Revert r25827 and instead copy over the host aclocal files of the stuff
needed for auto* (libtool + pkgconfig). These logically belong with the
target stuff, as they are used for the target auto* toolchain.
Long term we should probably consider using
aclocal --acdir=$(STAGING_DIR)/usr/share/aclocal, so the host .m4 files
cannot interfere with the target build.
Change the definition of TARGET_LDFLAGS to use --sysroot
$(STAGING_DIR) instead of -L$(STAGING_DIR)/lib
-L$(STAGING_DIR)/usr/lib. It fixes the following failure while trying
to build mtd-utils :
/usr/local/xtools/arm-unknown-linux-uclibcgnueabi/bin/arm-unknown-linux-uclibcgnueabi-gcc -L/home/thomas/local/buildroot-output/build_arm/staging_dir/lib -L/home/thomas/local/buildroot-output/build_arm/staging_dir/usr/lib -o /home/thomas/local/buildroot-output/build_arm/mtd_orig/flash_eraseall /home/thomas/local/buildroot-output/build_arm/mtd_orig/crc32.o /home/thomas/local/buildroot-output/build_arm/mtd_orig/flash_eraseall.o
/usr/local/xtools/arm-unknown-linux-uclibcgnueabi/lib/gcc/arm-unknown-linux-uclibcgnueabi/4.3.2/../../../../arm-unknown-linux-uclibcgnueabi/bin/ld: cannot find /lib/libc.so.0
collect2: ld returned 1 exit status
make[1]: *** [/home/thomas/local/buildroot-output/build_arm/mtd_orig/flash_eraseall] Error 1
At the same time, simplify the definition of TARGET_CFLAGS, because
the -I$(STAGING_DIR)/include -I$(STAGING_DIR)/usr/include
-I$(TOOLCHAIN_EXTERNAL_PATH)/$(TOOLCHAIN_EXTERNAL_PREFIX)/include are
no longer necessary since we sysroot the toolchain in $(SYSROOT_DIR).
This patch has no effect on non-external toolchain builds.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
LIVE555 was installed on the build system and was recognized by configure:
Checking for LIVE555 Streaming Media libraries ... yes (using distribution version)
Later this led to a linker error:
ld: cannot find -lliveMedia
Disable LIVE555, since it's not available in Buildroot.
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
pango likes to rerun it's configure script at make time, and it then
fails to find the correct pkg-config files (it finds the ones ins
STAGING_DIR) causing problems if they are not compatible.
Fix it by setting PKG_CONFIG_PATH in it's make environment.
Move stamp (dependency) files outside the (version specific) source
directories, so other packages can hardcode dependencies on them instead
of having to use <PACKAGE>_VERSION variables.
This is important as the variables in the make rules are evaluated when
the rules is seen, which might be before the dependent makefile is parsed
(and hence <PACKAGE>_VERSION variable is known, screwing up stuff.
The downside of this is that the package isn't automatically rebuilt
when the version changes (E.G. by a svn update) and you now also have to
remove the stamp files next to $(BUILD_DIR)/<PACKAGE>-* to force a rebuild.
This matches upstream tarball, doesn't screw up existing .config's with
BR2_PACKAGE_PKGCONFIG and makes sure the patch gets applied for target
compilation.
Generate pango.modules using the host tools instead of setting up an init
script to do it at run time. Also apply patches to host build to get
identical pango.modules file.
The host versions shouldn't be visible in Kconfig, so remove the
reference to BR2_PACKAGE_PKGCONFIG everywhere and prefix the host targets
with host-.
At the same time add pkgconfig for the target (E.G. for development) and
let BR2_PACKAGE_PKGCONFIG control that package.
Notice: all defconfigs in the tree have been updated, but make sure to
disable the pkgconfig package (unless you want it) if you use an external
config, otherwise you'll end up with pkgconfig and glib2 in the target.
Choosing between GPL or LGPL doesn't change the build process, so it's
sufficient to have a selection for the commercial license.
Also change the URLs to the licenses to the Qt web page, where the
exception for the LGPL is included.
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
Not particular useful until we add dbus-python (which depends on a fixed up
dbus-glib, which in turn depends on us starting to compile glib/dbus/dbus-glib
for the host as well) - But that's for another day.
