Re: NEW: net/i2p
On Sat, October 3, 2020 13:00, Solene Rapenne wrote: > On Tue, 15 Sep 2020 13:16:37 - > "Dimitri Karamazov" : > > >> Can't seem to get any response on this, is there nobody that has >> time to verify this? > > Hello, > > > the ports looks fine and work on amd64. > > Although, at the end of the build process there is this message, > could you look at the build script to disable this check producing the > following error? Made changes with some input from sthen@ and solene@ *modified existing patch to disable the test *changed pkgname to java-tanukiwrapper *set MODJAVA_VER to 1.8 for both ports Any comments/OK's? i2p.tar.gz Description: application/gzip diff Description: Binary data tanukiwrapper.tar.gz Description: application/gzip
Re: Add missing dependency for audio/rgain
On Sat, Oct 03, 2020 at 10:23:36PM -0700, Jeremy Evans wrote:
> I think audio/rgain is missing a runtime dependency. Running it after
> install, I get:
> /usr/local/bin/replaygain -h
> Traceback (most recent call last):
> File "/usr/local/bin/replaygain", line 3, in
> from rgain3.script.replaygain import replaygain
> File "/usr/local/lib/python3.8/site-packages/rgain3/script/__init__.py",
> line 29, in
> import rgain3.rgio # noqa isort:skip
> File "/usr/local/lib/python3.8/site-packages/rgain3/rgio.py", line 23, in
>
> import mutagen
> ModuleNotFoundError: No module named 'mutagen'
> Patch below fixes it. OKs?
The new audio/py-mutagen${MODPY_FLAVOR} line should got on the line with
RUN_DEPENDS and bump devel/py-gobject3${MODPY_FLAVOR} to a line on its own
below it. (Sorting of DEPENDS should be alphabetical).
With that changed, ok kmos@
--Kurt
> Thanks,
> Jeremy
>
> Index: Makefile
> ===
> RCS file: /cvs/ports/audio/rgain/Makefile,v
> retrieving revision 1.8
> diff -u -p -r1.8 Makefile
> --- Makefile 3 Oct 2020 15:39:58 - 1.8
> +++ Makefile 4 Oct 2020 05:21:00 -
> @@ -5,7 +5,7 @@ COMMENT= multi-format tools and Python m
> MODPY_EGG_VERSION= 1.0.0
> DISTNAME=rgain3-${MODPY_EGG_VERSION}
> EPOCH= 0
> -REVISION=0
> +REVISION=1
>
> #WRKDIST=${WRKDIR}/fk-rgain-8df1a846a12c
>
> @@ -23,6 +23,7 @@ MODPY_VERSION= ${MODPY_DEFAULT_VERSION_3
> BUILD_DEPENDS= textproc/py-docutils${MODPY_FLAVOR}
>
> RUN_DEPENDS= devel/py-gobject3${MODPY_FLAVOR} \
> + audio/py-mutagen${MODPY_FLAVOR} \
> multimedia/gstreamer1/core \
> multimedia/gstreamer1/plugins-libav \
> multimedia/gstreamer1/plugins-base \
>
Re: Review for cabal.port.mk
I'll have to look more closely, but I have a few concerns.
On Sat, Oct 03, 2020 at 11:03:43PM -0700, Greg Steuck wrote:
> MASTER_SITES0 = https://hackage.haskell.org/package/
That's a minor one.
This sets a precedent of hijacking a MASTER_SITESn for a specific module.
I don't think it's a big concern because cabal is somewhat isolated.
We'll deal with that if we must.
> DIST_SUBDIR ?=hackage
>
> # The .cabal files are explicitly copied over the ones extracted from
> # archives by the normal extraction rules.
> EXTRACT_CASES += *.cabal) ;;
I'm not fond of this solution to the PKGSTEM issue.
I would prefer that you would reconstitute DISTNAME from MODCABAL variables
exactly like stuff is done for github
> MODCABAL_HACKAGE_NAME = ${DISTNAME:C,-[0-9.]*$,,}
> MODCABAL_HACKAGE_VERSION =${DISTNAME:C,.*-([0-9.]*)$,\1,}
and I think those variables might be a bit long, though very descriptive
what's wrong with MODCABAL_STEM and MODCABAL_VERSION ?
> .if !target(do-build)
> do-build:
> @${_MODCABAL_BUILD_TARGET}
> .endif
>
> .if !target(do-install)
> do-install:
> @${_MODCABAL_INSTALL_TARGET}
> .endif
let's think these through. _MODCABAL_{BUILD,INSTALL}_TARGET are internal
so not reusable from outside, strictly speaking.
(this is probably taken from ruby.port.mk or something similar)
If you define do-build/do-install manually, you might want to reference
these.
The reason the redirection exist is to allow non trivial build/install to
happen when several modules are combined.
in general, you may need to make these visible.
