Age | Commit message (Collapse) | Author |
|
Seems this only fails on the Intel ABIs? According to the bug
history, I first marked this as failing when tested on an AOSP device
and some armeabi-v7a emulators. Perhaps it's just an issue with
emulated devices?
Bug: http://b/24380035
Change-Id: I07e5231e28da30a592c4d3c11beffd543b112849
|
|
This test is actually only broken for ARM's 32-bit ABIs.
Bug: http://b/26031639
Change-Id: Idd166f1f11012565cc3f7ad723557796108561ac
|
|
GCC is deprecated, so why is it still the default testing toolchain?
Change-Id: I6540370e5917b5b7e74f2869b4ab8d177e4ef68f
|
|
The libgcc and LLVM unwinders like to fight. This test will fail if
the unwinder is not linked to be tolerant of this.
Bug: http://b/26194502
Change-Id: I0d03f67edae1e53dac44fff3557666eb08e77a4a
|
|
Change-Id: Iad9352405062b27b052db086149dc3a9a96054f3
|
|
Change-Id: Iee3182513369fdb2b38c58407291aa041b3294b5
|
|
Change-Id: I02d2ac056bfecfc4e2f1e292ba7be7c476a4a79f
|
|
Change-Id: Id188be6933455c768aa15fd369a125b63cdf073c
|
|
This doesn't actually check that any optimization was performed, just
that the compiler exits successfully.
Change-Id: I33c6e9d9b6b3fed3420ebd0e1b6d868529837548
|
|
Change-Id: I560bcd005025a151acace682c0453ae9b965151b
|
|
|
|
|
|
|
|
Change-Id: Idb21d74f4d470c8f26b84224fb881a2bf40e9e5d
|
|
Change-Id: Ia88c2f4879e8a10a398b7d86316b45c560cb6049
|
|
wait3 was removed from the NDK
Bug: http://b/26234924
Change-Id: I5c2c1cd0d9097151478936b72ccfffc1976ac4c5
|
|
We haven't been able to finalize a new package layout, and we don't
want to change it twice. Restore the old layout until we can figure
out the new one.
Change-Id: I3082a0354b5f1f34456ba3a769b37257484e5f45
|
|
|
|
https://twitter.com/sonicadvance1/status/679201171781713920?s=03
Apparently gnustl doesn't include std::copysign on LP32 platforms.
Haven't looked in to why yet, but this is the test case I used to
narrow down the bug (since the tweets make terrible bug reports).
Bug: http://b.android.com/74835
Bug: http://b/26318616
Change-Id: I55bc9baf7259646a3322dfee64a189805c0dedcd
|
|
Flag __mips__ is defined for both mips 32 and 64 arch.
Use __mips64 flag to differentiate.
Flag __mips64__ is not exported by GCC.
Change-Id: Iada7bde624d227eebf055ad6a6847bc298e5a2b4
|
|
Change-Id: I8acf4a06cf116c63f85b554acdc3bfcab55cf11b
|
|
Bug: http://b/25981507
Change-Id: Id3b46ff76fe5aff8d3d605669affcc25c22731a4
|
|
The NDK used to support defining `_WCHAR_IS_8BIT` to get an 8-bit
wchar (this was for compatibility with Froyo and earlier releases). We
don't support targeting anything before Gingerbread any more, so I
don't think we need this compatibility flag now (it's already gone
from the NDK sys/cdefs.h).
Change-Id: I13e748a9038b2c7c80d417e3ddb0c6bd9ae57bd5
|
|
mm_malloc.h depends on posix_memalign, which wasn't added until
android-16.
Change-Id: I247d275f9065b397c8e2c64149eefba93360d517
|
|
Previously the test wasn't actually checking for wsbh like it claimed.
We need to check that we emit a wsbh followed by a rotate.
Also, rotr is a synonym for ror that we need to catch.
Change-Id: Ib3a14e1534b00b3d63af222ebf5c56c563d46000
|
|
Clang does not emit rev16 for __builtin_bswap16, but rev + lsr
instead. Previously we were implementing __swap16 ourselves, but now
we're relying on the compiler builtin.