Patch by Gustavo Zacarias <gustavo@zacarias.com.ar>
Closes#151.
* Avoid fips directory completely since it just installs source file cruft
inlib
* Point openssldir to a more friendly and common /etc/ssl rather than
/usr/lib/ssl
Patch by Gustavo Zacarias <gustavo@zacarias.com.ar>
Closes#145.
Current bind package is version 9.3.2 which is from the 9.3 branch and is
EOLed. It has many security bugs probably fixed in 9.3.6-P1 but since it
won't be supported for long it's probably metter to move on to a supported
branch. CVE-2009-0025, CVE-2008-1447, CVE-2008-0122, CVE-2007-2926 and
probably more. While at it migrate to Makefile.autotools.in too. Also
introduced an option for/not to install userland tools (dig, host, nslookup,
nsupdate).
[ Peter: don't install into staging ]
Patch by Gustavo Zacarias <gustavo@zacarias.com.ar>
Closes#5, #77, #141 and #143.
* Migrate php package to Makefile.autotools.in
* Make PCRE regex an option (previously forced on)
* Add calendar option from bug #77
* Add external sqlite3 option from bug #141
New option BR2_PACKAGE_PHP_PDO_SQLITE_EXTERNAL
* Add alternate php.ini configuration option from bug #5
Named BR2_PACKAGE_PHP_CONFIG and defaults to shipped config.
In most scenarios the shipped config isn't good enough, it has
a high memory limit for embedded for example.
* Changed some options from depends to select
Namely openssl, libxml2, zlib, gettext, gmp.
* Disabled some 'y' defaults to make things lighter by default
Namely libxml2, sqlite, pdo and pdo_sqlite.
* Made some of the extensions help text more verbose
Patch by Olaf Rempel <razzor@kopf-tisch.de>.
- bump version to 2.6.28
- fix download
- change compile ordering (build tc after ip)
- remove both binaries from target when cleaning up
[Peter: fixed target install strip]
This was forgotten when adding config options for zlib and bzip2. The
matroska plugin doesn't need these dependencies, they are optional.
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
- php comes with it's own embedded sqlite, so no need to select sqlite
- disable largefile support for !BR2_LARGEFILE
- php somehow forgets to link with -ldl, breaking the sqlite extension.
Fix it by disabling the (unused?) modload support in sqlite.
php.net only offers downloads though a strange php script with the file name
embedded in the *MIDDLE* of the URL, which isn't compatible with the DOWNLOAD
macro. Fix it by reverting php.mk hunk of r24689 to go back to $(WGET).
This ofcourse means that the primary site / fallback mirror stuff isn't used.
Patch by Maxim Grigoriev.
GCC 4.3.x would link with the correct shared library versions, but without
any RPATH, so it ended up looking for then in /usr/lib/ at runtime.
This is less than good, as the version there might not be compatible, and
even worse for distributions like Ubuntu, which doesn't install those
libs by default, the build breaks.
Fix it by forcing static linking instead.
As discussed on the list. I suspect the ARM abi thing isn't needed anymore
now we use the correct abi in the first place, but I'll leave it in until
after the release.
I know that changing something as fundamental as mpfr this close to the
release isn't really ideal, but it's listed as the minimum version
for gcc 4.3.x:
http://gcc.gnu.org/install/prerequisites.html
Bernhard's tree has been using 2.3.2 since October, so the risk is probably
not that big though.
The GNU_TARGET_NAME symlink and target_utils location were not correctly
adjusted to match the move of the toolchain to $(STAGING_DIR)/usr,
creating dangling symlinks.
Make sure the developer is informed why a package is not
visable and that they need to have a NPTL enabled
toolchain.
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
Buildroot is adopting depends on for toolchain features and
select for packages. Make this change in this file.
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
Have been having various issues with multiple jobs (-jX) and building
autoconf. Turns out that v2.63 has fixed this. So bounce version
to a working version.
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
The avahi-autoipd target install step adds /var/lib/avahi-autoipd as a
symlink to /tmp/avahi-autoipd, which is fine, except if your /var/lib is
already a symlink to /tmp (E.G. like the default target_skeleton is).