aarch64 bulk build report
bulk build on arm64.ports.openbsd.org started on Thu Oct 1 08:22:11 MDT 2020 finished at Sun Oct 4 02:25:25 MDT 2020 lasted 2D18h03m done with kern.version=OpenBSD 6.8 (GENERIC.MP) #826: Tue Sep 29 17:21:59 MDT 2020 built packages:10848 Oct 1:3885 Oct 2:946 Oct 3:3329 Oct 4:2687 critical path missing pkgs: http://build-failures.rhaalovely.net/aarch64/2020-10-01/summary.log build failures: 14 http://build-failures.rhaalovely.net/aarch64/2020-10-01/converters/wv2.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/editors/calligra.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/editors/xwpe.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/emulators/vice.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/games/shockolate.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/lang/pfe.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/sysutils/nomad.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/sysutils/rclone.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/sysutils/telegraf.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/sysutils/terragrunt.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/x11/e17/elementary.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/x11/elementary/dock.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/x11/kde-applications/marble.log http://build-failures.rhaalovely.net/aarch64/2020-10-01/x11/qt5/qtwebkit.log recurrent failures failures/emulators/vice.log failures/games/shockolate.log failures/lang/pfe.log failures/sysutils/nomad.log failures/sysutils/rclone.log failures/sysutils/telegraf.log failures/sysutils/terragrunt.log failures/x11/e17/elementary.log failures/x11/qt5/qtwebkit.log new failures +++ ls-failures Sun Oct 4 02:25:38 2020 +failures/x11/elementary/dock.log +failures/x11/kde-applications/marble.log resolved failures --- ../old/aarch64/last//ls-failuresTue Sep 29 18:02:41 2020 -failures/net/ntp.log -failures/security/vault.log -failures/sysutils/packer.log -failures/x11/qt5/qtwebengine.log
Re: Review for cabal.port.mk
On Sun, Oct 04, 2020 at 09:56:22AM +0200, Marc Espie wrote: > I'll have to look more closely, but I have a few concerns. > > On Sat, Oct 03, 2020 at 11:03:43PM -0700, Greg Steuck wrote: > > MASTER_SITES0 = https://hackage.haskell.org/package/ > That's a minor one. > This sets a precedent of hijacking a MASTER_SITESn for a specific module. > I don't think it's a big concern because cabal is somewhat isolated. > We'll deal with that if we must. just a side note. technically lang/rust and lang/go already do something like that, but in a more flexible way: the number is configurable, and by default is a high number (9) to limit conflict with possible user usage. -- Sebastien Marie
Re: [macppc] Don't build x11/kitty
On 10/4/2020 1:01 AM, Thomas Frohwein wrote: On Wed, Sep 30, 2020 at 05:51:05PM +0200, Charlene Wendling wrote: Hi, (This can wait post-release, the wasted time is low) Kitty is broken in the current macppc bulk: cc -MMD -DNDEBUG -Wextra -Wfloat-conversion -Wno-missing-field-initializers -Wall -Wstrict-prototypes -std=c11 -pedantic-errors -Werror -fwrapv -fstack-protector-strong -pipe -fvisibility=hidden -D_FORTIFY_SOURCE=2 -fPIC -O2 -pipe -flto -pthread -Ikitty -I/usr/local/include/python3.8 -c kittens/choose/score.c -o build/subseq_matcher-score.c.obuild/fast_data_types-charsets.c.o: file not recognized: File format not recognized My first guess was to disable LTO by doing a debug build (see setup.py:289 for rationale), and i managed to package kitty like that. At runtime i'm greeted with: [274 17:24:38.489771] [glfw error 65543]: GLX: Failed to create context: GLXBadFBConfig [274 17:24:38.496529] Failed to create GLFW temp window! This usually happens because of old/broken OpenGL drivers. kitty requires working OpenGL 3.3 drivers. And it quits. No macppc machine should be OpenGL 3.3 capable using hardware acceleration. With software rendering it leads to floating point exceptions or system lockups, and would be very slow anyway. I'm proposing to not build it on macppc. ok thfr@ Comments/feedback are welcome, Charlène. Index: Makefile === RCS file: /cvs/ports/x11/kitty/Makefile,v retrieving revision 1.2 diff -u -p -u -p -r1.2 Makefile --- Makefile7 Sep 2020 12:53:00 - 1.2 +++ Makefile30 Sep 2020 15:37:32 - @@ -1,5 +1,8 @@ # $OpenBSD: Makefile,v 1.2 2020/09/07 12:53:00 sthen Exp $ +# Requires OpenGL>=3.3 +NOT_FOR_ARCHS =powerpc + COMMENT = fast, feature full, GPU-based terminal emulator MODPY_EGG_VERSION = 0.18.3 Looking at the system requirements it would be better to flip this around to using ONLY_FOR_ARCHS with amd64 and i386 for now.
Re: Remove: games/spacehulk and games/taxipilot
On Sun, Oct 04, 2020 at 08:28:01AM +0200, Rafael Sadowski wrote: > The story is long. Time to let go. Ok to get rid of these last two Qt3 > applications/games. OK with me. -- Antoine
sparc64 bulk build report
Bulk build on sparc64-0.ports.openbsd.org Started : Wed Sep 30 21:19:52 MDT 2020 Finished: Sun Oct 4 04:29:45 MDT 2020 Duration: 3 Days 7 hours 10 minutes Built using OpenBSD 6.8 (GENERIC.MP) #489: Tue Sep 29 21:10:07 MDT 2020 Built 9762 packages Number of packages built each day: Sep 30: 3104 Oct 1: 4589 Oct 2: 1357 Oct 3: 701 Oct 4: 11 Critical path missing pkgs: http://build-failures.rhaalovely.net/sparc64/2020-09-30/summary.log Build failures: 8 http://build-failures.rhaalovely.net/sparc64/2020-09-30/games/fs2open.log http://build-failures.rhaalovely.net/sparc64/2020-09-30/graphics/inkscape.log http://build-failures.rhaalovely.net/sparc64/2020-09-30/graphics/mypaint.log http://build-failures.rhaalovely.net/sparc64/2020-09-30/productivity/gnucash.log http://build-failures.rhaalovely.net/sparc64/2020-09-30/sysutils/libvirt.log http://build-failures.rhaalovely.net/sparc64/2020-09-30/www/purritobin.log http://build-failures.rhaalovely.net/sparc64/2020-09-30/x11/grantlee-qt5.log http://build-failures.rhaalovely.net/sparc64/2020-09-30/x11/picom.log Recurrent failures: failures/games/fs2open.log failures/graphics/mypaint.log failures/productivity/gnucash.log New failures: +failures/graphics/inkscape.log Resolved failures: Packages newly built: Packages not built this time: -graphics/inkscape -meta/mate,-extras -sysutils/gamin,-python -x11/mate/utils
Re: Remove: games/spacehulk and games/taxipilot
On Sun, Oct 04, 2020 at 08:28:01AM +0200, Rafael Sadowski wrote: > The story is long. Time to let go. Ok to get rid of these last two Qt3 > applications/games. OK kn
Re: Remove: graphics/k3dsurf OR help to update to MathMod
On Sun, Oct 04, 2020 at 08:24:44AM +0200, Rafael Sadowski wrote: > Doas anyone want to port MathMod, otherwise I would suggest to remove > x11/k3dsurf? > > OK to remove or does anyone volunteer? OK to remove if noone wants to do the work.
