Thomas Petazzoni [Sun, 22 May 2011 11:56:08 +0000 (13:56 +0200)]
fakeroot: disallow selection for the target
fakeroot for the target fails to build against glibc, due to largefile
related problems :
libtool: link: /home/test/outputs/test-503/host/usr/bin/powerpc-linux-gnu-gcc -pipe -Os -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -o faked faked.o ./.libs/libcommunicate.a -ldl
{standard input}: Assembler messages:
{standard input}:2592: Error: symbol `__fxstatat64' is already defined
{standard input}:2627: Error: symbol `__fxstat64' is already defined
{standard input}:2663: Error: symbol `__xstat64' is already defined
{standard input}:2698: Error: symbol `__lxstat64' is already defined
make[3]: *** [libfakeroot.lo] Error 1
As fakeroot for the target is not really needed, and was only added
when we packaged fakeroot for the host, just make fakeroot not
selectable for the target.
We have to keep the AUTOTARGETS call for fakeroot on the target,
because the host variant does not work otherwise, due to issues in the
package infrastructure (that will be fixed post-2011.05).
If the scripttools are enabled in the Qt configuration, but not the
gui, then the scripttools library is not built by Qt, and the
Buildroot build process fails with:
cp: cannot stat `/home/test/outputs/test-502/host/usr/sh4-unknown-linux-gnu/sysroot/usr/lib/libQtScriptTools.so.*': No such file or directory
Thomas Petazzoni [Sat, 21 May 2011 15:47:50 +0000 (17:47 +0200)]
psmisc: do not install pstree.x11 symbolic link
The psmisc-no-__progname.patch removes the usage of __progname, which
removes any difference between pstree and pstree.x11. Therefore, we
make sure to not create the pstree.x11 symbolic link in the target.
Yann E. MORIN [Sat, 21 May 2011 22:05:42 +0000 (00:05 +0200)]
toolchain/crosstool-NG: fix up ct-ng config file to saner defaults
Peter on IRC reported some build failures for different targets.
They were of two kinds:
- missing/unknown stack unwinding support
- missing *_chk functions
The first is about configure not being able to automagically determine
if stack unwinding support is available for the target. The second is
about fortified build forgetting to build the fortified functions.
This applies to both glibc and eglibc.
After some discussions on IRC with Jacmet, it appears that we can
safely assume both of the following:
- virtually all targets of buildroot will have stack unwinding support
- we do not care about fortified builds (so far)
So, update the bundled crosstool-NG .config file to saner defaults:
- force unwind support
- disable fortified builds
[Peter: fix 'force non-fortified build' sed invocation] Reported-by: Peter Korsgaard <[email protected]> Signed-off-by: "Yann E. MORIN" <[email protected]> Signed-off-by: Peter Korsgaard <[email protected]>
Yann E. MORIN [Sat, 21 May 2011 22:05:41 +0000 (00:05 +0200)]
toolchain/crosstool-NG: update to 1.11.3
1.11.3 is the latest stable version, and it has a few improvements
we'll find usefull. For example, it hasa a boolean option for
disabling the {,e}glibc fortified build (instead of requiring us
to shoehorn it in extra cflags.
Thomas Petazzoni [Thu, 19 May 2011 17:23:29 +0000 (19:23 +0200)]
bind: give path to OpenSSL when OpenSSL is enabled
bind needs to be specifically told where OpenSSL is, otherwise, the
build fails with:
checking for OpenSSL library... configure: error: OpenSSL was not found in any of /usr /usr/local /usr/local/ssl /usr/pkg /usr/sfw; use --with-openssl=/path
If you don't want OpenSSL, use --without-openssl
Thomas Petazzoni [Wed, 18 May 2011 10:01:24 +0000 (12:01 +0200)]
gtk2-engines: bump, rename config option and convert to AUTOTARGETS
gtk2-engines is bumped from 2.9.1 to 2.20.2 (the latest available
version compatible with Gtk2). The package is converted to the
AUTOTARGETS infrastructure.
The config option is renamed so that it matches the package name
(gtk2-engines instead of libgtk2-engines).