If that's the case, then you end up adding a /tmp/avahi-autoipd symlink
pointing to itself, causing subsequent builds to fail
(Too many levels of symbolic links in the ln -sf call). Fix it by simply
first removing the symlink if it's there.
This seems to be required to compile some of the inline asm.
This solves the following compiler errors:
h264.c: In function 'decode_cabac_residual':
h264.c:6158: warning: passing argument 4 of 'decode_significance_8x8_x86' discards qualifiers from pointer target type
cabac.h: In function 'get_cabac_noinline':
cabac.h:513: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm'
cabac.h:513: error: 'asm' operand has impossible constraints
While searching several weeks ago, I hit the following URL, which
unfortunately currently doesn't seem to work anymore and I can't
remember what the bug was about.
Also see http://bugs.uclibc.org/view.php?id=4604
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
This include file is only available, if uClibc is compiled with
UCLIBC_SUSV3_LEGACY defined and ftime() is not needed in this part of
mplayer at all.
This solves the following compiler error:
win32.c:67:23: error: sys/timeb.h: No such file or directory
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
Also fix whitespace and use an existing variable for qmake.conf, which
was forgotten in r23828, though it's only commented out.
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
package/multimedia/gst-plugins-good/gst-plugins-good.mk
Always disable ESD
Move JPEG/PNG
Add BZIP2/ZLIB options as posted by Markus
Add LIBXML2 dependency for annodex
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
Use select instead of depends on as it makes clear what people
need rather than hiding options based on choices they dont
know they need to make.
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
package/multimedia/gst-plugins-good/Config.in
gst-plugins-good does not build unless some package dependencies are forced.
This patch allows you to turn off some of the plugins with these dependencies and hides
others that require some packages (like PNG/JPEG)
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
This adds a patch to fix compilation against alsa-lib 1.0.18. It is basically a
grab of the change from mplayer repository. It also refreshes the other patches
so they apply clean.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
gcc < 4.2.0 doesn't support -Wno-overlength-stings, but gcc-4.3.x configure
fails to detect that, breaking the build.
Work around it by detecting the host gcc version (and store in HOSTCC_VERSION)
and set the proper configure variables for gcc < 4.2.0.
UCLIB_HAS_.* is not defined in Buildroot space, so the conditions always
evaluate to the same result. Use appropriate Buildroot variables.
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
The VFP is only available for a few ARM CPUs at the moment,
so this breaks the liboil build.
A patch is available upstream which only enables "-mfpu=vfp"
if "--enable-vfp" is given to "configure".
Autotools needs to be run for liboil for this to take effect.
A new configuration BR2_VFP_FLOAT is added to allow enabling vfp.
If this is "yes", then "-mfpu=vfp" is added to CFLAGS.
:= causes the function parameters to get evaluated at definition time
instead of implementation time, which is kind of silly for a function.
Fixes make source-check / external-deps
This is the long requested backup mirror support. So far only enabled
for Makefile.autotools.in, but there's no reason why we cannot use
DOWNLOAD everywhere.
Notice, I'm still syncing http://buildroot.uclibc.org/downloads/sources/
Found that the editing of freetypr-config was not quite enough.
Changed it and looked at freetype-config --cflags, --libs and got the values
I expected.
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
package/multimedia/gst-plugins-*: I do not want docs installed in rootfs.
Stop it installing the docs in the ROOTFS by using the same fix
as has been applied to cairo.
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
GNU Libidn is a fully documented implementation of the Stringprep,
Punycode and IDNA specifications. Libidn's purpose is to encode
and decode internationalized domain names. The native C, C#
and Java libraries are available under
the GNU Lesser General Public License version 2.1 or later.
Not quite sure I've put it in the correct menu but it will work.
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
package/gettext/gettext.mk: Both of these components look for iconv when configuring so need to be dependent.
Both of these packages can be built with libiconv support. If libiconv is enabled and these are built first
then some configure variables get stored in the system cache and then libiconv is not used properly.
If libiconv is selected by user make sure it is built before either of these packages are built.
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
This does not buld because its "configure" requires
that -llzo contains "lzolx_decompress".
"vtun" does not build, evenm if liblzo is available.