[UPDATE MAINTAINER] sysutils/fff -> 2.2
Update fff to 2.2 Works on amd64 Any comments/Ok's? Index: Makefile === RCS file: /cvs/ports/sysutils/fff/Makefile,v retrieving revision 1.1.1.1 diff -u -p -r1.1.1.1 Makefile --- Makefile15 Jul 2020 15:36:32 - 1.1.1.1 +++ Makefile4 Oct 2020 08:25:08 - @@ -4,7 +4,7 @@ COMMENT = simple and fast file manager GH_ACCOUNT = dylanaraps GH_PROJECT = fff -GH_TAGNAME = 2.1 +GH_TAGNAME = 2.2 CATEGORIES = sysutils Index: distinfo === RCS file: /cvs/ports/sysutils/fff/distinfo,v retrieving revision 1.1.1.1 diff -u -p -r1.1.1.1 distinfo --- distinfo15 Jul 2020 15:36:32 - 1.1.1.1 +++ distinfo4 Oct 2020 08:25:08 - @@ -1,2 +1,2 @@ -SHA256 (fff-2.1.tar.gz) = d2hw0RwCL6QEaNXVgoMcCrW+ztVzSJCX3qr13WkOfqs= -SIZE (fff-2.1.tar.gz) = 15573 +SHA256 (fff-2.2.tar.gz) = RfbhCRmGyJLqReGsgvLX9UF8+zQ9xWnSYltZgOa8+2I= +SIZE (fff-2.2.tar.gz) = 16426 Index: pkg/PLIST === RCS file: /cvs/ports/sysutils/fff/pkg/PLIST,v retrieving revision 1.1.1.1 diff -u -p -r1.1.1.1 PLIST --- pkg/PLIST 15 Jul 2020 15:36:32 - 1.1.1.1 +++ pkg/PLIST 4 Oct 2020 08:25:08 - @@ -1,3 +1,5 @@ @comment $OpenBSD: PLIST,v 1.1.1.1 2020/07/15 15:36:32 bket Exp $ bin/fff @man man/man1/fff.1 +share/doc/fff/ +share/doc/fff/README.md diff Description: Binary data
Re: [macppc] Don't build x11/kitty
On Sun, 4 Oct 2020 01:12:35 -0400 Brad Smith wrote: > On 10/4/2020 1:01 AM, Thomas Frohwein wrote: > > On Wed, Sep 30, 2020 at 05:51:05PM +0200, Charlene Wendling wrote: > >> Hi, > >> > >> (This can wait post-release, the wasted time is low) > >> > >> Kitty is broken in the current macppc bulk: > >> > >>> cc -MMD -DNDEBUG -Wextra -Wfloat-conversion > >>> -Wno-missing-field-initializers -Wall -Wstrict-prototypes -std=c11 > >>> -pedantic-errors -Werror -fwrapv -fstack-protector-strong -pipe > >>> -fvisibility=hidden -D_FORTIFY_SOURCE=2 -fPIC -O2 -pipe -flto > >>> -pthread -Ikitty -I/usr/local/include/python3.8 -c > >>> kittens/choose/score.c -o > >>> build/subseq_matcher-score.c.obuild/fast_data_types-charsets.c.o: > >>> file not recognized: File format not recognized > >> My first guess was to disable LTO by doing a debug build (see > >> setup.py:289 for rationale), and i managed to package kitty like > >> that. > >> At runtime i'm greeted with: > >> > >>> [274 17:24:38.489771] [glfw error 65543]: GLX: Failed to create > >>> context: GLXBadFBConfig > >>> [274 17:24:38.496529] Failed to create GLFW temp window! This > >>> usually happens because of old/broken OpenGL drivers. kitty > >>> requires working OpenGL 3.3 drivers. > >> And it quits. No macppc machine should be OpenGL 3.3 capable using > >> hardware acceleration. With software rendering it leads to floating > >> point exceptions or system lockups, and would be very slow anyway. > >> > >> I'm proposing to not build it on macppc. > > ok thfr@ > > > >> Comments/feedback are welcome, > >> > >> Charlène. > >> > >> > >> Index: Makefile > >> === > >> RCS file: /cvs/ports/x11/kitty/Makefile,v > >> retrieving revision 1.2 > >> diff -u -p -u -p -r1.2 Makefile > >> --- Makefile 7 Sep 2020 12:53:00 - 1.2 > >> +++ Makefile 30 Sep 2020 15:37:32 - > >> @@ -1,5 +1,8 @@ > >> # $OpenBSD: Makefile,v 1.2 2020/09/07 12:53:00 sthen Exp $ > >> > >> +# Requires OpenGL>=3.3 > >> +NOT_FOR_ARCHS = powerpc > >> + > >> COMMENT =fast, feature full, GPU-based terminal > >> emulator > >> MODPY_EGG_VERSION = 0.18.3 > > > Looking at the system requirements it would be better to flip this > around to using ONLY_FOR_ARCHS with amd64 and i386 for now. > You're right, i added aarch64 as i did in games/taisei, that has the same requirement. Index: Makefile === RCS file: /cvs/ports/x11/kitty/Makefile,v retrieving revision 1.2 diff -u -p -u -p -r1.2 Makefile --- Makefile7 Sep 2020 12:53:00 - 1.2 +++ Makefile4 Oct 2020 13:44:56 - @@ -1,5 +1,8 @@ # $OpenBSD: Makefile,v 1.2 2020/09/07 12:53:00 sthen Exp $ +# Requires OpenGL>=3.3 +ONLY_FOR_ARCHS = aarch64 amd64 i386 + COMMENT = fast, feature full, GPU-based terminal emulator MODPY_EGG_VERSION =0.18.3
Re: Review for cabal.port.mk
On Sun, Oct 04, 2020 at 10:26:50AM +0200, Sebastien Marie wrote: > On Sun, Oct 04, 2020 at 09:56:22AM +0200, Marc Espie wrote: > > I'll have to look more closely, but I have a few concerns. > > > > On Sat, Oct 03, 2020 at 11:03:43PM -0700, Greg Steuck wrote: > > > MASTER_SITES0 = https://hackage.haskell.org/package/ > > That's a minor one. > > This sets a precedent of hijacking a MASTER_SITESn for a specific module. > > I don't think it's a big concern because cabal is somewhat isolated. > > We'll deal with that if we must. > > just a side note. > > technically lang/rust and lang/go already do something like that, but > in a more flexible way: the number is configurable, and by default is > a high number (9) to limit conflict with possible user usage. yep... making it configurable might be a bit complex, but it should work.
Re: Review for cabal.port.mk
Sebastien Marie writes: > On Sun, Oct 04, 2020 at 09:56:22AM +0200, Marc Espie wrote: >> I'll have to look more closely, but I have a few concerns. >> >> On Sat, Oct 03, 2020 at 11:03:43PM -0700, Greg Steuck wrote: >> > MASTER_SITES0 =https://hackage.haskell.org/package/ >> That's a minor one. >> This sets a precedent of hijacking a MASTER_SITESn for a specific module. >> I don't think it's a big concern because cabal is somewhat isolated. >> We'll deal with that if we must. > > just a side note. > > technically lang/rust and lang/go already do something like that, but > in a more flexible way: the number is configurable, and by default is > a high number (9) to limit conflict with possible user usage. I'm happy to switch to 9. I'm reticent to introduce a knob considering neither MODGO_MASTER_SITESN nor MODCARGO_MASTER_SITESN attracted any usage I could find. The latter has been around for 4 years. YAGNI? Thanks Greg
[pretest] openvpn-2.5_rc2
Updated version of the beta1 diff I sent earlier, for those who care.
I did not receive feedback so far but it works fine for my client-only
setup, so I'll probably commit this when the release is out.
