2018-05-28 21:12:29 +02:00
|
|
|
#!/bin/bash
|
2009-08-26 17:00:38 +02:00
|
|
|
|
Tue Jul 18 19:58:10 UTC 2023
a/tar-1.35-x86_64-1.txz: Upgraded.
x/mesa-23.2.0_rc1-x86_64-1.txz: Upgraded.
OK, usually I won't use rc versions even in -current, but in this case I'm
going to. Some time ago my desktop machine with RS880 / Radeon HD 4290
graphics on the motherboard began acting up with the screen going black
for a few seconds before returning. This after an hour or so of uptime,
usually, then becoming more frequent with more uptime. Eventually I'd lose
mouse and/or keyboard too, and have to reboot. Here's a couple of errors
from dmesg:
[ 9942.677675] [drm:r600_ib_test [radeon]] *ERROR* radeon: fence wait
timed out.
[ 9942.677741] [drm:radeon_ib_ring_tests [radeon]] *ERROR* radeon: failed
testing IB on GFX ring (-110).
I also noticed that the backtrace started with ttm_bo_release, and seeing
this in recent kernel patches had been chalking this up to a kernel bug.
I *still* think it could be, and there are a bunch of kernel patches coming
soon to -stable from Alex Deucher that could address the underlying causes
(not for 6.1.39 though, unfortunately). Anyway, I'd recently figured out
that reverting Mesa sufficiently made the issue go away. And now it seems
this 23.2.0 release candidate also fixes the issue.
Yes, I could go search for the commits to cherry-pick, but we'll be moving
to mesa-23.2.0 when it's released, so we might as well start testing now.
2023-07-18 21:58:10 +02:00
|
|
|
# Copyright 2006, 2007, 2008, 2009, 2010, 2011, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020, 2021, 2022, 2023 Patrick J. Volkerding, Sebeka, MN, USA
|
2009-08-26 17:00:38 +02:00
|
|
|
# All rights reserved.
|
|
|
|
#
|
|
|
|
# Redistribution and use of this script, with or without modification, is
|
|
|
|
# permitted provided that the following conditions are met:
|
|
|
|
#
|
|
|
|
# 1. Redistributions of this script must retain the above copyright
|
|
|
|
# notice, this list of conditions and the following disclaimer.
|
|
|
|
#
|
|
|
|
# THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
|
|
|
|
# WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
|
|
|
|
# MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
|
|
|
|
# EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
|
|
|
|
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
|
|
|
|
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
|
|
|
|
# OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
|
|
|
|
# WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
|
|
|
|
# OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
|
|
|
|
# ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
|
|
|
2018-05-28 21:12:29 +02:00
|
|
|
cd $(dirname $0) ; CWD=$(pwd)
|
|
|
|
|
2009-08-26 17:00:38 +02:00
|
|
|
PKGNAM=mesa
|
Tue Jul 18 19:58:10 UTC 2023
a/tar-1.35-x86_64-1.txz: Upgraded.
x/mesa-23.2.0_rc1-x86_64-1.txz: Upgraded.
OK, usually I won't use rc versions even in -current, but in this case I'm
going to. Some time ago my desktop machine with RS880 / Radeon HD 4290
graphics on the motherboard began acting up with the screen going black
for a few seconds before returning. This after an hour or so of uptime,
usually, then becoming more frequent with more uptime. Eventually I'd lose
mouse and/or keyboard too, and have to reboot. Here's a couple of errors
from dmesg:
[ 9942.677675] [drm:r600_ib_test [radeon]] *ERROR* radeon: fence wait
timed out.
[ 9942.677741] [drm:radeon_ib_ring_tests [radeon]] *ERROR* radeon: failed
testing IB on GFX ring (-110).
I also noticed that the backtrace started with ttm_bo_release, and seeing
this in recent kernel patches had been chalking this up to a kernel bug.
I *still* think it could be, and there are a bunch of kernel patches coming
soon to -stable from Alex Deucher that could address the underlying causes
(not for 6.1.39 though, unfortunately). Anyway, I'd recently figured out
that reverting Mesa sufficiently made the issue go away. And now it seems
this 23.2.0 release candidate also fixes the issue.
