The test for clock_gettime() in configure.in doesn't work properly
when a previous package has loaded the shared configuration cache with
informations about the availability of clock_gettime. A package such
as ctorrent does so, which means that compiling ctorrent *then*
libglib2 currently fails.
According to people on the Autoconf mailing list, the libglib2 test is
likely the one that needs to be fixed. The problem is that the
AC_CHECK_FUNCS() test assumes that if it finds clock_gettime() it
means that there's no need to add any -lrt flag to the
build. Unfortunately, due to the shared configuration cache, this test
is already done with -lrt, so the test succeeds, and libglib2 does not
know that it needs to add -lrt to G_THREAD_LIBS and
G_THREAD_LIBS_FOR_GTHREAD.
So instead, we remplace the test with an AC_SEARCH_LIBS() test,
followed by a test on the result of this AC_SEARCH_LIBS() test to add
the necessary -lrt to G_THREAD_LIBS and
G_THREAD_LIBS_FOR_GTHREAD. Therefore, even if the result for the
AC_SEARCH_LIBS() test is cached due to the prior execution ctorrent
./configure script, libglib2 ./configure will properly add -lrt to the
appropriate variables.
Obviously, as this patch modifies the configure.in file, we enable the
autoreconf step for the libglib2 package.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
For some reason, the imagemagick Buildroot .mk file creates a
"datefile" file in the Buildroot source directory, probably an ancient
debugging thing that has been left here for no reason. Let's get rid
of it.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The problem fixed by 60f945e47a is in
fact not limited to the AVR32 architecture, as reported by Will Newton
on the list. The issue is the combination uClibc 0.9.31 with gcc 4.2,
C++ support and locales.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The CRIS architecture support in Buildroot hasn't been updated since a
long time. Even a toolchain with recent kernel headers does not build
due to missing patches.
Moreover, the CRIS architecture has been discontinued by Axis, as
visible at http://www.axis.com/products/dev/index.htm. We will remove
it from Buildroot at the next release cycle.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Check in toolchain/dependencies/dependencies.sh if an UTF-8 locale is
properly present on the system before trying to build a locale enabled
toolchain. As this test is only needed when a locale enabled toolchain
is going to be built, we pass the configuration file path to the
dependencies.sh script so that it can grep for the current value of
various options.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When no UTF-8 locale is available on the host system, uClibc can't
generate some stuff it needs to compile a C library with locale
support. Unfortunately, as gen_wc8bit message is shown on stdout and
the stdout of gen_wc8bit is redirected to a file, the user don't see
anything, as reported at
http://lists.busybox.net/pipermail/buildroot/2010-May/034177.html.
Those two patches fix the problem for uClibc 0.9.31 and 0.9.30.3. It
has been submitted upstream:
http://lists.uclibc.org/pipermail/uclibc/2010-August/044256.html
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
util-linux can build without ncurses, but when ncurses is available,
additional features can be built (such as the more
command). Therefore, in util-linux.mk, when ncurses is available, mark
it as a dependency.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
As reported in bug #635, util-linux doesn't build due to missing
constant definitions related to the a.out binary format. We fix this
by hardcoding these constant definitions, as done in newer versions of
util-linux.
Obviously, the long term fix is to upgrade to util-linux-ng, but this
is probably not acceptable for 2010.08.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
* Remove $(TARGET_DIR)/usr/share/aclocal from target-finalize when not
installing devfiles and
* Remove some (now) redundant cleanup from individual packages
Signed-off-by: Malte Starostik <m-starostik@versanet.de>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
* Don't install libglib2 development binaries and to target unless
BR2_HAVE_DEVFILES is set
Signed-off-by: Malte Starostik <m-starostik@versanet.de>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The <tuple>/lib* symlinking added by 3c77bab2ee needs to
be moved up to the gcc-intermediate step now the NPTL stuff is merged,
otherwise 64bit builds fails (lib64 already created).
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Since 5575d205c (toolchain: remove multilib) we were no longer passing
--disable-multilib, which broke builds for multilib-capable archs (like
x86-64, ppc, ..).
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The iconv library can only be present when locale are disabled in the
toolchain. When locale are enabled in the toolchain, iconv is directly
implemented by the C library.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The ./configure script of libcurl includes <arpa_inet.h> instead of
<arpa/inet.h> when testing for inet_pton(). The test fails, but it
doesn't prevent libcurl to build as it can work without inet_pton().
However, it fills the configure cache with the fact that inet_pton()
does not exist. And later, tcpreplay reads this from the configure
cache and fails to build, because tcpreplay really need inet_pton().