Bug: http://b/26091410
Change-Id: I2c75ceeb8c52fd499d9fba26ce4d7ce5f042df49
|
|
Might use rolw instead of rorw, and bswapl instead of bswap.
Change-Id: Ic06fb1ceca7f7a125d1bb5fbc3c3c22a646d467b
|
|
If we're trying to check that we don't have arm5 definitions in arm7,
test that directly.
Change-Id: I8fd29b56cd37a64e731236203ff136c268872ec1
|
|
Haven't investigated the problem yet. Failures go away with -O0. Seems
to not have been a problem for r10e.
Bug: http://b/26085687
Change-Id: I9fb5f4473e9286059e0ff7c453e87ada0b475dc3
|
|
We're still suffering from the x86_64 long double issues in clang.
Bug: http://b.android.com/177432
Change-Id: I57fe770c6738666347f80f6ccbb19a04e9f7d03b
|
|
Clang supports LTO through gold, and we don't support gold for mips.
Change-Id: Id64f63bf17ea0e85c5e541c4e1187ac09bbe87ac
|
|
Clang LTO is only supported with gold. ARM64 still uses bfd by
default, so make sure this test uses gold when we're using clang.
Change-Id: Iac8080eff23e4fcdd5f7be4e2687d938472843d2
|
|
Clang doesn't know all these builtins.
Change-Id: I89fd731df955685160997e7f5fe1221cc50c8a6a
|
|
Change-Id: I4eec8450302e2d82d1f29c159baaf2e9f5464674
|
|
Change-Id: I81f359d7452ee8902959fee8ee2d4c2e378f58c3
|
|
Change-Id: I0edaff5e393bee10b8fa9c117fe388a24a3f23e8
|
|
Change-Id: Ic90944c8cb0dbb7ca534ee098e8048c0eaba6a50
|
|
Looks like stlport's filter1 is actually only broken with GCC.
Bug: http://b/24403847
Change-Id: Id8823717b2305f2b2d07844c05f28e5bf4fb1c98
|
|
Clang doesn't have this optimization yet.
Bug: http://b/26031639
Change-Id: Ia2d74330fef2213cddc7b5dfe3d6d0a7f7c9fa5f
|
|
The test is making sure that LIBCXX_FORCE_REBUILD works, and I really
don't care about supporting that for ARM5 (and it currently doesn't
work).
Change-Id: I832d9d1d88ff06669c72f03f8ea6907c3aa6c23f
|
|
Change-Id: Icc02192572e392b0ff5acb14e3572ed4ce1e54b3
|
|
Change-Id: Ia79c68a8990df63eb2df6fd081d989e4aa297470
|
|
Doesn't seem to be any reason to build for 18 in the original bug, and
removing that lets us run on more targets.
Change-Id: I2a7fc921aa70edac1f012635336991a6e8d38664
|
|
Bug: http://b/26015756
Change-Id: I1edc0536c08fad8be4dee4ff61b5a692a8ab083a
|
|
Wow. Gingerbread didn't support `mkdir -p` or `rm -f`. Avoid using
either of those.
Change-Id: I9f79613465e8608f89bad9a77c232620a3ca8b4e
|
|
Change-Id: I63cb52c17bef1ae83e75df896e1743239758622d
|
|
Bug: http://b/26014163
Change-Id: I301423c2306be641c813891411f3bd95e128d4a7
|
|
Apparently not all versions of Android supported the getcwd(3)
extension of mallocing memory when passed NULL, so this would create
the string "(null)/libbar.so" on JB.
Change-Id: I0174762c53337577a25893402fe74ae3481539f5
|
|
This test is checking symbols that weren't present before API 18.
Change-Id: Id44a8c7c9654018c6d5bb578dcf84354a0bd3cdc
|
|
This test passes before M.
Bug: http://b/26002885
Change-Id: I92b0cac936e74d0036aeb42d99481ed17e4c389b
|