Yes, I could go search for the commits to cherry-pick, but we'll be moving
to mesa-23.2.0 when it's released, so we might as well start testing now.
2023-07-18 21:58:10 +02:00
|
|
|
VERSION=${VERSION:-$(echo $PKGNAM-[0-9]*.tar.?z | rev | cut -f 3- -d . | rev | cut -f 2- -d -)}
|
2022-08-09 21:25:22 +02:00
|
|
|
AMBERVERS=${AMBERVERS:-21.3.9}
|
2022-08-09 01:29:31 +02:00
|
|
|
DEMOVERS=${DEMOVERS:-8.5.0}
|
2023-09-22 23:51:12 +02:00
|
|
|
BUILD=${BUILD:-2}
|
2022-08-09 01:29:31 +02:00
|
|
|
BUILD_AMBER=${BUILD_AMBER:-YES}
|
2019-10-09 22:33:48 +02:00
|
|
|
BUILD_DEMOS=${BUILD_DEMOS:-YES}
|
2022-09-30 19:52:21 +02:00
|
|
|
CODECS=${CODECS:-h264dec,h264enc,h265dec,h265enc,vc1dec}
|
2009-08-26 17:00:38 +02:00
|
|
|
|
2018-09-21 20:51:07 +02:00
|
|
|
NUMJOBS=${NUMJOBS:-" -j$(expr $(nproc) + 1) "}
|
2011-04-25 15:37:00 +02:00
|
|
|
|
|
|
|
# Be sure this list is up-to-date:
|
2021-08-22 20:53:28 +02:00
|
|
|
GALLIUM_DRIVERS="nouveau,r300,r600,svga,radeonsi,swrast,virgl,iris,crocus,zink"
|
2010-05-19 10:58:23 +02:00
|
|
|
|
|
|
|
if [ -z "$ARCH" ]; then
|
|
|
|
case "$( uname -m )" in
|
2016-06-30 22:26:57 +02:00
|
|
|
i?86) export ARCH=i586 ;;
|
2010-05-19 10:58:23 +02:00
|
|
|
arm*) export ARCH=arm ;;
|
|
|
|
*) export ARCH=$( uname -m ) ;;
|
|
|
|
esac
|
|
|
|
fi
|
|
|
|
|
2018-05-28 21:12:29 +02:00
|
|
|
# If the variable PRINT_PACKAGE_NAME is set, then this script will report what
|
|
|
|
# the name of the created package would be, and then exit. This information
|
|
|
|
# could be useful to other scripts.
|
|
|
|
if [ ! -z "${PRINT_PACKAGE_NAME}" ]; then
|
Tue Jul 18 19:58:10 UTC 2023
a/tar-1.35-x86_64-1.txz: Upgraded.
x/mesa-23.2.0_rc1-x86_64-1.txz: Upgraded.
OK, usually I won't use rc versions even in -current, but in this case I'm
going to. Some time ago my desktop machine with RS880 / Radeon HD 4290
graphics on the motherboard began acting up with the screen going black
for a few seconds before returning. This after an hour or so of uptime,
usually, then becoming more frequent with more uptime. Eventually I'd lose
mouse and/or keyboard too, and have to reboot. Here's a couple of errors
from dmesg:
[ 9942.677675] [drm:r600_ib_test [radeon]] *ERROR* radeon: fence wait
timed out.
[ 9942.677741] [drm:radeon_ib_ring_tests [radeon]] *ERROR* radeon: failed
testing IB on GFX ring (-110).
I also noticed that the backtrace started with ttm_bo_release, and seeing
this in recent kernel patches had been chalking this up to a kernel bug.
I *still* think it could be, and there are a bunch of kernel patches coming
soon to -stable from Alex Deucher that could address the underlying causes
(not for 6.1.39 though, unfortunately). Anyway, I'd recently figured out
that reverting Mesa sufficiently made the issue go away. And now it seems
this 23.2.0 release candidate also fixes the issue.
