| Age | Commit message (Collapse) | Author |
|---|
|
OpenTTD doesn't depend on libicudata:
$ readelf -d /usr/bin/openttd | grep "Shared library:" | awk '{print $5}'
[libpthread.so.0]
[libSDL-1.2.so.0]
[libz.so.1]
[liblzma.so.5]
[liblzo2.so.2]
[libpng16.so.16]
[libfontconfig.so.1]
[libfreetype.so.6]
[libicui18n.so.67]
[libicuuc.so.67]
[libfluidsynth.so.2]
[libstdc++.so.6]
[libm.so.6]
[libc.so.6]
[libgcc_s.so.1]
but libicuuc does depend on libicudata:
$ readelf -d /usr/lib/libicuuc.so.67 | grep "Shared library:" | awk '{print $5}'
[libicudata.so.67]
[libpthread.so.0]
[libdl.so.2]
[libstdc++.so.6]
[libm.so.6]
[libgcc_s.so.1]
[libc.so.6]
[ld-linux-x86-64.so.2]
The ldd utility doesn't show this level of details.
The lddtree utility which is part of the pax-utils package can show a tree of
dependencies, but on x86_64 it didn't work for me due to some python issue:
# lddtree
Traceback (most recent call last):
File "/usr/bin/lddtree", line 53, in <module>
from elftools.elf.elffile import ELFFile
ImportError: No module named elftools.elf.elffile
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
The Arch Linux icu PKGBUILD has now the following:
provides=(libicu{data,i18n,io,test,tu,uc}.so)
and the resulting packages are already in Parabola.
Thanks to that, adding such libraries as a dependency will
make sure that the package depends on the precise version
of the library like libicui18n.so=67-64.
For some reason even if libicudata.so.67 is used by the
OpenTTD binary:
$ ldd /usr/bin/openttd | grep libicudata.so
libicudata.so.67 => /usr/lib/libicudata.so.67
When building the package, it didn't add manage to find the
automatic libicudata.so dependency:
==> Checking for packaging issues...
==> WARNING: Package contains reference to $srcdir
usr/bin/openttd
==> Creating package "openttd"...
-> Generating .PKGINFO file...
==> WARNING: Library listed in 'depends' is not required by any files: libicudata.so
-> Generating .BUILDINFO file...
-> Generating .MTREE file...
-> Compressing package...
This also shows up in the package:
$ pacman -Q -i openttd
Depends On: [...] libicui18n.so=67-64 libicuuc.so=67-64 libicudata.so
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
|
|
|
|
|
|
Following the same pattern as some AUR packages (icu65 and so on), make
this package just provide compat libs. There's no need to try building
packages against an old icu version, we just want to make sure programs
which are linked to an older libicu*.so can use it.
The libprovides recently added in Arch Linux are introduced as well.
They won't break anything, they simply add another way to depend on this
package, which is more convenient than the current use in package()
functions of:
local _icu_ver
_icu_ver=$(pacman -S --print-format='%v' icu)
depends+=("icu>=${_icu_ver}" "icu<$((${_icu_ver%%.*} + 1))")
(which has other problems, like not being the version which is built
against, but rather the version that a pacman -S icu would install or
upgrade to -- partial upgrades are bad, but that doesn't mean break it
if someone does so)
Expected use of this package:
depends+=('icu' 'libicui18n.so')
... added to a Parabola package which uses icu, to make it depend on the
Arch icu package (and ensure it is preferred in all cases) and
additionally to depend on any package providing the correct version of
the libicu*.so dependent libraries. If all goes well, the icu package
will be pulled in as a dependency. If all doesn't go well, then packages
fall back on additionally installing this package for compatibility
purposes, and all still goes well.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
|
|
The .install was only added to run ldconfig, but since
ldconfig is already run by pacman, it is unnecessary.
Reported-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
|
|
|
|
Bump version
Add the rest of the logic for ARM builds
- ARM builds with error: semop(1): encountered an error: Function not implemented
Possibly unrelated to the PKGBUILD.
|
|
|
|
|
|
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
texlive-bin 2020.54586-3 is already in archlinux testing, but for
now we only sync with extra.
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
|
|
|
|
|
|
|
|
|
|
Reason: Linux-libre{-hardened} enables all LSM by default in i686 and x86_64 since a long time.
Although it kinda made sense because it had AppArmor enabled by default, it's not worth to maintain
it since anyone can enable it by passing a kernel parameter, as explained in the AppArmor page in
ArchWiki. This isn't the case for armv7h, but will see if users want AppArmor for ARM in the future.
Signed-off-by: David P <megver83@parabola.nu>
|
|
Signed-off-by: David P <megver83@parabola.nu>
|
|
Signed-off-by: David P <megver83@parabola.nu>
|
|
Signed-off-by: David P <megver83@parabola.nu>
|
|
Signed-off-by: David P <megver83@parabola.nu>
|
|
|
|
|
|
|
|
|
|
Signed-off-by: David P <megver83@parabola.nu>
|
|
this is due to rcn patch, which introduces a "backport" from linux 5.4.18
Signed-off-by: David P <megver83@parabola.nu>
|
|
Signed-off-by: David P <megver83@parabola.nu>
|
|
linux-libre 5.6+ replaces it
Signed-off-by: David P <megver83@parabola.nu>
|
|
On armv7h we have:
$ mplayer
mplayer: error while loading shared libraries:
libdvdread.so.7: cannot open shared object file: No
such file or directory
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
|
|
|
|
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
|
|
|
|
|
|
|
|
Signed-off-by: David P <megver83@parabola.nu>
|
|
Signed-off-by: David P <megver83@parabola.nu>
|
|
Signed-off-by: David P <megver83@parabola.nu>
|
|
|
|
|