Index: Makefile
===
RCS file: /d/cvs/ports/net/openvpn/Makefile,v
retrieving revision 1.101
diff -u -p -r1.101 Makefile
--- Makefile17 May 2020 08:53:27 - 1.101
+++ Makefile4 Oct 2020 21:07:54 -
@@ -2,8 +2,8 @@
COMMENT= easy-to-use, robust, and highly configurable VPN
-DISTNAME= openvpn-2.4.9
-REVISION= 0
+DISTNAME= openvpn-2.5_rc2
+PKGNAME= openvpn-2.5rc2
CATEGORIES=net security
@@ -18,14 +18,18 @@ WANTLIB += c lz4 lzo2
MASTER_SITES= https://swupdate.openvpn.org/community/releases/
+BUILD_DEPENDS= textproc/py-docutils,python3
LIB_DEPENDS= archivers/lzo2 \
archivers/lz4
+TEST_DEPENDS= devel/cmocka
-SEPARATE_BUILD=Yes
+#SEPARATE_BUILD= Yes
CONFIGURE_STYLE= gnu
-CONFIGURE_ENV= CPPFLAGS="-I${LOCALBASE}/include" \
- LDFLAGS="-L${LOCALBASE}/lib ${LDFLAGS}"
+CONFIGURE_ENV= CPPFLAGS="-I${LOCALBASE}/include" \
+ LDFLAGS="-L${LOCALBASE}/lib ${LDFLAGS}" \
+ RST2HTML="${LOCALBASE}/bin/rst2html-3" \
+ RST2MAN="${LOCALBASE}/bin/rst2man-3"
DEBUG_PACKAGES=${BUILD_PACKAGES}
Index: distinfo
===
RCS file: /d/cvs/ports/net/openvpn/distinfo,v
retrieving revision 1.44
diff -u -p -r1.44 distinfo
--- distinfo21 Apr 2020 23:43:55 - 1.44
+++ distinfo4 Oct 2020 21:08:12 -
@@ -1,2 +1,2 @@
-SHA256 (openvpn-2.4.9.tar.gz) = RrJo74jmfKbeLp8ZlD655ayFROVfXB869ncpjQPmS24=
-SIZE (openvpn-2.4.9.tar.gz) = 1455736
+SHA256 (openvpn-2.5_rc2.tar.gz) = yUHlNgqNsMkHD1V4KG7H9IIqeeuJKHUtkdRCXiGtnxk=
+SIZE (openvpn-2.5_rc2.tar.gz) = 1799431
Index: patches/patch-configure
===
RCS file: /d/cvs/ports/net/openvpn/patches/patch-configure,v
retrieving revision 1.22
diff -u -p -r1.22 patch-configure
--- patches/patch-configure 21 Apr 2020 23:43:55 - 1.22
+++ patches/patch-configure 23 Aug 2020 11:04:42 -
@@ -2,7 +2,7 @@ $OpenBSD: patch-configure,v 1.22 2020/04
Index: configure
--- configure.orig
+++ configure
-@@ -18146,7 +18146,7 @@ else
+@@ -18249,7 +18249,7 @@ else
fi
Index: patches/patch-include_Makefile_in
===
RCS file: /d/cvs/ports/net/openvpn/patches/patch-include_Makefile_in,v
retrieving revision 1.11
diff -u -p -r1.11 patch-include_Makefile_in
--- patches/patch-include_Makefile_in 21 Apr 2020 23:43:55 - 1.11
+++ patches/patch-include_Makefile_in 23 Aug 2020 11:04:42 -
@@ -2,7 +2,7 @@ $OpenBSD: patch-include_Makefile_in,v 1.
Index: include/Makefile.in
--- include/Makefile.in.orig
+++ include/Makefile.in
-@@ -336,7 +336,7 @@ host_cpu = @host_cpu@
+@@ -339,7 +339,7 @@ host_cpu = @host_cpu@
host_os = @host_os@
host_vendor = @host_vendor@
htmldir = @htmldir@
Index: patches/patch-sample_sample-config-files_static-home_conf
===
RCS file: patches/patch-sample_sample-config-files_static-home_conf
diff -N patches/patch-sample_sample-config-files_static-home_conf
--- patches/patch-sample_sample-config-files_static-home_conf 29 Nov 2016
09:22:02 - 1.2
+++ /dev/null 1 Jan 1970 00:00:00 -
@@ -1,17 +0,0 @@
-$OpenBSD: patch-sample_sample-config-files_static-home_conf,v 1.2 2016/11/29
09:22:02 jca Exp $
sample/sample-config-files/static-home.conf.orig Thu Nov 3 09:49:49 2016
-+++ sample/sample-config-files/static-home.confFri Nov 18 17:49:59 2016
-@@ -40,10 +40,10 @@ cipher AES-256-CBC
- ; port 1194
-
- # Downgrade UID and GID to
--# "nobody" after initialization
-+# "_openvpn" after initialization
- # for extra security.
--; user nobody
--; group nobody
-+user _openvpn
-+group _openvpn
-
- # If you built OpenVPN with
- # LZO compression, uncomment
Index: patches/patch-sample_sample-config-files_static-office_conf
===
RCS file: patches/patch-sample_sample-config-files_static-office_conf
diff -N patches/patch-sample_sample-config-files_static-office_conf
--- patches/patch-sample_sample-config-files_static-office_conf 29 Nov 2016
09:22:02 - 1.2
+++ /dev/null 1 Jan 1970 00:00:00 -
@@ -1,17 +0,0 @@
-$OpenBSD: patch-sample_sample-config-files_static-office_conf,v 1.2 2016/11/29
09:22:02 jca Exp $
sample/sample-config-files/static-office.conf.orig Thu Nov 3 09:49:49 2016
-+++ sample/sample-config-files/static-office.conf Fri Nov 18 17:49:59 2016
-@@ -37,10 +37,10 @@ cipher AES-256-CBC
- ; port 1194
-
- # Downgrade UID and GID to
--# "nobody" after initialization
Re: Review for cabal.port.mk
Thanks for the prompt review!
Marc Espie writes:
> I'll have to look more closely, but I have a few concerns.
>
> On Sat, Oct 03, 2020 at 11:03:43PM -0700, Greg Steuck wrote:
>> MASTER_SITES0 = https://hackage.haskell.org/package/
> That's a minor one.
> This sets a precedent of hijacking a MASTER_SITESn for a specific module.
> I don't think it's a big concern because cabal is somewhat isolated.
> We'll deal with that if we must.
Moved to :9 to reduce the chance of collision. Unless you believe an
indirection is warranted I'll keep it at this.