Yes, I could go search for the commits to cherry-pick, but we'll be moving
to mesa-23.2.0 when it's released, so we might as well start testing now.
2023-07-18 21:58:10 +02:00
|
|
|
echo "$PKGNAM-$(echo $VERSION | tr - _)-$ARCH-$BUILD.txz"
|
2018-05-28 21:12:29 +02:00
|
|
|
exit 0
|
|
|
|
fi
|
|
|
|
|
2021-10-15 22:47:13 +02:00
|
|
|
## COMMENTED OUT as swr seems to be broken with llvm13:
|
|
|
|
## Build swr for the following architectures (not for i?x6, as this will cause
|
|
|
|
## an Illegal Instruction startup failure on many otherwise supported CPUs):
|
|
|
|
#for swr_arch in x86_64 ; do
|
|
|
|
# if [ "$ARCH" = "x86_64" ]; then
|
|
|
|
# GALLIUM_DRIVERS="$GALLIUM_DRIVERS,swr"
|
|
|
|
# fi
|
|
|
|
#done
|
2019-07-01 00:45:12 +02:00
|
|
|
|
2009-08-26 17:00:38 +02:00
|
|
|
TMP=${TMP:-/tmp}
|
|
|
|
PKG=$TMP/package-mesa
|
|
|
|
|
2016-06-30 22:26:57 +02:00
|
|
|
if [ "$ARCH" = "i586" ]; then
|
|
|
|
SLKCFLAGS="-O2 -march=i586 -mtune=i686"
|
2009-08-26 17:00:38 +02:00
|
|
|
LIBDIRSUFFIX=""
|
|
|
|
elif [ "$ARCH" = "x86_64" ]; then
|
|
|
|
SLKCFLAGS="-O2 -fPIC"
|
|
|
|
LIBDIRSUFFIX="64"
|
|
|
|
else
|
|
|
|
SLKCFLAGS="-O2"
|
|
|
|
LIBDIRSUFFIX=""
|
|
|
|
fi
|
|
|
|
|
|
|
|
rm -rf $PKG
|
|
|
|
mkdir -p $TMP $PKG
|
|
|
|
cd $TMP
|
2016-06-30 22:26:57 +02:00
|
|
|
rm -rf ${PKGNAM}-${VERSION}
|
2011-04-25 15:37:00 +02:00
|
|
|
|
2020-02-15 03:42:28 +01:00
|
|
|
tar xvf $CWD/${PKGNAM}-${VERSION}.tar.?z || exit 1
|
2018-05-28 21:12:29 +02:00
|
|
|
cd ${PKGNAM}-$VERSION || exit 1
|
2009-08-26 17:00:38 +02:00
|
|
|
|
2013-11-04 18:08:47 +01:00
|
|
|
# Let's kill the warning about operating on a dangling symlink:
|
|
|
|
rm -f src/gallium/state_trackers/d3d1x/w32api
|
|
|
|
|
2009-08-26 17:00:38 +02:00
|
|
|
# Make sure ownerships and permissions are sane:
|
|
|
|
chown -R root:root .
|
|
|
|
find . \
|
|
|
|
\( -perm 777 -o -perm 775 -o -perm 711 -o -perm 555 -o -perm 511 \) \
|
2019-09-30 23:08:32 +02:00
|
|
|
-exec chmod 755 {} \+ -o \
|
2009-08-26 17:00:38 +02:00
|
|
|
\( -perm 666 -o -perm 664 -o -perm 600 -o -perm 444 -o -perm 440 -o -perm 400 \) \
|
2019-09-30 23:08:32 +02:00
|
|
|
-exec chmod 644 {} \+
|
2009-08-26 17:00:38 +02:00
|
|
|
|
2011-04-25 15:37:00 +02:00
|
|
|
# Apply patches from git (and maybe elsewhere):
|
|
|
|
# Patches obtained by:
|
2018-05-28 21:12:29 +02:00
|
|
|
# git checkout origin/17.2
|
|
|
|
# git format-patch 93c2beafc0a7fa2f210b006d22aba61caa71f773 # 17.2.6 release
|
2011-04-25 15:37:00 +02:00
|
|
|
if /bin/ls $CWD/patches/*.patch 1> /dev/null 2> /dev/null ; then
|
|
|
|
for patch in $CWD/patches/*.patch ; do
|
Fri Dec 25 00:52:08 UTC 2020
Here are some updates to provide a little holiday cheer. The kernels are
rebuilt (yeah, we'll probably have new ones tomorrow but whatever) to build in
HWMON making NVMe temperature monitoring available, and there's also a massive
cleanup of polkit/dbus related packages to move most of the config files out
of /etc and not install them as .new. Local config files that override the
shipped ones may be placed in the directories in /etc to override the ones that
we ship. Thanks to Robby Workman for this!
