If LOCAL_WHOLE_STATIC_LIBRARIES contains a library that has
two files that have the same name but are in different
directories, only the first gets included.
This fix detects this case, and uses the m option to ar to force the
duplicate object to the end of the archive. After this, using the p
option gets the correct object file.
Change-Id: I2e183f48cef3c79499d4ab8ff147444611ff938b
This will be necessary to support -std=gnu99 mode for clang 3.6, which
defaults to C11 mode (unlike prior releases that use C99).
Change-Id: Iea84582f9f12ba76b988463cbc0a20bd61042538
This caused a lot of failures in the libc++ tests, since linking
libgcc.a before libc++.so means that the wrong unwinder will be used.
This patch is still correct, but I will revert for now while I find a
good way to apply it without breaking unwinding.
This reverts commit 741b36e922.
Change-Id: I57e0b89534adf5c1677dbf2f64ac750e74bd5fbf
Previously we store tag's installed modules in a separate variable
ALL_MODULE_TAGS.$(tag). However we only record the main installed file
for a module, but omit affiliated files like .odex.
With this change, we handle the tagged module the same way as modules in
PRODUCT_PACKAGES.
Change-Id: I7972528a4df5a4ba8bc98930864da9672d32d7fe
The previous position of libgcc.a/libatomic.a on the link line causes
the linker to prefer satisfying dependencies from these libraries from
other libraries that might include them, rather than from libgcc.a (or
libatomic.a) itself. This imposes an ABI requirement that those
intermediate shared libraries _always_ export those symbols, which is
undesirable.
Change-Id: Ib593236b475d3e98356b2b1be6f96cee2b67378f
Previously, there was one generic definition for each of the
transform-o-to-* functions in definitions.mk, and one target specific
one in each combo/TARGET_*.mk. The generic one was entirely unused,
and the target specific ones were all nearly identical.
Changing anything in these functions was tedious at best, and often
error prone. The differences between any 32-bit arch and its 64-bit
equivalent were restricted to the name of the linker, and the ARM and
MIPS definitions were identical. The few differences between ARM and
x86 looked to be compatibility for an old (ca. 2008) toolchain issue
with --gc-sections, and a bug (LDFLAGS coming first rather than
later).
To simplify things, I've moved the definitions for these out of
combo/TARGET_*.mk and back into definitions.mk. The differences
between ARM and x86 have been scrapped. Anything that really does
still need to be target specific can be handled as I have the linker:
add a TARGET_FOO variable to the given target and then add it to the
generic definition.
Change-Id: I54dc1bffc32ac39f27f0b87247dd6a6dbaf0b162
build/core/Makefile looks for build tasks in either device/*/build/tasks
or vendor/*/build/tasks. Do the same with vendor-specific definitions.mk.
Change-Id: Ib4fd00a1d0effe8e347382a922da101ce26bf696
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Bug: 14416410
Our linker uses "start" as the primary entry point, so we need to adjust this
for mingw-based Windows compiles, or nothing will execute correctly.
Change-Id: I6e99f43e075ef9f00500099ce34ec4425c996454
So result of get-prebuilt-src-arch can be passed to
LOCAL_MODULE_TARGET_ARCH, which takes empty string as "any" actually.
Change-Id: I916c9738ccce4a94ac084fb4141d54659e896a1f
Improves performance for LOCAL_WHOLE_STATIC_LIBS by copying the first .a
to the new .a rather than extracting and recreating
Change-Id: Iecdb5e4bb2ce987bb41a70c3393d18a6d72ae689
This fixes mips unbundled build since we switched to mips64el toolchain
for both mips and mips64.
TODO: multilib build support.
Change-Id: I7add92d2cecfc3ab739785ceef6700240a25093a
This fixes mips unbundled build since we switched to mips64el toolchain
for both mips and mips64.
TODO: multilib build support.
Change-Id: I7add92d2cecfc3ab739785ceef6700240a25093a
We've been using -fPIC and -fPIE together in the global cflags all this
time. These options are incompatible. The only reason we haven't been
hit by this before is because of the forced -Bsymbolic in GCC. To fix
this, pass -fpic when compiling objects for shared libraries and -fpie
when compiling objects for executables. For static libraries, also use
-fpic. We have to do this because static libraries might be included in
either a shared library or an executable. Code compiled with -fpie
cannot be included in a shared library, but code compiled with -fpic
may be included in an executable.
We've also been using -fpic and -fPIC together. These are different
options, and only the latter will take effect.
http://stackoverflow.com/a/967010
The final thing this fixes is that we had -f(PIC|PIE) flags being passed
to link commands. These are compile time flags, and don't do anything at
link time.
