Use a tagged constructor for FlatTuple instead.
Some versions of MSVC are getting confused with that constructor and generating invalid code.
PiperOrigin-RevId: 342050957
Fix *_THROW tests under MSVC, which produce slightly different type
names from std::type_info (e.g. "class std::runtime_error" instead of
"std::runtime_error")
PiperOrigin-RevId: 341874178
Improve DoubleNearPredFormat output on bad epsilons
DoubleNearPredFormat will happily accept epsilon values (abs_error) that
are so small that they are meaningless. This turns EXPECT_NEAR into a
complicated and non-obvious version of EXPECT_EQ.
This change modifies DoubleNearPredFormat) so that when there is a
failure it calculates the smallest meaningful epsilon value, given the
input values, and then prints a message which explains what happened.
If a true equality test is wanted either pass a literal 0.0 as abs_error
or use EXPECT_EQ. If a check for being almost equal is wanted consider
using EXPECT_DOUBLE_EQ which, contrary to its name, verifies that the
two numbers are *almost* equal (within four ULPs).
With this change the flaky test mentioned in crbug.com/786046 gives this
output:
The difference between 4.2934311416234112e+18 and 4.2934311416234107e+18 is 512, where
4.2934311416234112e+18 evaluates to 4.2934311416234112e+18,
4.2934311416234107e+18 evaluates to 4.2934311416234107e+18.
The abs_error parameter 1.0 evaluates to 1 which is smaller than the minimum distance between doubles for numbers of this magnitude which is 512, thus making this EXPECT_NEAR check equivalent to EXPECT_EQUAL. Consider using EXPECT_DOUBLE_EQ instead.
Tested:
I confirmed that this change detects the bad epsilon value that caused
crbug.com/786046 in Chromium and added a test for the desired output.
PiperOrigin-RevId: 332946880
Add millisecond precision to start timestamp in XML/JSON output
- Previous timestamp had format YYYY-MM-DDThh:mm:ss, now YYYY-MM-DDThh:mm:ss.sss
- This conforms to the ISO 8601 standard
PiperOrigin-RevId: 329503623
This change updates testing::internal::IsAProtocolMessage to return true not
just for full proto messages but also for lite ones (i.e. those inheriting
directly from MessageLite).
PiperOrigin-RevId: 304286535
Add --gtest_fail_fast support to googletest.
- Analogous functionality to to golang -test.failfast and python --failfast
- Stops test execution upon first test failure.
- Also add support Bazel equivalent env var (TESTBRIDGE_TEST_RUNNER_FAIL_FAST)
PiperOrigin-RevId: 302488880
This reverts commit 20b5b8ecc7a81d23b4716e22a2b35fd53379c8c6.
Reason for revert: Breaks existing applications, such as ANGLE
(angleproject.org), requires adding an extra prefix that needs to be
typed for no reason (when testing from command line) and increases the
logs' sizes by a non-trivial amount due to the very large number of
tests ANGLE runs.
Original commit message:
Add a compile time check to ensure that the parameters to TEST_P and
INSTANTIATE_TEST_SUITE_P are not empty. Some compilers already fail
in that case and, even where it works, it's likely to result in
technically invalid code by virtue of creating reserved identifiers:
https://en.cppreference.com/w/cpp/language/identifiers
First, every project is perfectly capable of adding a prefix if they
want to support such a compiler. This change penalizes every
project.
Second, using a prefix such as `_p` also results in reserved
identifiers, so this change is not really solving the problem. For that
matter, instead of generating `gtest_##prefix##...`, you can generate
`gtest_x##prefix##...` to correctly fix the issue, including when empty
prefixes are used.
Change googletest to notice failures during SetUpTestSuite() and TearDownTestSuite().
Previously, errors that occurred during those functions were logged but otherwise ignored. After this change, such failures will cause the test to fail and a brief summary will be printed at the bottom of the test log.
See https://github.com/google/googletest/issues/2330.
PiperOrigin-RevId: 284033342
Use standard C++11 integer types in gtest-port.h.
Remove testing::internal::{Int,Uint}{32,64} in favor of types
guaranteed to be in <cstdint> since C++11.
Tests for built-in integer type coverage are switched from
{Int,Uint}64 to [unsigned] long long, which is guaranteed by
C++11 to exist and be at least 64-bit wide.
PiperOrigin-RevId: 281565263
Mark more functions with "override" keyword, just like
it was done in commit 2460f97152c.
This should prevent compiler from complaining while compiling both
user code, and the googletest code itself with the -Wsuggest-override
option turned on; with the exception of:
* calls to new MOCK_METHOD() in test/gmock-function-mocker_test.cc
* calls to old MOCK_METHODx()/MOCK_CONST_METHODx() in other
unit test files.
Closes#2493
Add a compile time check to ensure that the parameters to TEST_P and INSTANTIATE_TEST_SUITE_P are not empty. Some compilers already fail in that case and, even where it works, it's likely to result in technically invalid code by virtue of creating reserved identifiers:
https://en.cppreference.com/w/cpp/language/identifiers
PiperOrigin-RevId: 273832263
Add GTEST_FAIL_AT, an equivalent to ADD_FAILURE_AT but that
spawns a fatal failure rather than a non-fatal, eg. the equivalent
of an ASSERT* failure rather than an EXPECT* failure.
PiperOrigin-RevId: 244746609
Cast some values as their unsigned equivalents or `size_t` to match the
parameter type used for the template object under test. Also, provide
UInt32 equivalent delegate methods for some callers (with
int-equivalents for backwards compatibility).
This closes#2146.
Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>