Gustavo Zacarias [Mon, 16 May 2011 14:18:26 +0000 (11:18 -0300)]
libnl: minor package fixes
* libnl fails to build with multiple jobs so keep make at -j1
* Multiple libraries are installed so update uninstall target for this
[Peter: remove libnl/libnl-*.* to catch subdir + any new libs in the future] Signed-off-by: Gustavo Zacarias <[email protected]> Signed-off-by: Peter Korsgaard <[email protected]>
Thomas Petazzoni [Sun, 15 May 2011 19:58:50 +0000 (21:58 +0200)]
gcc: remove references to arm-softfloat.patch.conditional
The arm-softfloat.patch.conditional patches are no longer in Buildroot
since 9af188d883b2657bd967e6621011d4bbe1bb3441 (October 2009), which
removed support for gcc 3.x and 4.0.x.
Thomas Petazzoni [Sun, 15 May 2011 19:55:02 +0000 (21:55 +0200)]
uClibc: remove AVR32 patches for 0.9.29
The 0.9.29 version of uClibc can no longer be selected on AVR32 since
December 2009 (b642f7eefbf8c7f0111d74528390de31aaacf203). Therefore,
those two patches are now completely useless.
Peter Korsgaard [Thu, 12 May 2011 15:15:36 +0000 (17:15 +0200)]
target/generic: only show getty port option if default skeleton is used
The special marker in etc/inittab might not be present with a custom
skeleton. At the same time make the option always active, remove the
hardcoded tty1/tty2 gettys and reword the option description slightly.
Allow several device tables and split in two parts our device table
This allows to have a device table for all directories/files and
another device table for the device files themselves. Both are needed
for static /dev, but only the first one is needed when
devtmpfs/mdev/udev are used.
We take this opportunity to move the documentation of the device table
format in a common location, package/makedevs/README.
[Peter: simplify code slightly, fix indentation] Signed-off-by: Thomas Petazzoni <[email protected]> Signed-off-by: Peter Korsgaard <[email protected]>
The S10udev startup script did useless things such as mounting sysfs,
mounting a tmpfs filesystem in /dev/, etc. Those things are already
done during in the default inittab.
Create menu entry to select device creation method
Four methods for the creation of device files in /dev are now
proposed:
- static method uses device table as before
- devtmpfs method enables this feature in kernel
- mdev method adds mdev starting script to the file system
and selects mdev itself for installation
- udev method selects udev for installation
All dynamic methods are based on devtmpfs, so one doesn't need to care
about /dev folder.
The recent commit adding the external toolchain wrapper has broken the
support for external toolchain. The check_arm_eabi, check_cplusplus
and check_cross_compiler_exists functions were using TARGET_CC, which
points to the toolchain wrapper, but at the moment those functions are
called, the wrapper hasn't been generated yet.
We fix this by passing to these functions the path to the C or C++
compiler they should use for their tests.
Peter Korsgaard [Mon, 2 May 2011 21:58:20 +0000 (23:58 +0200)]
Get rid of unneeded CFLAGS
Now that we use a wrapper for external toolchains (and internal ones
default to the correct setting), we no longer need to explicitly pass
sysroot/march/mtune/mabi/floating point mode in TARGET_CFLAGS.
Peter Korsgaard [Fri, 29 Apr 2011 11:09:26 +0000 (13:09 +0200)]
Add toolchain wrapper for external toolchains
Add a simple toolchain wrapper for external toolchains, which forces the
correct sysroot/march/mtune/floating point options needed to use it
with buildroot.
With this in place the external toolchain behaves similar to the internal
ones, and the special handling can be removed. This also means that the
toolchain is usable outside buildroot without having to pass any special
compiler flags.
Also adjust the downloadable external toolchain support to install under
HOST_DIR so it can be used after the temporary build files are removed.
Bump bind to version 9.6-ESV-R4 since the 9.5 series is no longer
supported.
While at it switch to the ESV (Extended Support Version) branch instead
of going for the latest.
Also adjust the package to build against OpenSSL when available for
proper DNSSEC support and libxml2 too.
projects / buildroot-mgba.git / log