None of the packages included in Slackware place any files in
/etc/polkit-1/rules.d/ or /etc/dbus-1/system.d/ any more, but there may be
third-party packages that still do. However, since Slackware's packages
originally installed these as .new files, the leftover configs will remain.
You'll probably want to take a look in those directories to clean out any
config files that you haven't modified yourself or that belong to packages that
aren't part of Slackware itself.
Hope everyone has a great day. :-)
a/dbus-1.12.20-x86_64-3.txz: Rebuilt.
Added local config override directory /etc/dbus-1/system.d/ since no other
package provides it now.
a/kernel-generic-5.10.2-x86_64-2.txz: Upgraded.
a/kernel-huge-5.10.2-x86_64-2.txz: Upgraded.
a/kernel-modules-5.10.2-x86_64-2.txz: Upgraded.
a/udisks-1.0.5-x86_64-6.txz: Rebuilt.
Move D-Bus configs from /etc to the system location. Thanks to Robby Workman.
a/udisks2-2.9.1-x86_64-2.txz: Rebuilt.
Include 20-plugdev-group-mount-override.rules, installed in the system
location. Thanks to Robby Workman.
ap/cups-2.3.3-x86_64-3.txz: Rebuilt.
Move D-Bus configs from /etc to the system location. Thanks to Robby Workman.
ap/hplip-3.20.6-x86_64-4.txz: Rebuilt.
Use --enable-qt5.
Move D-Bus configs from /etc to the system location.
Drop HAL support (LOL!)
Thanks to Robby Workman.
ap/mpg123-1.26.4-x86_64-1.txz: Upgraded.
d/gnucobol-3.1.2-x86_64-1.txz: Upgraded.
d/kernel-headers-5.10.2-x86-2.txz: Upgraded.
k/kernel-source-5.10.2-noarch-2.txz: Upgraded.
HWMON m -> y
+NVME_HWMON y
+POWER_SUPPLY_HWMON y
+THERMAL_HWMON y
Thanks to Daedra.
kde/sddm-0.19.0-x86_64-4.txz: Rebuilt.
Don't include the /etc/dbus-1/system.d/ directory.
l/GConf-3.2.6-x86_64-5.txz: Rebuilt.
Move D-Bus configs from /etc to the system location. Thanks to Robby Workman.
l/accountsservice-0.6.55-x86_64-2.txz: Rebuilt.
Move D-Bus configs from /etc to the system location. Thanks to Robby Workman.
l/cryptopp-8.3.0-x86_64-1.txz: Upgraded.
l/harfbuzz-2.7.3-x86_64-1.txz: Upgraded.
l/polkit-0.118-x86_64-3.txz: Rebuilt.
Remove D-Bus rules for udisks2 and NetworkManager (moved to those packages).
Move D-Bus configs from /etc to the system location. Thanks to Robby Workman.
l/pulseaudio-14.0-x86_64-2.txz: Rebuilt.
Move D-Bus configs from /etc to the system location. Thanks to Robby Workman.
l/system-config-printer-1.5.12-x86_64-6.txz: Rebuilt.
Move D-Bus configs from /etc to the system location. Thanks to Robby Workman.
n/ModemManager-1.14.8-x86_64-2.txz: Rebuilt.