Bug: 16823325
Change-Id: Ic76f47e63dc2c81b7e1a8058bae1b3dc8565d606
(cherry picked from commit 4803ce2696)
We've been using -fPIC and -fPIE together in the global cflags all this
time. These options are incompatible. The only reason we haven't been
hit by this before is because of the forced -Bsymbolic in GCC. To fix
this, pass -fpic when compiling objects for shared libraries and -fpie
when compiling objects for executables. For static libraries, also use
-fpic. We have to do this because static libraries might be included in
either a shared library or an executable. Code compiled with -fpie
cannot be included in a shared library, but code compiled with -fpic
may be included in an executable.
We've also been using -fpic and -fPIC together. These are different
options, and only the latter will take effect.
http://stackoverflow.com/a/967010
The final thing this fixes is that we had -f(PIC|PIE) flags being passed
to link commands. These are compile time flags, and don't do anything at
link time.
Bug: 16823325
Change-Id: Ic76f47e63dc2c81b7e1a8058bae1b3dc8565d606
- Better messaging if the expected split apk isn't generated by the aapt
command in the base apk rule; Remove the built base apk, so make will
rerun aapt after the user changes the splitting parameters.
- Use cleaner static pattern rules instead of running $(foreach) with
$(eval).
Bug: 16319961
Change-Id: If6ae302e1a39d2e0db8f784d4e1cf292ec855281
Support LOCAL_PACKAGE_SPLITS, which accepts a list of resource lables
and generates multiple apks. The build system sets up rules to sign and
zipalign the split apks.
Bug: 16319961
Change-Id: I344b3d1c7eb158c6d0df879093d666a89870aadd
Apk's path is changed to <parent_dir>/MyApp/MyApp.apk;
JNI path is changed to <parent_dir>/MyApp/lib/<arch_name>/libfoo.so.
Symlinks of JNIs are changed accordingly.
Bug: 16319961
Change-Id: Ib3b2309c95fa9aea27837fcc29e28d990b04747b
Dx is now capable of generating multiple dex file in one run.
This capability is not compatible with --output=*.dex so this CL
changes the --output argument to be a directory. This CL also
includes the packaging of the multiple dex files when dx generates
more than one dex file.
(cherry picked from commit 8ffe9c3e79)
Conflicts:
core/java.mk
Change-Id: I1ddc91b416de5929d4d4b41824c6012b32f94deb
Use "LOCAL_MULTILIB := both" to install jni libraries of both archs in
multilib build.
The build system will package jni of both archs to the apk, or install
them to the right location on the system image and create symlinks,
extract .so files from prebuilt apk, etc if appropriate.
Bug: 15849902
Change-Id: I7e147b5a47db476584c38250de7b36c75ea40d81
- Remove the incorrectly used (and unused) 3rd arg to assert-max-image-size
- Remove test for fstype in assert-max-file-size() as it was always
set to a value leading to 0.
Change-Id: I0a1182fcc85c6c3801d45731e691b9bb38657606
Signed-off-by: JP Abgrall <jpa@google.com>
Previously we only expanded product_MODULES with LOCAL_REQUIRED_MODULES,
but not modules introduced by LOCAL_SHARED_LIBRARIES; Later we did a further
shared libary expansion in vendor_module_check.mk.
It couldn't track C in the following case:
A : B, by LOCAL_SHARED_LIBRARIES; B : C, by LOCAL_REQUIRED_MODULES.
With this change, we transformed the LOCAL_SHARED_LIBRARIES dependencies
into LOCAL_REQUIRED_MODULES dependencies before doing the required
module expansion and the loophole is closed.
All module names are now expanded to product_MODULES now and it makes
vendor_module_check.mk simpler.
Change-Id: I8835a478d2ce0ce10601a8449f446f07b01c2b7f
Previously the RS cpp files are generated by the timestamp rule. Though
we have the generated RS cpp files depend on the timestamp file, we
don't have a build recipe. In such case gmake does some "optimization"
that it skip recompiling the generated cpp files, because it assumes the
generated cpp files are already up to date even if the rs files have
been updated.
Bug: 15313144
Change-Id: Ie69ecd2c788057d3619f9c7d2a125d44c4a534a1
Allow LOCAL_JAR_PACKAGES to work on subpackages. Previously it
only worked on top-level packages due to a bug in how the directories
were recursively deleted.
Add LOCAL_JAR_EXCLUDE_PACKAGES to allow specific subpackages to
be excluded. This rule applies after LOCAL_JAR_PACKAGES has
selected which packages to include. It may also be used independently.
Change-Id: Ibd0e495be1a20c84b59c9da132e92100ef3f8705
