.10.0 - default > chicken-bootstrap-4.9.0.1 does not match > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2112 > '/usr/ports/pobj/chicken-4.10.0/.dep-STEM-ge-4.10.0-lang-chicken-bootstrap') > *** Error 1 in /usr/ports/mystuff/lang/chicken/core > (/usr/ports/infrastructure/mk/bsd.port.mk:2488 'depends') > $ pkg_info | grep chicken > chicken-bootstrap-4.10.0 practical and portable Scheme system > $ > > I have the new bootstrap installed but 'make depends' doesn't seem to find it. > > I then tested it a bit by changing the core/Makefile BUILD_DEPENDS like > follows: > -BUILD_DEPENDS= lang/chicken/bootstrap>=${V} > +BUILD_DEPENDS= lang/chicken/bootstrap>=4.9.0 > > The build proceeded a bit further: > $ make depends > ===> chicken-4.10.0 depends on: chicken-bootstrap->=4.9.0 -> > chicken-bootstrap-4.10.0 > ===> chicken-4.10.0 depends on: gmake-* -> gmake-4.1p0 > ===> Verifying specs: c m pthread > ===> found c.81.0 m.9.0 pthread.19.0 > ===> Verifying specs: c m pthread > ===> found c.81.0 m.9.0 pthread.19.0 > ===> chicken-4.10.0 depends on: chicken-* - not found > ===> Verifying install for chicken-* in lang/chicken/core > ===> Checking files for chicken-4.9.0.1p1 > `/usr/ports/distfiles/chicken-4.9.0.1.tar.gz' is up to date. > >> (SHA256) chicken-4.9.0.1.tar.gz: OK > ===> chicken-4.9.0.1p1 depends on: chicken-bootstrap->=4.9.0.1 -> > chicken-bootstrap-4.10.0 > ... > > The odd thing here is that now chicken-4.10.0 seems to depend on chicken-* and > tries to build the old chicken version although it should only require the > bootstrap to build. I'm a bit lost why the ports behaves like this. chicken-core uses chicken-bootstrap from lang/chicken instead of mystuff/lang/chicken. The ports framework will pick always the dependencies from ${PORTSDIR} which is usually /usr/ports. Move your patches to lang/chicken and the port will work fine. Yesterday, I tested the update on sparc64 and it passed the tests. -- Juan Francisco Cantero Hurtado http://juanfra.info
Re: update chicken-4.10.0
Hmm, I was remaking my changes but I can't seem to get the core package to work. I've made changes to the bootstrap package, built and installed it. When I tried to extract the chicken core I got following: $ cd /usr/ports/mystuff/lang/chicken/core $ make depends ===> chicken-4.10.0 depends on: chicken-bootstrap->=4.10.0 - default chicken-bootstrap-4.9.0.1 does not match *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2112 '/usr/ports/pobj/chicken-4.10.0/.dep-STEM-ge-4.10.0-lang-chicken-bootstrap') *** Error 1 in /usr/ports/mystuff/lang/chicken/core (/usr/ports/infrastructure/mk/bsd.port.mk:2488 'depends') $ pkg_info | grep chicken chicken-bootstrap-4.10.0 practical and portable Scheme system $ I have the new bootstrap installed but 'make depends' doesn't seem to find it. I then tested it a bit by changing the core/Makefile BUILD_DEPENDS like follows: -BUILD_DEPENDS= lang/chicken/bootstrap>=${V} +BUILD_DEPENDS= lang/chicken/bootstrap>=4.9.0 The build proceeded a bit further: $ make depends ===> chicken-4.10.0 depends on: chicken-bootstrap->=4.9.0 -> chicken-bootstrap-4.10.0 ===> chicken-4.10.0 depends on: gmake-* -> gmake-4.1p0 ===> Verifying specs: c m pthread ===> found c.81.0 m.9.0 pthread.19.0 ===> Verifying specs: c m pthread ===> found c.81.0 m.9.0 pthread.19.0 ===> chicken-4.10.0 depends on: chicken-* - not found ===> Verifying install for chicken-* in lang/chicken/core ===> Checking files for chicken-4.9.0.1p1 `/usr/ports/distfiles/chicken-4.9.0.1.tar.gz' is up to date. >> (SHA256) chicken-4.9.0.1.tar.gz: OK ===> chicken-4.9.0.1p1 depends on: chicken-bootstrap->=4.9.0.1 -> chicken-bootstrap-4.10.0 ... The odd thing here is that now chicken-4.10.0 seems to depend on chicken-* and tries to build the old chicken version although it should only require the bootstrap to build. I'm a bit lost why the ports behaves like this. Timo Juan Francisco Cantero Hurtado writes: > Sorry for the double mail. > > Can you remove the patch bootstrap/patches/patch-Makefile_bsd and to > modify