Added elogind support.
Move D-Bus configs from /etc to the system location.
Thanks to Robby Workman.
n/NetworkManager-1.28.0-x86_64-2.txz: Rebuilt.
Include 10-org.freedesktop.NetworkManager.rules, installed in the system
location. Thanks to Robby Workman.
n/bluez-5.55-x86_64-2.txz: Rebuilt.
Move D-Bus configs from /etc to the system location. Thanks to Robby Workman.
n/netatalk-3.1.12-x86_64-5.txz: Rebuilt.
Patched to fix failure to build from source with gcc10.
Move D-Bus configs from /etc to the system location.
Thanks to Robby Workman.
n/wpa_supplicant-2.9-x86_64-3.txz: Rebuilt.
Move D-Bus configs from /etc to the system location. Thanks to Robby Workman.
x/mesa-20.3.1-x86_64-2.txz: Rebuilt.
[PATCH] vulkan/device_select: Store Vulkan vendorID and deviceID as uint32_t.
Thanks to sasha69 for the link to the upstream discussion/commit.
xap/blueman-2.1.4-x86_64-3.txz: Rebuilt.
Move D-Bus configs from /etc to the system location. Thanks to Robby Workman.
xap/xgames-0.3-x86_64-6.txz: Rebuilt.
Fixed spider failing to build from source with gcc10.
Removed the ancient version of xsnow.
xap/xscreensaver-5.45-x86_64-2.txz: Rebuilt.
Make the description of the xscreensaver settings be clearly for xscreensaver
as opposed to a generic "screensaver" settings. Thanks to Robby Workman.
xap/xsnow-3.1.9-x86_64-1.txz: Added.
Merry Christmas / Happy Holidays! :-)
xfce/mousepad-0.5.1-x86_64-1.txz: Upgraded.
xfce/xfce4-screensaver-0.1.11-x86_64-2.txz: Rebuilt.
Make the description of the xfce4-screensaver settings be clearly for
xfce4-screensaver as opposed to a generic "screensaver" settings.
Thanks to Robby Workman.
xfce/xfce4-session-4.14.2-x86_64-2.txz: Rebuilt.
Use xfce4-screensaver by default and don't ship the xscreensaver.desktop
autostart file. Thanks to Robby Workman.
extra/pure-alsa-system/*: Removed.
This served a purpose during the pure ALSA to PulseAudio transition, but
it's time for it to go away. If the latency is an issue for your use case,
a look around should turn up documentation on how to reduce it, or how to
send output directly to ALSA without having to uninstall PulseAudio or
recompile anything that's linked to it. As an example, see this one (which
was mentioned here before):
https://wiki.archlinux.org/index.php/PulseAudio/Examples#PulseAudio_as_a_minimal_unintrusive_dumb_pipe_to_ALSA
isolinux/initrd.img: Rebuilt.
kernels/*: Upgraded.
usb-and-pxe-installers/usbboot.img: Rebuilt.