The LZO package currently used by buildroot
does not contain ANY reference to "lzolx_decompress"
"vtun" has been upgraded to 3.0.2 and now builds OK,
but is yet to be tested on a target.
The previous patch containing three diffs,
has been broken up into three files.
The second patch fails.
This patch tries to replace a perl script
($(VTUN_DIR)/scripts/vtund.rc.debian)
with a shell script with the same name.
In vtun-3.0.2, vtund.rc.debian is a shell script which is
fairly similar to the shell script provided by the patch.
For now, it has been decided not to replace this shell
script with the script generated by the patch for 2.6
vtun will thus be built with the 3.0.2 vtund.rc.debian.
The start-stop-daemon parameters and other things
in this script may be inappropriate for something based on busybox.
I will leave the decision which script to use,
the 2-6 script or the 3.0.2 script to someone else.
Both files are kept in the directory, but should
be removed once it has been decided what to do
about this script.
Signed-Off by: Ulf Samuelsson <ulf.samuelsson@atmel.com>
This patch updates the AVR32 arch patches for strace 4.5.18 and refreshes all
the other patches so they apply clean.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
git-svn (and git) doesn't handle empty directories, so add .empty files
to those dirs like elsewhere in buildroot.
Those empty directories are normally not a big deal, but the recent changes
to u-boot broke the build.
* When installing pci.ids[.gz] the -D means create all components leading up, using $(@D) does not include the filename, the last directory component is then /usr/share. pci.ids[.gz] is therefore created as /usr/share/misc (misc being the filename). Using $@ fixes this, the directory path is created and the file is installed as /usr/share/misc/pci.ids[.gz] .
* Do not install to the staging dir inside the target install target,
since it has already been done in the staging install target.
* Strip the binaries installed in the target.
* Properly create the target installation stamp.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
TARGET_PATH didn't contain $(STAGING_DIR)/usr/bin, which means that
programs installed in $(STAGING_DIR)/usr/bin were not considered for
execution during Buildroot build process. This was a problem with host
automake/autoconf/libtool, which could not be found.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
* In toolchain/external-toolchain/ext-tool.mk, copy the contents of
the sysroot directory to the staging dir.
* In package/Makefile.in, add a --sysroot CFLAGS pointing to the
staging dir
* Remove the CFLAGS and LDFLAGS definition from
TARGET_CONFIGURE_OPTS. I haven't investigated exactly why, but with
these options, DirectFB fails to build because it cannot find
PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP, even if DirectFB's Makefile
properly sets -D_GNU_SOURCE.
I have already sent this patch on December, 2nd to the mailing-list,
but got no feedback. So let's commit and see what happens :-)
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch fixes parallel build of Samba. The trick is to make proto before
executing a full make.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
This patch bumps Samba version to 3.2.5. In the process Samba now
depends on libiconv. This dependency is needed because Samba wants to use
various formats when talking to the different parts of the system.
Two new patches are introduced, one to double check if BSD functions are
available, and one to skip checking for glibc version. The use unsigned
enum patch could be removed since this can be supplied with a autoconf
variable.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
This patch is mainly for size reduction, removing more than half the
needed footprint.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
This patch will install zlib keeping the symlinks. The usage of
$(INSTALL) here did not take into account symbolic links.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
This patch adds a kconfig option to enable Phonon when configuring Qt.
Phonon is the multimedia API from KDE 4 which has been integrated into
Qt 4.4. For playback under Linux the GStreamer multimedia framework is
used.
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
The ARM uClinux BFLT format apparently doesn't support shared libs nor
-fPIC. So let's distinguish between shared and non shared builds by
having split rules for libz.a and libz.so based on BR2_PREFER_STATIC_LIB.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Currently, udev won't start for me, as we've just mounted a new /dev,
which is completely empty, and udevd requires a /dev/null.
This change manually creates the three base device nodes (zero, null
and console) so that udev can start.
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
If the udev extras are built without large file support in uClibc,
I get the following error:
build_powerpc/staging_dir/usr/bin/powerpc-linux-uclibc-gcc -c -g -Wall
-pipe -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -Wstrict-prototypes
-Wsign-compare -Wshadow -Wchar-subscripts -Wmissing-declarations
-Wnested-externs -Wpointer-arith -Wcast-align -Wsign-compare
-Wmissing-prototypes -Os ext.c -o ext.o
In file included from build_powerpc/staging_dir/usr/include/stdio.h:28,
from ext.c:19:
build_powerpc/staging_dir/usr/include/features.h:382:4: error:
#error It appears you have defined _FILE_OFFSET_BITS=64. Unfortunately,
uClibc was built without large file support enabled.