>> DIST_SUBDIR ?= hackage
>>
>> # The .cabal files are explicitly copied over the ones extracted from
>> # archives by the normal extraction rules.
>> EXTRACT_CASES += *.cabal) ;;
>
> I'm not fond of this solution to the PKGSTEM issue.
Is this comment about EXTRACT_CASES above or does it belong with
DISTNAME below?
> I would prefer that you would reconstitute DISTNAME from MODCABAL variables
> exactly like stuff is done for github
There's a minor tension which I'm not sure how to resolve
best. Implementing your suggestion may complicate porting the
self-hosted main binary packages that aren't published to hackage, yet
grab the dependencies there.
So far all the cases I have are hackage packages. Which means I should
follow YAGNI and implement your suggestion. I'll give it a try, but
wanted to explain why I didn't do this before.
>> MODCABAL_HACKAGE_NAME = ${DISTNAME:C,-[0-9.]*$,,}
>> MODCABAL_HACKAGE_VERSION = ${DISTNAME:C,.*-([0-9.]*)$,\1,}
> and I think those variables might be a bit long, though very descriptive
> what's wrong with MODCABAL_STEM and MODCABAL_VERSION ?
Perfect, done.
>> .if !target(do-build)
>> do-build:
>> @${_MODCABAL_BUILD_TARGET}
>> .endif
>>
>> .if !target(do-install)
>> do-install:
>> @${_MODCABAL_INSTALL_TARGET}
>> .endif
>
> let's think these through. _MODCABAL_{BUILD,INSTALL}_TARGET are internal
> so not reusable from outside, strictly speaking.
>
> (this is probably taken from ruby.port.mk or something similar)
I was reading bsd.port.mk(5) and port-modules(5) instead. They provided
enough guidance to not require much peeking into examples. I probably
cribbed a thing or two from ghc.port.mk, but that's fair game :)
> If you define do-build/do-install manually, you might want to reference
> these.
>
> The reason the redirection exist is to allow non trivial build/install to
> happen when several modules are combined.
>
> in general, you may need to make these visible.
I erred on the side of hiding everything that I must not expose. It's
easier to expose than take away later. Do you feel I'm being too
cautious?
Thanks
Greg
Re: Review for cabal.port.mk
Marc Espie writes:
> I would prefer that you would reconstitute DISTNAME from MODCABAL variables
> exactly like stuff is done for github.
Does the diff below look like what you had in mind? I like it as the
data becomes more structured (parsing is almost never the best choice).
I can imagine adding some error checks to help the users.
diff --git devel/alex/Makefile devel/alex/Makefile
index 77a2f8fec5f..e19050f77f0 100644
--- devel/alex/Makefile
+++ devel/alex/Makefile
@@ -2,7 +2,9 @@
COMMENT = lexical analyser generator for Haskell
-DISTNAME = alex-3.2.5
+MODCABAL_STEM = alex
+MODCABAL_VERSION = 3.2.5
+MODCABAL_DATA_DIR = data
CATEGORIES = devel
HOMEPAGE = http://www.haskell.org/alex/
@@ -13,7 +15,6 @@ PERMIT_PACKAGE = Yes
WANTLIB = c charset ffi gmp iconv m pthread util
MODULES = devel/cabal
-MODCABAL_DATA_DIR= data
LIB_DEPENDS = converters/libiconv \
devel/gmp \
diff --git devel/cabal/cabal.port.mk devel/cabal/cabal.port.mk
index 18a082206d9..3a5859266c9 100644
--- devel/cabal/cabal.port.mk
+++ devel/cabal/cabal.port.mk
@@ -3,8 +3,10 @@
# Module for building Haskell programs with cabal-install.
# Inspired by FreeBSD cabal.mk by Gleb Popov.
-# Available input variables:
+# Input variables:
#
+# MODCABAL_STEM - the name of the package on hackage.
+# MODCABAL_VERSION - the version of the package.
# MODCABAL_MANIFEST - hackage dependencies required by this package, tripples
#of space separate . Typically generated
#with cabal-bundler program from cabal-extras. The patch pending review
@@ -46,8 +48,7 @@ DIST_SUBDIR ?=hackage
# archives by the normal extraction rules.
EXTRACT_CASES += *.cabal) ;;
-MODCABAL_STEM =${DISTNAME:C,-[0-9.]*$,,}
-MODCABAL_VERSION = ${DISTNAME:C,.*-([0-9.]*)$,\1,}
+DISTNAME ?=${MODCABAL_STEM}-${MODCABAL_VERSION}
HOMEPAGE ?=${MASTER_SITES9}${MODCABAL_STEM}
MASTER_SITES ?=${MASTER_SITES9}${DISTNAME}/
DISTFILES ?= ${DISTNAME}.tar.gz
diff --git devel/happy/Makefile devel/happy/Makefile
index fbe2ab0a1bb..77b0888eba0 100644
--- devel/happy/Makefile
+++ devel/happy/Makefile
@@ -2,7 +2,10 @@
COMMENT= parser generator for Haskell
-DISTNAME= happy-1.19.12
+MODCABAL_STEM= happy
+MODCABAL_VERSION= 1.19.12
+MODCABAL_DATA_DIR= data
+
CATEGORIES=devel
HOMEPAGE= http://www.haskell.org/happy/
@@ -14,8 +17,6 @@ WANTLIB= c ffi gmp iconv m pthread
MODULES= devel/cabal
-MODCABAL_DATA_DIR= data
-
LIB_DEPENDS= converters/libiconv \
devel/gmp \
devel/libffi
UPDATE: astro/gnuastro 0.12 => 0.13
Hello ports --
Attached is an update to GNU Astro. The NEWS file contains all the
major changes:
https://git.savannah.gnu.org/cgit/gnuastro.git/plain/NEWS?id=gnuastro_v0.13
Note that the shared library major is bumped due to removed symbols.
While here, I'd like to take maintainer and also add a homepage that
was missing from import.
All tests pass on amd64.
OK?
(Now to the correct mailing list--thanks solene!)
