The Makefile would still force the use of and linking against uClibc++
even though libstdc++ may have been chosen, which would result in the
package depending on libstdcpp ($(CXX_DEPENDS) but we would still be
missing an libuClibc++ library depdency.
Fix this by looking at CONFIG_USE_UCLIBCXX to adjust the configure
script variables.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Passing GNU_TARGET_NAME implies that we have a 4-group triplet (e.g:
arm-openwrt-unknown-linux), and this works fine with pciutils'
lib/configure operating system detection. If we have an external
toolchain (e.g: mipsel-linux-gnu), the configure script is all confused
and ends-up picking up i386-ports, which is not valid on the target
architecture.
Finally, we are not passing TARGET_CPPFLAGS as we should, so let's do
that correctly.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
We would first not be able to properly detect libxml2 because it's not
automatically found, fix that by explicitly linking against it (and zlib
a dependency). Then if we don't don't define X_OPEN_SOURCE=500, we would
run into the following build error:
clish/shell/shell_execute.c: In function 'clish_shell_exec_action':
clish/shell/shell_execute.c:309:18: error: 'sigignore' undeclared (first
use in this function)
sa.sa_handler = sigignore; /* Empty signal handler */
^
clish/shell/shell_execute.c:309:18: note: each undeclared identifier is
reported only once for each function it appears in
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Update tmux to 2.5
Drop patch
Update homepage URL
Use toolchain functionality for configuring and compiling
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
Fixes CVE-2017-6318
Patches where refreshed and the following removed:
- 020-fix_pieusb.patch: fixed upstream
- 030-musl.patch: accepted upstream
PS: there was no sane-backends 1.0.26
Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
Mitigate stack exhaustion when printing a very deeply nested term.
Fix heap buffer overflow in tokenadd()
Signed-off-by: Marko Ratkaj <marko.ratkaj@sartura.hr>
There is a gcc related bug that causes the following issue on MIPS:
Assertion failed: jv_get_kind(a) == JV_KIND_STRING (jv.c: jvp_string_ptr: 435)
This patch will disable SRA optimizations on MIPS platform and prevent
the above issue.
Signed-off-by: Marko Ratkaj <marko.ratkaj@sartura.hr>
This removes unnecessary dependancy on xmlto which can be problematic on macOS build environment.
Signed-off-by: David Thornley <david.thornley@touchstargroup.com>
Changes to NTFS-3G:
Delegated processing of special reparse points to external plugins
Allowed kernel cacheing by lowntfs-3g when not using Posix ACLs
Enabled fallback to read-only mount when the volume is hibernated
Made a full check for whether an extended attribute is allowed
Moved secaudit and usermap to ntfsprogs (now ntfssecaudit and ntfsusermap)
Enabled encoding broken UTF-16 into broken UTF-8
Autoconfigured selecting <sys/sysmacros.h> vs <sys/mkdev>
Allowed using the full library API on systems without extended attributes support
Fixed DISABLE_PLUGINS as the condition for not using plugins
Corrected validation of multi sector transfer protected records
Denied creating/removing files from $Extend
Returned the size of locale encoded target as the size of symlinks
http://www.tuxera.com/community/release-history/
Signed-off-by: BangLang Huang <banglang.huang@foxmail.com>
Make ntfs-3g-utils into a single package due to upstream package changes.
Signed-off-by: Ted Hess <thess@kitschensync.net>
Enable collectd-mod-cpufreq also for armvirt to make it
available for ipq806x devices in LEDE buildbot snapshots.
LEDE phase2 buildbot for arm_cortex-a15_neon-vfpv4 that is
the package arch for ipq806x, uses armvirt SDK instead of
ipq806x SDK.
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
Longest package titles are not shown in the menuconfig,
which decreases manuconfig usability.
Shorten some of the longest titles.
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
As lcd4linux is all kinds of BROKEN lately, and of unknown status.
Signed-off-by: Philip Prindeville <philipp@redfish-solutions.com>
Reviewed-by: Harald Geyer <harald@ccbib.org>
If you have a firstboot script which seeds a passwd based on
run-time information (like MAC addresses, hostname, etc) then
you need to be able to pass in a cleartext string via chpasswd.
Other applets are similarly potentially useful in other corner
cases.
Signed-off-by: Philip Prindeville <philipp@redfish-solutions.com>
Several problems were found with building against an external toolchain:
* TARGET_CPPFLAGS was missing which would prevent finding the headers
* TARGET_LDFLAGS also was not passed correctly
* when building against glibc, since we use strlcpy() we need to link
against libbsd
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
External toolchains don't automatically have STAGING_DIR in their
default search path, so make sure we pass TARGET_LDFLAGS down. Since the
unrar makefile does not allow overriding, we also need to preserve the
original LDFLAGS which include lpthread.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Instead of assigning various variables ourself, rely on the common logic
provided by "packaging framework". This brings this package in sync with
other ones which also use git repos as source, dates and commit ids as
version etc. This results also in using xz for tarballs. While at,
populate the PKG_MIRROR_HASH.
Thanks @diizzyy for suggesting these improvements.
Signed-off-by: Michael Heimpold <mhei@heimpold.de>
cryptodev log by default unecessary debug message
With some app (like aria2) the syslog get spammed with lots of this message
cryptodev: aria2c[3231] (adjust_sg_array:106): reallocating from 32 to 512 pages
With this we disable logging of debug message as they are just for info purpose and they are not error at all.
Signed-off-by: Ansuel Smith ansuelsmth@gmail.com
Update to 4.1
Remove 0001-mkfs-Default-to-64-32-heads-sectors-for-targets-smal.patch
as it's included in upstream repo.
Refresh patch
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
For consistency, use full name instead of $(PKG_NAME) in define and eval
lines for all packages.
I've seen reviews that asked to do this before, and I am asking the same
during reviews now. To avoid this in the future, fix this treewide so
when people use existing packages as example, we will not have to
request this change anymore.
This makes all packages consistent with both LEDE and OpenWrt base
repositories.
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
aTanW
Florian Fainelli
Ansuel Smith
Daniel Golle
Hauke Mehrtens
Daniel Engberg
Luiz Angelo Daros de Luca
Daniel Albers
Hannu Nyman
Marko Ratkaj
Christian Simon
David Thornley
Ted Hess
Matthew McClintock
Shuanglei Tao
Philip Prindeville
Nikil Mehta
Michael Heimpold
Vladimir Ulrich
Yegor Yefremov
Bruno Randolf
Stijn Tintel
Günther Kelleter
Dirk Brenken