Although the base udev package builds correctly, the extras pacakges
don't adjust their CFLAGS depending on BR2_LARGEFILE.
This change adds the required CFLAGS modifications to make the
extras build.
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
According to udev.mk, the scsi_id package depends on the libvolume_id
library, so add this dependency to Config.mk
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
The configure script tries to find out if enums can be negative, this is not
supported when cross compiling, since it tries to run the binary. This commit
sets the test to yes, i.e. negative enums are possible.
Avahi uses autoconf during ./configure, make sure it is available and used
before building avahi.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
install-strip unconditionally overwrites $(TARGET_DIR)/etc/init.d/ifplugd.*,
so the check in _HOOK_POST_INSTALL is useless.
Fix it by using install-exec and manually stripping the target binaries
instead.
CPUs. It is required by various GStreamer plugins.
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
Modifications by Thomas Petazzoni:
- Remove the glib2 dependency, which is only necessary to build a few
examples. If glib2 is enabled in the configuration, then we still
depend on it in order to build the examples, but that's just
pedantic, not so useful.
- Remove the oil-bugreport program installed to the target by
install-strip.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The ac_cv_path_install override is needed to strip the target binary for
haserl 0.8.0, but it doesn't interact nicely with the shared cache, so
disable the cache for now.
Longer term we should probably forget about install-strip and do the
strip/cleanup/mklibs stuff as a post processing step before the target
filesystem rules are run.
This patch will add needed libraries with $(if ...) syntax. This should ease
readability and make it easy for others to spot where the depencies are added.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
This patch removes -I and -L from ./configure since it breaks cross-compiling.
The inclusion of -I and -L will confuse the host compiler when compiling the
host tools.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
This patch adds alsa-lib to a dependency for mplayer if it is selected in
kconfig.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
This patch will put dependency libraries in a variable and add this to the
dependencies to build mplayer. This is done to ease adding new libraries in the
future.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
This patch disabled ivtv video output, since it does not compile with Linux
kernel 2.6.27 headers.
Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
The following changes allow for use of a central configure cache
file. This speeds up configuration of packages.
Its use is configurable at the top level (BR2_CONFIG_CACHE - default n).
Old style makefiles can use it if they use the following MACRO in makefiles:
$(AUTO_CONFIGURE_TARGET) see my change to directfb.mk.
New style Autotools.in will use it if you set the global option.
However you can enable the global option and on a per package overrule it by doing
the following: $(PKGNAME)_USE_CONFIG_CACHE = NO see fontconfig.mk for an example
of this.
Finally I have removed a few config variable settings which indicated no CXX compiler
as this is wrong and breaks the build when using this central cache.
Config.in | 8 ++++++++
package/Makefile.autotools.in | 5 ++++-
package/Makefile.in | 28 +++++++++++++++++++++++++++-
package/atk/atk.mk | 2 +-
package/directfb/directfb.mk | 7 +------
package/fontconfig/fontconfig.mk | 3 +++
package/libglib2/libglib2.mk | 2 +-
package/libgtk2/libgtk2.mk | 1 -
8 files changed, 45 insertions(+), 11 deletions(-)
I would appreciate feedback on this change (I have been testing for 2-3 weeks)
But I can never test all cases! If you enable the BR2_CONFIG_CACHE option some
Makefile.autotools.in based packages may now break - I cannot build them all.
In this case you may need to remove config options that are being hardcoded all
over the place (like gtk saying we have 2 CXX compiler) or disable the use
of CONFIG CACHE file like I have done in fontconfig.
I can build all packages required to get WebKit on DirectFB up and running
and it runs fine.
I will try to resolve any issues this creates as fast as I can.
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
The id applet in 1.13.0 only compiles with uclibc < 0.9.30 if the
busybox internal passwd/grp functions are used.