C_COMPILER_OPTIONS within core/patches/patch-Makefile_bsd? > > The GCC version in base supports -fwrapv. > > > On Wed, Aug 26, 2015 at 06:00:33PM +0200, Juan Francisco Cantero Hurtado > wrote: >> Can you update the patch? >> >> Your core/patches/patch-data-structures_scm is outdated. >> >> On Wed, Aug 12, 2015 at 09:54:56PM +0300, Timo Myyrä wrote: >> > Hi, >> > >> > Here's an chicken update waiting for ports unlock. >> > Lightly tested on my thinkpad. >> > >> > Timo >> > >> > Index: Makefile.inc >> > === >> > RCS file: /cvs/ports/lang/chicken/Makefile.inc,v >> > retrieving revision 1.4 >> > diff -u -u -p -r1.4 Makefile.inc >> > --- Makefile.inc 4 Jul 2014 18:26:49 - 1.4 >> > +++ Makefile.inc 7 Aug 2015 19:38:47 - >> > @@ -2,7 +2,7 @@ >> > >> > COMMENT= practical and portable Scheme system >> > >> > -V=4.9.0.1 >> > +V=4.10.0 >> > DISTNAME= chicken-${V} >> > >> > MAINTAINER= Timo Myyra >> > @@ -30,6 +30,6 @@ FAKE_FLAGS= PREFIX=${TRUEPREFIX} >> > # Based on param.h:MACHINE_STACK_GROWS_UP >> > pre-build: >> > .if ${MACHINE_ARCH} == "hppa" || ${MACHINE_ARCH} == "hppa64" >> > - perl -pi -e 's/C_STACK_GROWS_DOWNWARD 1/C_STACK_GROWS_DOWNWARD 0/g' >> > "${WRKSRC}/Makefile.bsd" >> > + sed -i 's/C_STACK_GROWS_DOWNWARD 1/C_STACK_GROWS_DOWNWARD 0/g' >> > "${WRKSRC}/Makefile.bsd" >> > .endif >> > >> > Index: bootstrap/distinfo >> > === >> > RCS file: /cvs/ports/lang/chicken/bootstrap/distinfo,v >> > retrieving revision 1.2 >> > diff -u -u -p -r1.2 distinfo >> > --- bootstrap/distinfo 4 Jul 2014 18:26:49 - 1.2 >> > +++ bootstrap/distinfo 7 Aug 2015 19:38:47 - >> > @@ -1,2 +1,2 @@ >> > -SHA256 (chicken-4.9.0.1.tar.gz) = >> > BN98Q5w2/BZEa9+hhuenAlj5EdLYJrUhao5rHLKqKBU= >> > -SIZE (chicken-4.9.0.1.tar.gz) = 4023371 >> > +SHA256 (chicken-4.10.0.tar.gz) = >> > Dgf1q80RlhmGlQ2+qlpA20Ffihtl2v+cMA6bBbM0iZs= >> > +SIZE (chicken-4.10.0.tar.gz) = 4020442 >> > Index: core/Makefile >> > === >> > RCS file: /cvs/ports/lang/chicken/core/Makefile,v >> > retrieving revision 1.5 >> > diff -u -u -p -r1.5 Makefile >> > --- core/Makefile 16 Jun 2015 14:45:04 - 1.5 >> > +++ core/Makefile 7 Aug 2015 19:38:47 - >> > @@ -2,9 +2,7 @@ >> > >> > SHARED_ONLY= Yes >> > >> > -REVISION= 1 >> > - >> > -SHARED_LIBS= chicken 3.0 >> > +SHARED_LIBS= chicken 4.0 # 7 >> > >> > TEST_TARGET= check >> > TEST_DEPENDS= ${BUILD_PKGPATH} >> > Index: core/distinfo >> > === >> > RCS file: /cvs/ports/lang/chicken/core/distinfo,v >> > re
Re: [patch] Test target PATH
Can you try this in your port? TEST_ENV="PATH=${PATH}:YOURPATH" or TEST_ENV="PATH=YOURPATH:${PATH}" On Thu, Aug 27, 2015 at 12:37:22AM -0500, Matthew Martin wrote: > The zsh port fails an additional test when invoked with make clean fake > test instead of make clean test. This is because in its tests it > > touch zerolength > chgrp $EGID zerolength > > and then tests that zerolength has a group of EGID. Since make fake puts > a symlink in ${WRKDIR}/bin from chgrp to /bin/echo and ${WRKDIR}/bin is > first in the environment's PATH, the chgrp does nothing, and test > C02cond fails. With the below patch it completes successfully. > > This patch does slightly change behavior as PATH can no longer be set in > TEST_ENV (which no ports seem to do anyway) and install is no longer the > perl wrapper in ${WRKDIR}/bin; however, during tests we probably > shouldn't be messing with install either. > > - Matthew Martin > > > Index: infrastructure/mk/bsd.port.mk > === > RCS file: /cvs/ports/infrastructure/mk/bsd.port.mk,v > retrieving revision 1.1298 > diff -u -p -r1.1298 bsd.port.mk > --- infrastructure/mk/bsd.port.mk 19 Jul 2015 17:31:44 - 1.1298 > +++ infrastructure/mk/bsd.port.mk 27 Aug 2015 03:53:55 - > @@ -809,7 +809,7 @@ FAKE_TARGET ?= ${INSTALL_TARGET} > > TEST_TARGET ?= test > TEST_FLAGS ?= > -TEST_ENV ?= > +TEST_ENV += > PATH=/usr/bin:/bin:/usr/sbin:/sbin:${DEPBASE}/bin:${LOCALBASE}/bin:${X11BASE}/bin > ALL_TEST_FLAGS = ${MAKE_FLAGS} ${TEST_FLAGS} > ALL_TEST_ENV = ${MAKE_ENV} ${TEST_ENV} > TEST_LOGFILE ?= ${WRKDIR}/test.log > -- Juan Francisco Cantero Hurtado http://juanfra.info