| Commit message (Collapse) | Author | Age | Files | Lines |
|\
| |
| |
| | |
Android 5.1.0 release 1
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Refactor JniInvocation::GetLibrary to not return a stack-allocated
string. Instead, provide a char buffer.
Bug: 16404669
(cherry picked from commit 88b84ec200091bdf5754b435ef55dfcd30078a67)
Change-Id: Id5c1857eb0fdd413b87f839d0c57168d587b5c11
|
|/
|
|
| |
Change-Id: I0e0e63966502d31d7b7a63e71adedf68d1b24d27
|
|
|
|
| |
Bug: 16404669
Change-Id: I7bf7745684747bad8564f168d0def71878566e8c
|
|
|
|
|
|
|
| |
On device, only allow "libart.so" in non-debuggable (user) builds.
Bug: 16404669
Change-Id: Ie163c04ce40c82698dcc98ced651dafef094d8b5
|
|
|
|
| |
Change-Id: Ie162bb4969df7680cf88989a51862148d9779b83
|
|
|
|
| |
Change-Id: I375bff0f7f242f48fc2674c1d502501ff2197396
|
|
|
|
| |
Change-Id: Ib5510add3337992dc2a99affe671a01d39782237
|
|
|
|
| |
Change-Id: I8fe4d0a13eef77b19f3c29180cce26aa3bf90abe
|
|
|
|
| |
Change-Id: I9e91ac31f4c2d7b7ba7c2ed0fbdce1f2b01355e5
|
|
|
|
|
|
| |
This reverts commit 11ea128d72c0da41e773bfbf99ad13c08b0209af.
Change-Id: Id4c531493ebe763943d4fb6a0e89f0c0c6924b4c
|
|
|
|
|
|
| |
Not LOCAL_LDFLAGS.
Change-Id: Ia4e6508e7642897ab6b18e574dc8752c87a44b93
|
|
|
|
| |
Change-Id: I5f95fedfe64b465c8be1e02c482a2817740410a6
|
|
|
|
|
|
| |
Bug: 10680559
Change-Id: I18491baac6acc4486f2e87175c922b40b5c17179
|
|
|
|
| |
Change-Id: I15870937ad7bfa4757e21f83acbbb6a7826a0b0b
|
|
|
|
|
|
| |
(cherry picked from commit b66098568f22d672aac3949a39e92d8a474e4b94)
Change-Id: I60381a253af8ee738b1dac8ec124bf13aa7f5f82
|
|
|
|
|
| |
Bug: 6369821
Change-Id: I502211bddaf60d6097b81b35f606bbcdd1a3e092
|
|
|
|
|
|
|
| |
These aren't necessarily good abstractions, but they're no worse than what
we had, and having them factored out is a step in the right direction.
Change-Id: I5b839608317d2ca1ca54d8a38624fb686f2c37de
|
|
|
|
|
|
| |
Ouch.
Change-Id: I81c0457a95549f1bef7cc8d9ab23d6ca4475cdb5
|
|
|
|
|
| |
Bug: 2281992
Change-Id: Ibfb877bc78396bc0e947ae2f8f86e484843f07de
|
|
|
|
| |
Change-Id: I29a78ccb6312d12b3ef0ed02cfbc638a5b01db9c
|
|
|
|
| |
Change-Id: Id166ab03a01833e06ec1a9aa0059e4eb28ec04ce
|
|
|
|
| |
Change-Id: Ief1aa1c5f6da39f4d94cb68b285d8e02cc10b9ac
|
|
|
|
|
|
| |
them "*-host".
Change-Id: Ibf565c9c4b0709eb25edd71786b269895a1616a4
|
|
|
|
|
|
|
|
| |
The dalvikvm executable still doesn't quite build (fails in linker), which
is probably due to the core library native code not getting compiled and
linked. Unfortunately, it looks like that will be pretty hairy to untangle.
Change-Id: Ib5de623e43a40116c4b7459a45a246d76886edf3
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Most of this is just factoring out common make defs to include both on
the host and device side, though there is a little bit of code change
to be slightly POSIX friendlier and to avoid unnecessary sanity checks
when compiling without an asm version of the interpreter.
This patch should cause no difference in build results, except if you
turn on the newly-defined WITH_HOST_DALVIK build flag. If you do turn
it on, at this point you should probably expect the build to fail due
to insufficient dependencies, which I of course expect to address in
upcoming patches.
|
| |
|
| |
|
|
|