Unfortunately, just fixing the .m4 file doesn't work because the
autoreconfiguration of the package fails. Since the fix for this
problem is already upstream, the easiest solution is therefore to bump
libcurl.
The libcurl-7.19.2-fix-ssl-no-verbose.patch patch is no longer needed.
Since we're patching a m4 file, we must autoreconfigure the package.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
It fails to build with:
ctype_members.cc: In constructor 'std::ctype_byname<_CharT>::ctype_byname(const char*, size_t) [with _CharT = char]':
ctype_members.cc:59: error: invalid use of incomplete type 'struct __uclibc_locale_struct'
/home/test/avr32-br/usr/avr32-unknown-linux-uclibc/sys-include/bits/uClibc_locale.h:85: error: forward declaration of 'struct __uclibc_locale_struct'
ctype_members.cc:60: error: invalid use of incomplete type 'struct __uclibc_locale_struct'
/home/test/avr32-br/usr/avr32-unknown-linux-uclibc/sys-include/bits/uClibc_locale.h:85: error: forward declaration of 'struct __uclibc_locale_struct'
ctype_members.cc:61: error: invalid use of incomplete type 'struct __uclibc_locale_struct'
/home/test/avr32-br/usr/avr32-unknown-linux-uclibc/sys-include/bits/uClibc_locale.h:85: error: forward declaration of 'struct __uclibc_locale_struct'
make[5]: *** [ctype_members.lo] Error 1
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
As soon as PostScript, PNG or SVG support is enabled, PDF support is
required for Cairo to build properly. Otherwise, you get build
failures such as:
.libs/cairo-type3-glyph-surface.o: In function `_cairo_type3_glyph_surface_set_stream':
/home/thomas/local/buildroot-dl/cairo-1.8.10/src/cairo-type3-glyph-surface.c:337: undefined reference to `_cairo_pdf_operators_set_stream'
/home/thomas/local/buildroot-dl/cairo-1.8.10/src/cairo-type3-glyph-surface.c:337: undefined reference to `_cairo_pdf_operators_set_stream'
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Just as we do for U-Boot, error out in the Linux kernel makefile when
the defconfig name or the configuration file path are not
correct. What prompted me to implement this was a report on IRC from
an user using BR 2010.05 and not understand why the kernel build
process was failing. It was because he just forgot to set the path of
the configuration file.
Of course, it doesn't catch all mistakes (like pointing to a
non-existing defconfig or to a non-existing configuration file), but
it at least catches basic mistakes.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When U-Boot is enabled and no custom patch directory has been set,
then the current test:
ifneq ($(strip $(BR2_TARGET_UBOOT_CUSTOM_PATCH_DIR)),"")
works. However, when U-Boot is not enabled, but still gets compiled
because mkimage is needed to build the kernel,
BR2_TARGET_UBOOT_CUSTOM_PATCH_DIR is completely empty. It does not
even have quotes. So the test in fact needs to be:
ifneq ($(call qstrip,$(BR2_TARGET_UBOOT_CUSTOM_PATCH_DIR)),)
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When SWAT (the Web administration tool of Samba) is enabled, which is
the default when one enables samba in Buildroot, a lot of
documentation gets installed in /usr/swat (~15 MB). This patch fixes
this by removing the documentation when BR2_HAVE_DOCUMENTATION is not
set.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The upstream version of speech-tools does not build with GCC >= 4.3,
mainly due to changes in how C++ headers are included. This is fixed
in Debian, so let's use the Debian version and patches.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This is needed to avoid:
/home/test/mips-4.4/bin/mips-linux-gnu-ld --sysroot=/home/test/outputs/test-35/staging -shared --whole-archive -soname libdmallocxx.so -o libdmallocxx.so.t libdmallocxx.a
/home/test/mips-4.4/bin/mips-linux-gnu-ld: libdmalloc.a(arg_check.o): relocation R_MIPS_HI16 against `_dmalloc_flags' can not be used when making a shared object; recompile with -fPIC
It is fixed through a patch to Makefile.in instead of passing a CFLAGS
variable to ./configure environment in order to avoid cluttering the
configuration cache with incorrect values.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Also make the cpu counting routine more reliable (for ARM it's
"Processor" in cpuinfo rather than "processor").
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
This patch modifies current toolchain build sequence so that
NPTL enabled toolchain can be built. The new sequence works
well with linuxthreads as well.
It introduces a new pass for gcc cross compilation. The new
sequence is binutils->gcc-initial->linux-headers -> uclibc-configured
(some cheats to generate phony shared libc.so and libm.o)
-> gcc-intermediate(with shared lib support) -> uclibc -> gcc-final
I also added a new sample config arm_nptl_toolchain_defconfig which
builds the toolchain and busybox.