2020-12-25 01:52:08 +01:00
|
|
|
patch -p1 --verbose < $patch || exit 1 ;
|
2011-04-25 15:37:00 +02:00
|
|
|
done
|
|
|
|
fi
|
|
|
|
|
2021-03-12 02:46:50 +01:00
|
|
|
# Revert these patches from git (and maybe elsewhere):
|
|
|
|
if /bin/ls $CWD/patches-revert/*.patch 1> /dev/null 2> /dev/null ; then
|
|
|
|
for patch in $CWD/patches-revert/*.patch ; do
|
|
|
|
patch -p1 -R --verbose < $patch || exit 1 ;
|
|
|
|
done
|
|
|
|
fi
|
|
|
|
|
2019-06-25 20:32:03 +02:00
|
|
|
# Configure, build, and install:
|
|
|
|
export CFLAGS="$SLKCFLAGS"
|
|
|
|
export CXXFLAGS="$SLKCFLAGS"
|
|
|
|
mkdir meson-build
|
|
|
|
cd meson-build
|
|
|
|
meson setup \
|
2009-08-26 17:00:38 +02:00
|
|
|
--prefix=/usr \
|
2019-06-25 20:32:03 +02:00
|
|
|
--libdir=lib${LIBDIRSUFFIX} \
|
|
|
|
--libexecdir=/usr/libexec \
|
|
|
|
--bindir=/usr/bin \
|
|
|
|
--sbindir=/usr/sbin \
|
|
|
|
--includedir=/usr/include \
|
|
|
|
--datadir=/usr/share \
|
2009-08-26 17:00:38 +02:00
|
|
|
--mandir=/usr/man \
|
2019-06-25 20:32:03 +02:00
|
|
|
--sysconfdir=/etc \
|
|
|
|
--localstatedir=/var \
|
|
|
|
--buildtype=release \
|
2020-12-04 00:52:38 +01:00
|
|
|
-Dplatforms=x11,wayland \
|
2019-06-25 20:32:03 +02:00
|
|
|
-Dgallium-opencl=icd \
|
|
|
|
-Dgallium-nine=true \
|
2022-12-07 19:48:07 +01:00
|
|
|
-Dgallium-rusticl=true \
|
|
|
|
-Drust_std=2021 \
|
2021-03-12 02:46:50 +01:00
|
|
|
-Dosmesa=true \
|
2019-06-25 20:32:03 +02:00
|
|
|
-Dgallium-drivers=$GALLIUM_DRIVERS \
|
2023-09-22 23:51:12 +02:00
|
|
|
-Dvulkan-drivers=amd,intel,intel_hasvk,swrast \
|
2021-05-12 21:02:20 +02:00
|
|
|
-Dvulkan-layers=device-select,intel-nullhw,overlay \
|
2022-09-30 19:52:21 +02:00
|
|
|
-Dvideo-codecs="$CODECS" \
|
2019-06-25 20:32:03 +02:00
|
|
|
-Dglvnd=true \
|
2020-12-04 00:52:38 +01:00
|
|
|
-Dllvm=enabled \
|
|
|
|
-Dshared-llvm=enabled \
|
|
|
|
-Dshared-glapi=enabled \
|
|
|
|
-Degl=enabled \
|
|
|
|
-Dgles1=enabled \
|
|
|
|
-Dgles2=enabled \
|
2021-01-14 21:36:17 +01:00
|
|
|
-Dopengl=true \
|
|
|
|
-Dglx=dri \
|
2019-06-25 20:32:03 +02:00
|
|
|
.. || exit 1
|
2019-07-17 23:57:16 +02:00
|
|
|
"${NINJA:=ninja}" $NUMJOBS || exit 1
|
|
|
|
DESTDIR=$PKG $NINJA install || exit 1
|
2019-06-25 20:32:03 +02:00
|
|
|
cd ..
|
2018-05-28 21:12:29 +02:00
|
|
|
|
|
|
|
# Install /etc/drirc as a .new file:
|
2018-12-08 05:10:45 +01:00
|
|
|
if [ -r $PKG/etc/drirc ]; then
|
|
|
|
mv $PKG/etc/drirc $PKG/etc/drirc.new
|
|
|
|
fi
|
2018-05-28 21:12:29 +02:00
|
|
|
|
2021-01-14 21:36:17 +01:00
|
|
|
# Add a default provider for glvnd when the vendor cannot be determined:
|
|
|
|
( cd $PKG/usr/lib${LIBDIRSUFFIX}
|
|
|
|
if [ ! -r libGLX_system.so.0 ]; then
|
|
|
|
ln -sf libGLX_mesa.so.0 libGLX_system.so.0
|
|
|
|
fi
|
|
|
|
)
|
|
|
|
|
2022-08-09 01:29:31 +02:00
|
|
|
if [ "$BUILD_AMBER" = "YES" ]; then
|
|
|
|
. $CWD/mesa-amber.build
|
|
|
|
fi
|
|
|
|
|
2019-10-09 22:33:48 +02:00
|
|
|
if [ "$BUILD_DEMOS" = "YES" ]; then