~Brian
Index: Makefile
===
RCS file: /cvs/ports/astro/gnuastro/Makefile,v
retrieving revision 1.1.1.1
diff -u -p -r1.1.1.1 Makefile
--- Makefile 24 Jun 2020 19:26:25 - 1.1.1.1
+++ Makefile 3 Oct 2020 16:21:31 -
@@ -1,10 +1,13 @@
# $OpenBSD: Makefile,v 1.1.1.1 2020/06/24 19:26:25 bcallah Exp $
COMMENT = GNU Astronomy Utilities
-DISTNAME = gnuastro-0.12
+DISTNAME = gnuastro-0.13
CATEGORIES = astro
-SHARED_LIBS += gnuastro 0.0 # 10.0
+SHARED_LIBS += gnuastro 1.0 # 11.0
+
+HOMEPAGE = https://www.gnu.org/software/gnuastro/
+MAINTAINER = Brian Callahan
# GPLv3+
PERMIT_PACKAGE = Yes
Index: distinfo
===
RCS file: /cvs/ports/astro/gnuastro/distinfo,v
retrieving revision 1.1.1.1
diff -u -p -r1.1.1.1 distinfo
--- distinfo 24 Jun 2020 19:26:25 - 1.1.1.1
+++ distinfo 3 Oct 2020 16:21:31 -
@@ -1,2 +1,2 @@
-SHA256 (gnuastro-0.12.tar.gz) = Jy784mtsgLrIYkjrjqLgY4AbHlAd7PZflWusW5FIK5s=
-SIZE (gnuastro-0.12.tar.gz) = 5559331
+SHA256 (gnuastro-0.13.tar.gz) = YmT76h+z0f+Xi3Dm0w50tqqgpu+EFJvszknAQ5+HHiU=
+SIZE (gnuastro-0.13.tar.gz) = 5638209
Index: patches/patch-configure
===
RCS file: /cvs/ports/astro/gnuastro/patches/patch-configure,v
retrieving revision 1.1.1.1
diff -u -p -r1.1.1.1 patch-configure
--- patches/patch-configure 24 Jun 2020 19:26:25 - 1.1.1.1
+++ patches/patch-configure 3 Oct 2020 16:21:31 -
@@ -5,7 +5,7 @@ Remove hardcoded -O3
Index: configure
--- configure.orig
+++ configure
-@@ -19378,7 +19378,7 @@ fi
+@@ -20265,7 +20265,7 @@ fi
if test "x$enable_debug" = "xyes"; then :
cflags_add="-g -O0"; enable_shared=no
else
@@ -14,7 +14,7 @@ Index: configure
fi
CFLAGS="-Wall $cflags_add $CFLAGS"
CXXFLAGS="-Wall $cflags_add $CXXFLAGS"
-@@ -27470,7 +27470,7 @@ $as_echo_n "checking for shell to use with libtool...
+@@ -28419,7 +28419,7 @@ $as_echo_n "checking for shell to use with libtool...
cprog=libtool_shell_test.c
$as_echo "#include "[] > $cprog
$as_echo "int main(void){printf(\"success\\n\"); return 0;}" >> $cprog
Index: pkg/PLIST
===
RCS file: /cvs/ports/astro/gnuastro/pkg/PLIST,v
retrieving revision 1.1.1.1
diff -u -p -r1.1.1.1 PLIST
--- pkg/PLIST 24 Jun 2020 19:26:25 - 1.1.1.1
+++ pkg/PLIST 3 Oct 2020 16:21:31 -
@@ -32,6 +32,7 @@ include/gnuastro/fits.h
include/gnuastro/git.h
include/gnuastro/interpolate.h
include/gnuastro/jpeg.h
+include/gnuastro/kdtree.h
include/gnuastro/label.h
include/gnuastro/list.h
include/gnuastro/match.h
Re: [update] deve/py-llvmlite to 0.34.0
On 10/3/20 3:56 AM, Remi Pointel wrote:
> On 9/29/20 1:03 AM, Aisha Tammy wrote:
>> Hi,
>> Attached diff for py-llvmlite update to latest
>> It has been broken for a bit and am not sure if maintainer is too busy for
>> this.
>> If needed I can take this port, I am generally in touch with upstream to bug
>> them
>> about their JIT compiler numba.
>>
>> Would be nice to see it in 6.8 as its working now.
>> All tests are passing as well.
>>
>> (also attaching the patch just in case thunderbird still messes up)
>>
>> Aisha
>
> Hi,
>
> thank you for the diff.
>
> I think miasm will not work with your update, I will verify if recent version
> is compatible with Python 3.
>
> Best regards,
>
> Remi.
>
Hi,
I've attached a cleaner version, removing COMMON_ENV.
I've not had time to test py-miasm but it does seem to have python3.
Thanks,
Aisha
diff --git a/devel/py-llvmlite/Makefile b/devel/py-llvmlite/Makefile
index c20460f77e8..b530cf3321c 100644
--- a/devel/py-llvmlite/Makefile
+++ b/devel/py-llvmlite/Makefile
@@ -1,10 +1,8 @@
# $OpenBSD: Makefile,v 1.9 2020/08/22 22:06:01 naddy Exp $
-BROKEN = requires update to 0.34.0 for LLVM 10
COMMENT = lightweight LLVM-Python binding for writing JIT compilers
-MODPY_EGG_VERSION =0.29.0
-REVISION = 1
+MODPY_EGG_VERSION =0.34.0
GH_ACCOUNT = numba
GH_PROJECT = llvmlite
GH_TAGNAME = v${MODPY_EGG_VERSION}
@@ -19,20 +17,22 @@ HOMEPAGE = http://llvmlite.pydata.org/
# BSD
PERMIT_PACKAGE = Yes
-WANTLIB += ${COMPILER_LIBCXX} LLVM m
+WANTLIB += ${COMPILER_LIBCXX} LLVM m
COMPILER = base-clang
MODULES = lang/python
FLAVORS = python3
-FLAVOR ?=
+FLAVOR ?= python3
-.if !${FLAVOR:Mpython3}
-BUILD_DEPENDS += devel/py-enum34
-RUN_DEPENDS += devel/py-enum34
-.endif
+MAKE_ENV +=LLVM_CONFIG="/usr/bin/llvm-config" \
+ LDLIBS="`llvm-config --libs all`" \
+ CXXFLAGS="`llvm-config --cxxflags` -fPIC ${CXXFLAGS}" \
+ LDFLAGS="`llvm-config --ldflags`"
-MAKE_ENV = LLVM_CONFIG="/usr/bin/llvm-config"
+pre-build:
+ cd ${WRKSRC} && env -i ${MAKE_ENV} ${MAKE_PROGRAM} ${MAKE_FLAGS} \
+ -f LIB_MAKEFILE
do-test:
cd ${WRKSRC} && PYTHONPATH=. ${SETENV} ${MODPY_BIN} ./runtests.py
