The fltk build system has some logic that causes it to pass
-Wl,-rpath,/usr/lib when --libdir is not /usr/lib. However, in our
case, libdir is ${exec_prefix}/lib, and is not expanded to /usr/lib
before the rpath related test is done. Rather than fixing the fltk
build system, this commit works around the problem by explicitly
passing --libdir=/usr/lib.
Fixes:
http://autobuild.buildroot.org/results/8d1/8d1b202a182e3fb5dee21f20afc9f749c2defa1a/
and many other similar build failures that have been occuring since 1+
year.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
External toolchain can also have been generated by Buildroot previously, as
the list that follows demonstrates. Rephrase the paragraph describing what an
external toolchain is as suggested by Thomas Petazzoni, to make it clearer.
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The central package infrastructure handles all packages for quite some time
now, so this note is no longer needed.
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
NETSNMP_CAN_USE_SYSCTL is for BSD systems, setting it to yes on
Linux systems breaks compilation when the mibII mib is enabled.
Prior to commit 30bb1bdca4 ("netsnmp:
bump version") from November 2008, this variable was already set to
no, and was changed to yes by the commit, with no explanation. And the
previous code indeed had a comment saying:
# We set CAN_USE_SYSCTL to no and use /proc since the sysctl code
# in this thing is apparently intended for freebsd or some such
# thing...
[Thomas: improved commit log.]
Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fixes:
CVE-2015-0273 - Use after free vulnerability in unserialize() with
DateTimeZone.
CVE-2015-0235 - Mitigation for GHOST: glibc gethostbyname buffer
overflow.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fixes CVE-2015-1349 - Revoking a managed trust anchor and supplying an
untrusted replacement could cause namedto crash with an assertion
failure.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Some Erlang packages provide a bin directory for programs that are to be
invoked from the command line. An example of such a package is the Lisp
Flavored Erlang compiler. The Erlang OTP library includes several more
examples (it doesn't use rebar, though.) This change makes sure that the
bin directory gets installed too.
Signed-off-by: Frank Hunleth <fhunleth@troodon-software.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Disable the 'load' operation for static builds since it needs dlopen and
friends. Otherwise it will fail with an error message like this one:
load.o: In function `load_file':
load.c:(.text+0x2a8): undefined reference to `dlopen'
load.c:(.text+0x2c4): undefined reference to `dlerror'
load.c:(.text+0x308): undefined reference to `dlsym'
load.c:(.text+0x33c): undefined reference to `dlopen'
load.c:(.text+0x35c): undefined reference to `dlsym'
load.c:(.text+0x380): undefined reference to `dlopen'
load.c:(.text+0x39c): undefined reference to `dlopen'
load.c:(.text+0x3a0): undefined reference to `dlerror'
load.c:(.text+0x42c): undefined reference to `dlsym'
load.c:(.text+0x470): undefined reference to `dlsym'
load.c:(.text+0x48c): undefined reference to `dlerror'
load.o: In function `unload_file':
load.c:(.text+0x63c): undefined reference to `dlclose'
collect2: error: ld returned 1 exit status
Fixes:
http://autobuild.buildroot.net/results/9e3/9e39039b6db79a46990cd9bdcb179289e38d9f31/
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Don't use anymore the alternate download site beacause it does not
conatins anymore older versions, instead use the official download site
because now it contains even the older versions.
Signed-off-by: Fabio Porcedda <Fabio.Porcedda@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fixes CVE-2014-9680 - A user with sudo access may be able to exploit
parsing bugs in the time zone parsing functions of the system's C
library functions. The user may also be able to read arbitrary files,
potentially causing changes in system behavior when reading certain
device special files or simply causing the program run via sudo to
block.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
mplayer fails to compile with the following error message:
libmpdemux/demux_gif.c: In function 'demux_open_gif':
libmpdemux/demux_gif.c:260:3: error: too few arguments to function
'DGifOpen'
gif = DGifOpen(demuxer->stream, my_read_gif);
Backport an upstream patch to support newer versions of libgif in
mplayer. Unfortunately this patch is incomplete and mplayer stills
failing to compile with a new error message:
libvo/vo_gif89a.c: In function 'uninit':
libvo/vo_gif89a.c:374:3: error: too few arguments to function
'EGifCloseFile'
EGifCloseFile(new_gif); // also frees gif storage space.
So I have written a new patch and submitted it upstream to finally fix
the problem.
Upstream commit:
a0ddaef545
New submitted patch:
https://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/2015-February/072848.html
Fixes:
http://autobuild.buildroot.net/results/a51/a510a0ab2cb827bb91b4fdec43055f2bfda239b1/
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The sha1 used for the DT enabled kernel is no longer available
(presumably, the rpi-3.18.y branch was rebased recently.) This updates
it to the lastest sha1 in the rpi-3.18.y branch.
Signed-off-by: Frank Hunleth <fhunleth@troodon-software.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Add a patch to fix qt5connectivity for big endian platforms.
Building qtconnectivity fails for big endian platforms because the
bswap_16 function is not declared. This is the error message:
In file included from bluez/hcimanager_p.h:52:0,
from bluez/hcimanager.cpp:35:
./bluez/bluez_data_p.h: In function 'quint16 bt_get_le16(const void*)':
./bluez/bluez_data_p.h:172:60: error: 'bswap_16' was not declared in
this scope
return bswap_16(bt_get_unaligned((const quint16 *) ptr));
bswap_16 is defined in byteswap.h so we can include this file in order
to fix this problem.
This patch has been submitted upstream:
https://bugreports.qt.io/browse/QTBUG-44421
Fixes:
http://autobuild.buildroot.net/results/5b8/5b85c6819f94988abd8abfcdaad6226ceb2d790a/
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fixes:
CVE-2014-9297 - vallen is not validated in several places in ntp_crypto.c,
leading to a potential information leak or possibly a crash
CVE-2014-9298 - ::1 can be spoofed on some OSes (including "some versions" of
Linux), so ACLs based on IPv6 ::1 addresses can be bypassed
Drop a patch applied upstream, along with its accompanied AUTORECONF.
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