|
2022-08-09 01:29:31 +02:00
|
|
|
. $CWD/mesa-demos.build
|
2019-10-09 22:33:48 +02:00
|
|
|
fi
|
2016-06-30 22:26:57 +02:00
|
|
|
|
2009-08-26 17:00:38 +02:00
|
|
|
# Strip binaries:
|
|
|
|
find $PKG | xargs file | grep -e "executable" -e "shared object" | grep ELF \
|
|
|
|
| cut -f 1 -d : | xargs strip --strip-unneeded 2> /dev/null
|
|
|
|
|
2022-08-09 01:29:31 +02:00
|
|
|
# Compress man pages, if any:
|
|
|
|
if [ -d $PKG/usr/man ]; then
|
|
|
|
find $PKG/usr/man -type f -exec gzip -9 {} \+
|
|
|
|
for i in $( find $PKG/usr/man -type l ) ; do ln -s $( readlink $i ).gz $i.gz ; rm $i ; done
|
|
|
|
fi
|
2009-08-26 17:00:38 +02:00
|
|
|
|
|
|
|
# Compress info files, if any:
|
|
|
|
if [ -d $PKG/usr/info ]; then
|
2011-04-25 15:37:00 +02:00
|
|
|
rm -f $PKG/usr/info/dir
|
|
|
|
gzip -9 $PKG/usr/info/*
|
2009-08-26 17:00:38 +02:00
|
|
|
fi
|
|
|
|
|
2022-08-09 01:29:31 +02:00
|
|
|
mkdir -p $PKG/usr/doc/$PKGNAM-$VERSION
|
2011-04-25 15:37:00 +02:00
|
|
|
cp -a \
|
2022-08-09 01:29:31 +02:00
|
|
|
CODEOWNERS* README* docs/README* docs/license* docs/relnotes/${VERSION}.* \
|
2016-06-30 22:26:57 +02:00
|
|
|
$PKG/usr/doc/$PKGNAM-$VERSION
|
2009-08-26 17:00:38 +02:00
|
|
|
|
|
|
|
mkdir -p $PKG/install
|
|
|
|
cat $CWD/slack-desc > $PKG/install/slack-desc
|
2018-05-28 21:12:29 +02:00
|
|
|
zcat $CWD/doinst.sh.gz > $PKG/install/doinst.sh
|
2009-08-26 17:00:38 +02:00
|
|
|
|
|
|
|
cd $PKG
|
Tue Jul 18 19:58:10 UTC 2023
a/tar-1.35-x86_64-1.txz: Upgraded.
x/mesa-23.2.0_rc1-x86_64-1.txz: Upgraded.
OK, usually I won't use rc versions even in -current, but in this case I'm
going to. Some time ago my desktop machine with RS880 / Radeon HD 4290
graphics on the motherboard began acting up with the screen going black
for a few seconds before returning. This after an hour or so of uptime,
usually, then becoming more frequent with more uptime. Eventually I'd lose
mouse and/or keyboard too, and have to reboot. Here's a couple of errors
from dmesg:
[ 9942.677675] [drm:r600_ib_test [radeon]] *ERROR* radeon: fence wait
timed out.
[ 9942.677741] [drm:radeon_ib_ring_tests [radeon]] *ERROR* radeon: failed
testing IB on GFX ring (-110).
I also noticed that the backtrace started with ttm_bo_release, and seeing
this in recent kernel patches had been chalking this up to a kernel bug.
I *still* think it could be, and there are a bunch of kernel patches coming
soon to -stable from Alex Deucher that could address the underlying causes
(not for 6.1.39 though, unfortunately). Anyway, I'd recently figured out
that reverting Mesa sufficiently made the issue go away. And now it seems
this 23.2.0 release candidate also fixes the issue.
Yes, I could go search for the commits to cherry-pick, but we'll be moving
to mesa-23.2.0 when it's released, so we might as well start testing now.
2023-07-18 21:58:10 +02:00
|
|
|
/sbin/makepkg -l y -c n $TMP/${PKGNAM}-$(echo $VERSION | tr - _)-$ARCH-$BUILD.txz
|