diff --git a/devel/py-llvmlite/distinfo b/devel/py-llvmlite/distinfo
index 464df2c98f1..4e116070593 100644
--- a/devel/py-llvmlite/distinfo
+++ b/devel/py-llvmlite/distinfo
@@ -1,2 +1,2 @@
-SHA256 (llvmlite-0.29.0.tar.gz) = vO54HC3Ga+09tbqF9cBMTv/TjHwQ9Sh+6+qBy029zjQ=
-SIZE (llvmlite-0.29.0.tar.gz) = 196507
+SHA256 (llvmlite-0.34.0.tar.gz) = rqXDPVkUW5YlHRGXG60m2BgQliq2g9EHtu9KGEctPZo=
+SIZE (llvmlite-0.34.0.tar.gz) = 210438
diff --git a/devel/py-llvmlite/patches/patch-LIB_MAKEFILE
b/devel/py-llvmlite/patches/patch-LIB_MAKEFILE
new file mode 100644
index 000..0e491ddbf54
--- /dev/null
+++ b/devel/py-llvmlite/patches/patch-LIB_MAKEFILE
@@ -0,0 +1,13 @@
+$OpenBSD$
+
+upstream libllvmlite.so doesn't build nicely
+
+Index: config_makefile
+--- LIB_MAKEFILE.orig
LIB_MAKEFILE
+@@ -0,0 +1,5 @@
++SRC=ffi/*.cpp
++
++ffi/libllvmlite.so:
++ $(CXX) -shared $(CXXFLAGS) $(LDFLAGS) -o $@ $(SRC) $(LDLIBS)
++
diff --git a/devel/py-llvmlite/pkg/PLIST b/devel/py-llvmlite/pkg/PLIST
index dcbc4e9780d..1a56a66754b 100644
--- a/devel/py-llvmlite/pkg/PLIST
+++ b/devel/py-llvmlite/pkg/PLIST
@@ -5,7 +5,6 @@ lib/python${MODPY_VERSION}/site-packages/llvmlite/__init__.py
${MODPY_COMMENT}lib/python${MODPY_VERSION}/site-packages/llvmlite/${MODPY_PYCACHE}/
lib/python${MODPY_VERSION}/site-packages/llvmlite/${MODPY_PYCACHE}__init__.${MODPY_PYC_MAGIC_TAG}pyc
lib/python${MODPY_VERSION}/site-packages/llvmlite/${MODPY_PYCACHE}_version.${MODPY_PYC_MAGIC_TAG}pyc
-lib/python${MODPY_VERSION}/site-packages/llvmlite/${MODPY_PYCACHE}six.${MODPY_PYC_MAGIC_TAG}pyc
lib/python${MODPY_VERSION}/site-packages/llvmlite/${MODPY_PYCACHE}utils.${MODPY_PYC_MAGIC_TAG}pyc
lib/python${MODPY_VERSION}/site-packages/llvmlite/_version.py
lib/python${MODPY_VERSION}/site-packages/llvmlite/binding/
@@ -34,7 +33,7 @@
lib/python${MODPY_VERSION}/site-packages/llvmlite/binding/dylib.py
lib/python${MODPY_VERSION}/site-packages/llvmlite/binding/executionengine.py
lib/python${MODPY_VERSION}/site-packages/llvmlite/binding/ffi.py
lib/python${MODPY_VERSION}/site-packages/llvmlite/binding/initfini.py
-lib/python${MODPY_VERSION}/site-packages/llvmlite/binding/libllvmlite.so
+@so lib/python${MODPY_VERSION}/site-packages/llvmlite/binding/libllvmlite.so
lib/python${MODPY_VERSION}/site-packages/llvmlite/binding/linker.py
lib/python${MODPY_VERSION}/site-packages/llvmlite/binding/module.py
lib/python${MODPY_VERSION}/site-packages/llvmlite/binding/object_file.py
@@ -71,7 +70,6 @@
lib/python${MODPY_VERSION}/site-packages/llvmlite/llvmpy/${MODPY_PYCACHE}core.${
lib/python${MODPY_VERSION}/site-packages/llvmlite/llvmpy/${MODPY_PYCACHE}passes.${MODPY_PYC_MAGIC_TAG}pyc
lib/python${MODPY_VERSION}/site-packages/llvmlite/ll
Re: update games/jag 0.3.8
Nam Nguyen writes:
ping
> Here is an update for games/jag 0.3.8, released September 7, 2020.
> changelog: https://gitlab.com/coringao/jag/-/tags/0.3.8
>
> This update:
> - switches to -std=gnu++17 by default
> - removes gameprofile.cpp and displaywrapper.cpp patches, as they have
> already been added upstream
> https://gitlab.com/coringao/jag/-/commit/9edca4bdfdb8a049ad6ca0628b3d427e461d17f5
> https://gitlab.com/coringao/jag/-/commit/4f67546bc8e8b2ac60156605ad32fa0a3880d5dd
> - PERMIT_PACKAGE = Yes due to removal of all themes except for the
> default theme
> discussion: https://gitlab.com/coringao/jag/-/issues/1
>
> All questionable assets were removed so I propose PERMIT_PACKAGE =
> Yes. I still could not find a license for the default theme made by
> PixelMixer. Given that the default theme's assets are featured on the
> website and they were not made by Fasticon, they should fall under the
> JAG menu's statement: "Thanks to the authors of graphics, icons and
> sounds, which are used under free licenses in the game." Hopefully this
> is sufficient for PERMIT_PACKAGE = Yes.
>
> Some warnings remain and seem safe to ignore:
> - deprecated os info testing for windows vista
> - overloaded exec(const Qstring &qs) vs exec() in gamemenu.h and
> qdialog.h. gamemenu.h's version is always called.
>
> In file included from ../src/menucontrol.cpp:26:
> ../src/gamemenu.h:55:9: warning: 'ConfirmDialog::exec' hides overloaded
> virtual
> int exec(const QString &qs);
> ^
>
> /usr/local/include/X11/qt5/QtWidgets/qdialog.h:95:17: note: hidden
> overloaded virtual function 'QDialog::exec' declared here: different
> number of parameters (0 vs 1)
> virtual int exec();
>
> I tested jag and jag-editor runtimes. Feedback and tests are welcome.
