HOST_SED_BINARY is used 2 times during startup, and is relatively
expensive to run, so make it a simple expanded variable (:=) rather
than a recursively expanded one (=).
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Bash seems to be smart enough to source the file when execve returns
ENOEXEC, but other shells might not be.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
GNU_HOST_NAME is referred more than 100 times in the tree, used 5
times during startup, and is relatively expensive to run - so make
it a simple expanded variable (:=) rather than recursively expanded
one (=).
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
As noticed by Thomas, we call the UPPERCASE macro a lot, and it slows down
startup quite a bit.
Optimize it by implementing it in make, rather than forking a shell + tr.
The implementation is heavily based on the 'up' macro from gmsl
(http://gmsl.sf.net)
With this in place, startup time is ~5 times lower.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
On my target rootfs it is installed documentation for:
atk, cairo, gdk, gdk-pixbuf, gio, glib, gobject, gtk, pango
It ocupies 34M of space. Configure option --disable-gtk-doc doesn't
prevent installation of it. So we should remove them manually if
documentation is disabled.
[Peter: also remove manual rm's from hal/libxml2/libglade]
Signed-off-by: Paulius Zaleckas <paulius.zaleckas@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Closes#1525
Kconfig selects are not recursive, so ensure we select the dependencies
of our dependencies as well.
Cleanup target installation:
- Remove tasks handled by the generic post-build cleanup
- Hal doesn't install etc/rc.d, so don't try to delete it
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Fix a few typos/mistakes introduced in commit ea448fe:
- bzip2 soname is not the same as package version
- It is PKG_INSTALL_STAGING, not PKG_STAGING_INSTALL
- Typo: BZIP2_INSTALL_TARGET_CMDS, not B2IP_INSTALL_TARGET_CMDS
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Some packages like icu requires to be compiled against the host system
first to be able to compile against the target. This is due to the
usage of self generated binaries by the package to build itself. When
the generated tools also depends on generated libraries it is required
to add the path to these libraries in the library path
(LD_LIBRARY_PATH) especially for the configure step.
Adding $(HOST_DIR)/usr/lib to LD_LIBRARY_PATH for target compilation
might break the link step by mixing host libraries and target
binaries.
Signed-off-by: Lionel Landwerlin <llandwerlin@gmail.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
If devtmpfs (the kernel-maintained /dev filesystem) is used, no
/dev/pts directory is created, causing the devpts mount to fail, which
in term causes stuff like dropbear to fail.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Also convert to generic package infrastructure
Signed-off-by: Lionel Landwerlin <llandwerlin@gmail.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
These are ancient (2006) and upstream strongly discourage using them:
ftp://sourceware.org/pub/gdb/old-releases/README
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
It got broken by the introduction of the new autotools infrastructure
based on the generic infrastructure.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Instead of asking the user about the GNU target suffix, just compute
it automatically from the other configuration options.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The ROOTFS_SUFFIX thing has been removed in
325bfd1cba, so get rid of the last users.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
In both internal and external toolchain cases, KERNEL_CROSS was
defined to *exactly* the same value as TARGET_CROSS. It isn't modified
anywhere, and is just used by kernel compilation and pcmcia
compilation.
Therefore, get rid of KERNEL_CROSS and use TARGET_CROSS instead.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The definition of TOOLCHAIN_DIR is the same regardless of whether
external or internal toolchains are used. Moreover, move its
definition together with all the other *_DIR definitions.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Since new configuration options have been added in 0.9.31, the value
of these configuration options should be determined, either by the
default configuration file we provide, or by uclibc.mk process.
The locale generation process should probably be improved in order to
allow building other locales than just en_US.
[Peter: fixup locale handling, add PROGRAM__NAME to defconfig]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
UCLIBC_HAS_NFTW is a new knob in 0.9.31, which allows the obsolete and
deprecated ftw() to be compiled-out separatly from nftw(), which is
part of POSIX. nftw() should probably be enabled by default in uClibc,
and a bug has been opened about this on uClibc bug tracker
(https://bugs.busybox.net/show_bug.cgi?id=1597).
nftw() is, for example, used in Gtk+.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Paulius Zaleckas <paulius.zaleckas@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
When FOO_LIBTOOL_PATCH is defined, assume that the same value should
be used for HOST_FOO_LIBTOOL_PATCH.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
OpenSSL is not using the autotools as its build system. Therefore, we
must use the generic infrastructure instead of the autotools one.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Do not build against libgcrypt on the host, since we don't build
libgcrypt ourself, and it might fail even when libgcrypt is installed
on the host:
...configure...
checking for libgcrypt-config... /usr/bin/libgcrypt-config
Crypto extensions will be available.
...build...
gcc: /usr/lib/libgcrypt.so: No such file or directory
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
The autoreconf check was incorrect, missing a $ sign to properly
reference a package-specific variable. There was no visible effect
until now since :
* The existing syntax allowed to access the value defined in the
package specific .mk file, so when AUTORECONF was set to YES by a
package, it was working.
* The default value in Makefile.autotools.in was NO. In fact, when a
package .mkf file wasn't defining the AUTORECONF variable, the
Makefile.autotools.in test was testing the empty string against
'YES', which was false, leading to the AUTORECONF not being done,
which was the desired effect.
However, in a later patch, we intend to change this default value.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
install -D needs the destination filename, not just it's directory,
otherwise we end up with a pkgconfig file instead.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
autoheader is used by autoreconf, therefor we need to give the right
path to this tools for host/target autoreconfiguration.
Signed-off-by: Lionel Landwerlin <llandwerlin@gmail.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