Therefore, automatically enable CONFIG_USE_BB_PWD_GRP if that situation
is detected and warn the user.
Prepare for the merge of audio and video packages. Many packages cannot
properly be assigned to either audio or video, because they have support
for both (libogg, mplayer, vlc).
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
Also remove --enable-shared and --enable-static as it's default
and --disable-oggtest and $(DISABLE_NLS) as they are not supported.
Signed-off-by: Markus Heidelberg <markus.heidelberg@web.de>
This patch adds a patch to fix bug
https://bugs.freedesktop.org/show_bug.cgi?id=16464 affecting parallel
compilation of fontconfig.
The patch is the one proposed in the bugzilla entry, available at
https://bugs.freedesktop.org/attachment.cgi?id=17294.
Without this patch, the compilation (at BR2_JLEVEL > 1) of fontconfig
sometimes fails with:
In file included from fc-case.c:25:
../src/fcint.h:118:21: error: fcalias.h: No such file or directory
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch fixes two Qtopia build issues, encountered while trying to
use system implementation of zlib, freetype, jpeg and libpng :
* The build process doesn't look in $(STAGING_DIR)/usr/include for
includes and $(STAGING_DIR)/usr/lib. Same problem as the patch
currently floating around adding LDFLAGS to TARGET_CONFIGURE_OPTS,
but as Qtopia doesn't use TARGET_CONFIGURE_OPTS, we need a specific
fix here. So we use the -I and -L options of Qtopia's configure
script.
* The build process doesn't use pkg-config to get the header path for
Freetype headers (located in $(STAGING_DIR)/usr/include/freetype2
and not directly in $(STAGING_DIR)/usr/include/). There was already
a fix for this, consisting in adding $(FREETYPE_DIR)/include to the
-I path of Qtopia's configure. This patch modifies this fix to use
$(STAGING_DIR)/usr/include/freetype2 instead, which looks more
coherent with how all the packages are built (using $(STAGING_DIR)
as the reference to get headers and libraries).
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch is a new version of a patch already sent several times on
the mailing-list, committed and reverted a few times by Daniel Laird,
due to several imperfections. This version is a new try at finding a
solution that works for everybody. Hopefully it'll work :-)
The original problem is that external toolchain builds failed because
packages couldn't find their dependent libraries at configure time and
could not be linked with them. To fix these two problems, two things
are added:
* The TARGET_LDFLAGS variable was exposed as LDFLAGS at ./configure
time thanks to TARGET_CONFIGURE_OPTS. The TARGET_LDFLAGS variable
contains -L options with the path in the STAGING_DIR for the
libraries. It allows ./configure scripts to properly compile the
small test programs testing whether a dependency is properly
installed.
* The TARGET_CFLAGS contains a new -Wl,--rpath-link option for both
$(STAGING_DIR)/lib and $(STAGING_DIR)/usr/lib. It allows library
depending on other libraries to link properly. The TARGET_CFLAGS is
exposed as CFLAGS in TARGET_CONFIGURE_OPTS.
This new version fixes a problem encountered by hartleys
<hartleys@visionengravers.com> when building the kernel. The problem
was that the -Wl,--rpath-link options were added to LDFLAGS, while
there are options for the C compiler, not the ld linker. Moving them
to CFLAGS seems to fix the issue.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Bounce tiff into Autotools.in format.
Did not use install to target as this puts loads of executables into TARGET.
So just copied tiff.so instead.
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
Looking into adding a configure cache to the build (like the GIT buildroot version)
This means that freetype needs to know about zlib so make
sure it had it as a dependency.
Also remove install rule for staging as it matches default.
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
The following patch makes the MESSAGE Macro in Makefile.autotools.in
work.
I think it was originally intended to print the messages in bold type
but it doesn't appear to work correctly. This patch should work on all
platforms.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
(Fixed to not continously call tput)
Revert the rpath patch, it looked good up until someone tried
to build a kernel as well. This seems to break as a result.
Will post a new patch soon and see how that goes..
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
Apply the patch I posted some time ago that fixes
rpath issues with external toolchains.
Has been tested by users of buildroot and feedback looks good.
Signed-off-by: Thomas Petazzoni
Signed-off-by: Daniel Laird <daniel.j.laird@nxp.com>