Index: Makefile
===
RCS file: /cvs/ports/games/jag/Makefile,v
retrieving revision 1.12
diff -u -p -r1.12 Makefile
--- Makefile1 Aug 2020 05:34:08 - 1.12
+++ Makefile18 Sep 2020 22:20:57 -
@@ -1,14 +1,12 @@
# $OpenBSD: Makefile,v 1.12 2020/08/01 05:34:08 landry Exp $
COMMENT = arcade-puzzle game
-V =0.3.7
+V =0.3.8
DISTNAME = jag-${V}
CATEGORIES = games
HOMEPAGE = https://gitlab.com/coringao/jag
# GPLv3
-# https://gitlab.com/coringao/jag/-/issues/1
-PERMIT_PACKAGE = restrictively licensed themes
-PERMIT_DISTFILES = restrictively licensed themes
+PERMIT_PACKAGE = Yes
WANTLIB += ${COMPILER_LIBCXX} Qt5Core Qt5Gui Qt5OpenGL Qt5Widgets
WANTLIB += Qt5X11Extras SDL2 SDL2_mixer X11 Xrandr c
Index: distinfo
===
RCS file: /cvs/ports/games/jag/distinfo,v
retrieving revision 1.2
diff -u -p -r1.2 distinfo
--- distinfo1 Aug 2020 05:34:08 - 1.2
+++ distinfo18 Sep 2020 22:20:57 -
@@ -1,2 +1,2 @@
-SHA256 (jag-0.3.7.tar.gz) = jI+SYKL476N4ojzhUdWjoEDklJ+EFBClnRDUzeJH9t4=
-SIZE (jag-0.3.7.tar.gz) = 6971529
+SHA256 (jag-0.3.8.tar.gz) = NG07uNFmccGCsFaLkpg7jVM6Caz7PQPpSneO6HnS1gQ=
+SIZE (jag-0.3.8.tar.gz) = 3025319
Index: patches/patch-src_displaywrapper_cpp
===
RCS file: patches/patch-src_displaywrapper_cpp
diff -N patches/patch-src_displaywrapper_cpp
--- patches/patch-src_displaywrapper_cpp1 Aug 2020 05:34:08 -
1.1
+++ /dev/null 1 Jan 1970 00:00:00 -
@@ -1,23 +0,0 @@
-$OpenBSD: patch-src_displaywrapper_cpp,v 1.1 2020/08/01 05:34:08 landry Exp $
-
-Index: src/displaywrapper.cpp
src/displaywrapper.cpp.orig
-+++ src/displaywrapper.cpp
-@@ -305,8 +305,6 @@ void DisplayWrapper::dw_init(bool filter, int minWidth
- //short original_rate = XRRConfigCurrentRate(conf);
- //SizeID original_size_id = XRRConfigCurrentConfiguration(conf,
&original_rotation);
-
-- XCloseDisplay(dpy);
--
- for (int i = 0; i < num_sizes; i++)
- {
- int width = xrrs[i].width;
-@@ -325,6 +323,8 @@ void DisplayWrapper::dw_init(bool filter, int minWidth
- modes.append(mode);
- }
- }
-+
-+ XCloseDisplay(dpy);
-
- if (filter)
- modeNames.removeDuplicates();
Index: patches/patch-src_editor_jag-editor_pro
===
RCS file: /cvs/ports/games/jag/patches/patch-src_editor_jag-editor_pro,v
retrieving revision 1.1
diff -u -p -r1.1 patch-src_editor_jag-editor_pro
--- patches/patch-src_editor_jag-editor_pro 1 Aug 2020 05:34:08 -
1.1
+++ patches/patch-src_editor_jag-editor_pro 18 Sep 2020 22:20:57 -
@@ -5,7 +5,7 @@ install jag-editor to correct prefix. ja
Index: src/editor/jag-editor.pro
--- src/editor/jag-editor.pro.orig
+++ src/editor/jag-editor.pro
-@@ -22,11 +22,8 @@ QMAKE_CXXFLAGS += -g -std=gnu++14 -D_FORTIFY_SOURCE=2
+@@ -22,11 +22,8 @@ QMAKE_CXXFLAGS += -g -std=gnu++17 -D_FORTIFY_SOURCE=2
QMAKE_LFLAGS += -lpthread -fPIE -pie -Wl,-
Re: [macppc] Don't build x11/kitty
On 10/4/2020 9:45 AM, Charlene Wendling wrote: On Sun, 4 Oct 2020 01:12:35 -0400 Brad Smith wrote: On 10/4/2020 1:01 AM, Thomas Frohwein wrote: On Wed, Sep 30, 2020 at 05:51:05PM +0200, Charlene Wendling wrote: Hi, (This can wait post-release, the wasted time is low) Kitty is broken in the current macppc bulk: cc -MMD -DNDEBUG -Wextra -Wfloat-conversion -Wno-missing-field-initializers -Wall -Wstrict-prototypes -std=c11 -pedantic-errors -Werror -fwrapv -fstack-protector-strong -pipe -fvisibility=hidden -D_FORTIFY_SOURCE=2 -fPIC -O2 -pipe -flto -pthread -Ikitty -I/usr/local/include/python3.8 -c kittens/choose/score.c -o build/subseq_matcher-score.c.obuild/fast_data_types-charsets.c.o: file not recognized: File format not recognized My first guess was to disable LTO by doing a debug build (see setup.py:289 for rationale), and i managed to package kitty like that. At runtime i'm greeted with: [274 17:24:38.489771] [glfw error 65543]: GLX: Failed to create context: GLXBadFBConfig [274 17:24:38.496529] Failed to create GLFW temp window! This usually happens because of old/broken OpenGL drivers. kitty requires working OpenGL 3.3 drivers. And it quits. No macppc machine should be OpenGL 3.3 capable using hardware acceleration. With software rendering it leads to floating point exceptions or system lockups, and would be very slow anyway. I'm proposing to not build it on macppc. ok thfr@ Comments/feedback are welcome, Charlène. Index: Makefile === RCS file: /cvs/ports/x11/kitty/Makefile,v retrieving revision 1.2 diff -u -p -u -p -r1.2 Makefile --- Makefile7 Sep 2020 12:53:00 - 1.2 +++ Makefile30 Sep 2020 15:37:32 - @@ -1,5 +1,8 @@ # $OpenBSD: Makefile,v 1.2 2020/09/07 12:53:00 sthen Exp $ +# Requires OpenGL>=3.3 +NOT_FOR_ARCHS =powerpc + COMMENT =fast, feature full, GPU-based terminal emulator MODPY_EGG_VERSION = 0.18.3 Looking at the system requirements it would be better to flip this around to using ONLY_FOR_ARCHS with amd64 and i386 for now. You're right, i added aarch64 as i did in games/taisei, that has the same requirement. Index: Makefile === RCS file: /cvs/ports/x11/kitty/Makefile,v retrieving revision 1.2 diff -u -p -u -p -r1.2 Makefile --- Makefile7 Sep 2020 12:53:00 - 1.2 +++ Makefile4 Oct 2020 13:44:56 - @@ -1,5 +1,8 @@ # $OpenBSD: Makefile,v 1.2 2020/09/07 12:53:00 sthen Exp $ +# Requires OpenGL>=3.3 +ONLY_FOR_ARCHS = aarch64 amd64 i386 + COMMENT = fast, feature full, GPU-based terminal emulator MODPY_EGG_VERSION = 0.18.3 That doesn't make sense. There are no hardware accelerated drivers on aarch64.
