When the client exits (e.g. with Ctrl+C) the subprocess should be
notified as well so it can cleanup if needed.
Bug: http://b/23825725
Change-Id: Idb771710b293e0a9f7bebc9e2814b3a816e2c50e
It is reported that the registered fd can be bigger than FD_SETSIZE, and can't be
handled by select(). By moving to poll(), we can remove the limitation.
Although we can't ignore the possibility that there is a fd leak, but we can
still make the potential bug more explicit by moving to poll().
We didn't move to epoll() because it is not supported on mac.
Bug: 23820751
Change-Id: Icb39329c4984f1fef749472c9e088682ee8c3444
Some adb commands use the shell service but weren't using the
shell protocol, causing a protocol mismatch between the client and
the device. This CL uses the shell protocol whenever possible.
Change-Id: I4c9c75b7fd1d5bf3dc1f73900711840270527682
When calculating the crash_sender spread time, make sure that the
random number is not negative when converted to a shell int variable.
Bug: 24004011
Change-Id: I3b95dc244a26270ef2b93d5af4b0593a93eedcad
This test case ensures that we can log a metric with the metrics library
and upload it with the upload service.
BUG: 22879646
TEST: unit tests.
Change-Id: Idf4a1240d41745ebf7835089230608082eed19f5
If the metrics server is unreachable, we may need to resend a metrics
report later. Instead of keeping the staged report in memory, save it to
disk to avoid loosing data if the system restarts or crashes.
BUG: 23033262
TEST: unit tests.
Change-Id: Idd14964e40f022952469f47d675d8cda9586d7cd
Instead of passing different filenames around, only rely on the metrics
directory and infer the filepath from it. This makes testing easier.
BUG: 23939404
TEST: unit tests.
Change-Id: I79086acc3a546464114fa8ec4656ec04e1c43e35
Read the crash server URL from the crash_reporter.server
property. If it is not set, return a configuration error.
Bug: 22874192
Change-Id: Iac341b6352fe9c1b54cd2e8561ed4a5bbe8ddddc
- Added grep to the list of required modules.
- Run crash_sender with the 'system' group to allow calls to
metrics_client to read both metrics files owned by system and
crash_reporter files owned by root.
- Fix periodic_scheduler's check delay to actually be 5 minutes.
Bug: 23122375
Change-Id: Iff214c4e591a676a516162a92ea5aedcf0824f46
Add to whitelist.
For some reason, when compiling, parseint_test.cpp would encounter
errors like the following:
host cross C++: host_cross_libbase_test <= system/core/base/parseint_test.cpp
In file included from external/gtest/include/gtest/gtest.h:1929:0,
from system/core/base/parseint_test.cpp:19:
system/core/base/parseint_test.cpp: In member function ‘virtual void parseint_signed_smoke_Test::TestBody()’:
external/gtest/include/gtest/internal/gtest-internal.h:133:55: error: converting ‘false’ to pointer type for argument 1 of ‘char testing::internal::IsNullLiteralHelper(testing::internal::Secret*)’ [-Werror=conversion-null]
(sizeof(::testing::internal::IsNullLiteralHelper(x)) == 1)
^
external/gtest/include/gtest/gtest_pred_impl.h:77:52: note: in definition of macro ‘GTEST_ASSERT_’
if (const ::testing::AssertionResult gtest_ar = (expression)) \
^
external/gtest/include/gtest/gtest_pred_impl.h:166:3: note: in expansion of macro ‘GTEST_PRED_FORMAT2_’
GTEST_PRED_FORMAT2_(pred_format, v1, v2, GTEST_FATAL_FAILURE_)
^
external/gtest/include/gtest/gtest.h:1993:3: note: in expansion of macro ‘ASSERT_PRED_FORMAT2’
ASSERT_PRED_FORMAT2(::testing::internal:: \
^
external/gtest/include/gtest/gtest.h:1994:32: note: in expansion of macro ‘GTEST_IS_NULL_LITERAL_’
EqHelper<GTEST_IS_NULL_LITERAL_(expected)>::Compare, \
^
external/gtest/include/gtest/gtest.h:2011:32: note: in expansion of macro ‘GTEST_ASSERT_EQ’
# define ASSERT_EQ(val1, val2) GTEST_ASSERT_EQ(val1, val2)
^
system/core/base/parseint_test.cpp:23:3: note: in expansion of macro ‘ASSERT_EQ’
ASSERT_EQ(false, android::base::ParseInt("x", &i));
^
For whatever reason, the gcc for Windows isn't handling the template
metaprogramming correctly, so to work-around the issue, I
search-and-replaced "ASSERT_EQ(false, " to "ASSERT_FALSE(" and
"ASSERT_EQ(true, " to "ASSERT_TRUE(" and that seems to compile just fine
for Linux and Windows (and the tests pass on both platforms).
Change-Id: I05132909d1f4b18afff23139652a218649689f2d
Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
Adds missing #ifdef guards to shell_service_protocol_test.cpp so the
test compiles on Windows.
Also fixes a bug in Windows socketpair write implementation. Previously
it was only checking for a closed pipe if the write happened to block.
This adds an additional pre-check to exit immediately on a closed pipe.
These two changes allow the test to compile and pass on Windows.
Change-Id: Ib8853ed72f015fc0d623da47c32982cb3ffa4a3d
