Age | Commit message (Collapse) | Author | Files | Lines |
|
It turns out buildroot does not currently accept a toolchain where a dynamic
linker does not reside in the multi-os-directory. Unfortunately this is
how glibc installs itself on AArch64 without any extra tricks.
So, provide an option to force everything into /lib or /usr/lib; patch to
buildroot will be worked on separately.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
GCC does not distinguish the resulting binary by the CFLAGS (e.g. based
on which -march= was given). This means, while it will use the right
libraries for linking, at runtime they are all going to request the same
ld.so path and load the libraries from the same default path.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
... and default to 'n'.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
... as we are going through various multilibs in ld.so.conf
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Also a fix for CT_IterateMultilibs: it didn't pass multi_os_dir_gcc, so
it only worked if the caller did *not* declare it as a local variable.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Convert absolute targets to relative so that they are valid on the host,
too. The procedure is very similar to uclibc, so it is moved into a
common function.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
... and in addition to final toolchain aliasing, use it when configuring
multilibs for glibc/musl. Note that uClibc does not need it, it is
explicitly selecting the tools using CROSS_PREFIX.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
binutils 2.23 and earlier are not affected, they had the correct code.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
If libc ended up in a directory other than /lib, place GCC's libs
in that directory as well.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
SPARC glibc (possibly others) misconfigured on macos
|
|
Fix unifdef in Linux 2.6.32 on macos
|
|
... which fails to compile due to incompatible prototype for strlcpy()
which isn't even used.
2.6.33 dropped the prototype, so the patch is n/a for newer kernels.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Skip locales on macos
|
|
A couple of fixes for canadian samples on macos
|
|
FreeBSD build support
|
|
It picks up gettext string and results in [ERROR] messages from ct-ng
when gettext strings happen inside an error() call.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
--enable-hacker-mode is not sufficient, in 2.25 configure then
fails while checking for sysdeps fragments that apply to a given
configuration, and with that worked around, fails on binutils &
compiler version check.
In brief: if someone wants locales on cygwin/macos, you'd have to
implement cross-localedef (similar to cross-rpcgen) in glibc and
submit it upstream.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
I.e., only to CFLAGS_FOR_BUILD if the build is canadian.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
The latter does not prevent zlib's configure from overriding 'AR' with
/usr/bin/libtool on macos, and that breaks canadian crosses.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Sanitize CT_PREFIX_DIR, too
|
|
It's not needed now that it's been moved to glibc.sh
Signed-off-by: Dan McGregor <dan.mcgregor@usask.ca>
|
|
GMP's configure script tries to be too smart, and if it determines
that it's not cross-compiling it chooses gcc or cc instead of the
wrapper we create at the start of the build.
Signed-off-by: Dan McGregor <dan.mcgregor@usask.ca>
|
|
Having .. in it breaks GCC's relocatability.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Backport patches to glibc 2.14+
|
|
... fix building on ARM.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Final batch of updates before 1.23
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
... and updated cygwin patch.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
... when we're actually don't.
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
If glibc's configure finds the host c++ executable it assumes that
c++ should be enabled for the build. In case we don't have cross g++
built yet (ie, for headers), this causes the build to fail creating
C++ headers. So hide C++ from the build.
Signed-off-by: Dan McGregor <dan.mcgregor@usask.ca>
|
|
It's possible that "gcc" is not the compiler being used for the build,
so respect BUILD_PREFIX and BUILD_SUFFIX when finding its version.
Signed-off-by: Dan McGregor <dan.mcgregor@usask.ca>
|
|
configure.ac now finds how to count the CPUs in a system. Currently
the getconf method and sysctl methods are supported. Adding more is
easy enough.
Signed-off-by: Dan McGregor <dan.mcgregor@usask.ca>
|
|
Check for FreeBSD specific issues, it is mostly the same as Darwin.
Signed-off-by: Dan McGregor <dan.mcgregor@usask.ca>
|
|
Fix uClibc{,-ng} build on macos
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Backport the fix from 2.19 for cross-rpcgen on macos
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Skip localedef
|
|
No 'cp -T' on macos
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
Signed-off-by: Alexey Neyman <stilor@att.net>
|
|
... or it leaves them with no valid choices.
Signed-off-by: Alexey Neyman <stilor@att.net>
|