I have only tried it on arm. However it should work for other
architectures which support NPTL on uclibc e.g. mips, sh, x86, ppc, x86_64
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Since the reorganization of the variables in package/Makefile.in,
TARGET_CC and TARGET_CXX now directly contain the --sysroot= option in
addition to the compiler path. This is due to some ./configure scripts
using just $(TARGET_CC) for some tests instead of $(TARGET_CC)
$(TARGET_CFLAGS).
However, in the case of CMake, this fails as CMake really only wants
the path of the compiler in its CMAKE_C_COMPILER and
CMAKE_CXX_COMPILER variables. So here, we recompute proper values for
CMake by removing the --sysroot option from the compiler variables and
re-adding it to the flags variables.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When libeXosip fails to use pkg-config to find libosip, it defaults to
thinking that libosip is installed in $(prefix)/lib and
$(prefix)/include, which is of course wrong. There was an attempt to
fix this by passing OSIP_CFLAGS and OSIP_LIBS variables to libeXosip
./configure script, but it still does not work:
checking pkg-config is at least version 0.9.0... ./configure: line 21035: /home/test/outputs/test-41/host/usr/bin/pkg-config: No such file or directory
no
checking for OSIP... configure: WARNING: assuming osip can be found in -I${prefix}/include and -L${exec_prefix}/lib
Therefore, the correct fix is to depend on host-pkg-config.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This error should never show up if all Buildroot dependencies are
correct. However, rather than failing horribly later on, catch this
particular case early on and error out.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
All "select BR2_PACKAGE_LIBICONV" must use the "if !BR2_ENABLE_LOCALE"
condition, otherwise we can end up with a toolchain suppoting locales
*and* the libiconv package being compiled, which confuses other
packages. Example with glib:
gconvert.c:52:2: error: #error GNU libiconv in use but included
iconv.h not from libiconv
In addition to that, in xerces.mk, we add the dependency on libiconv
when it is available, to make sure it gets compiled before xerces.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The cross-gdb is supposed to be part of the external toolchain, so
Buildroot does not need to build it. Moreover, GDB_HOST build
currently fail with:
ln -snf ../../bin/arm-unknown-linux-gnueabi-gdb \
/home/test/outputs/test-48/staging/usr/arm-unknown-linux-gnueabi/bin/gdb
ln: creating symbolic link `/home/test/outputs/test-48/staging/usr/arm-unknown-linux-gnueabi/bin/gdb': No such file or directory
And even worse: they overwrite the cross-gdb of the external
toolchain!
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The current computation of REAL_GNU_TARGET_NAME is incorrect for
non-ARM glibc platforms because it generates something such as
mipsel-unknown-linux- as the REAL_GNU_TARGET_NAME.
So we correct this by :
* Adding "gnu" in the suffix when glibc is used, so that in the
previous case we will have mipsel-unknown-linux-gnu
* Improving the ARM_EABI code to correctly append "eabi" when glibc
is selected, so that we have arm-unknown-linux-gnueabi, and to
append "gnueabi" when uclibc is selected, so that we have
arm-unknown-linux-uclibcgnueabi. The little trick here is that LIBC
and ABI aren't completely orthogonal on ARM.
This fixes problems such as :
checking host system type... Invalid configuration
`mipsel-unknown-linux-': machine `mipsel-unknown-linux' not recognized
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
oprofile depends on binutils_target, but binutils_target fails to
build with external toolchains because the binutils version has not
been choosen. As the fix is not trivial, let's just disable oprofile
in external toolchain builds for the moment.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Now that two packages (tremor and libsvgtiny) are being downloaded
from svn, svn becomes a mandatory tool to run Buildroot.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The netsnmp package should depend on openssl when using it.
Otherwise netsnmp might get built before openssl and poison the
configure cache since it's not a mandatory dependency.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Supporting multilib is much more than just passing --enable-multilib
to gcc. You have to actually build the C library several times (once
for each multilib variant you want to support in your toolchain), and
to pass MULTILIB_OPTIONS/MULTILIB_EXCEPTIONS values to gcc to let it
know the set of multilib variants you're interested in.
Since we'll probably never support multilib toolchains in Buildroot,
just get rid of this BR2_ENABLE_MULTILIB option.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This is a very advanced option, and it seems, according to
http://choices.cs.uiuc.edu/exceptions.pdf that SJLJ exceptions aren't
really interesting.
Users really interested by this can always use the
BR2_EXTRA_GCC_CONFIG_OPTIONS is they want.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
