catch3/src/catch2/benchmark/catch_benchmark.hpp, branch v3.4.0 Progmasoft Catch3: an independently maintained C++17 test framework with a parallel runner, property checks, snapshots and structured results. 🛠 Add <exception> header where strictly necessary 2023-05-16T16:18:57+00:00 ThePhD 2023-05-16T16:13:49+00:00 05d7eb5a0028906e4c6270b2340ae0d88d3cfa2a 






Cleanup internal includes by splitting out some event structs
2023-03-31T17:31:51+00:00

Martin Hořeňovský

2023-03-31T08:44:29+00:00

9f08097f55eb29a61efcdb4af41952a67e155215

* Split out BenchmarkInfo and BenchmarkStats to their own header
* Outline BenchmarkStats<> declaration to separate header
* Split out TestRunInfo into its own header

These changes let us remove the large `interfaces_reporter.hpp`
include from `benchmark.hpp`, and replace it with
`interfaces_capture.hpp` in `run_context.hpp`.

I also cleaned out `interfaces_repoter.hpp` from reporter headers
that depend on `reporter_common_base.hpp`. This will not change
anything in the actual inclusion set, but makes it logically
more consistent.





* Split out BenchmarkInfo and BenchmarkStats to their own header
* Outline BenchmarkStats<> declaration to separate header
* Split out TestRunInfo into its own header

These changes let us remove the large `interfaces_reporter.hpp`
include from `benchmark.hpp`, and replace it with
`interfaces_capture.hpp` in `run_context.hpp`.

I also cleaned out `interfaces_repoter.hpp` from reporter headers
that depend on `reporter_common_base.hpp`. This will not change
anything in the actual inclusion set, but makes it logically
more consistent.







Cleanup benchmarking includes
2023-03-23T10:59:40+00:00

Martin Hořeňovský

2023-03-23T10:59:40+00:00

b3ebce715e9686f664e2756043a4c0e0e387a112












Move name and sample analysis in benchmarks into BenchmarkStats
2023-01-28T20:40:59+00:00

Martin Hořeňovský

2023-01-28T20:40:51+00:00

356dfc14391183aef80af7f23fe2297485311ae4

This always saves 1 allocation per benchmark, and another two
allocations if the benchmark name is longer than the SSO buffer.





This always saves 1 allocation per benchmark, and another two
allocations if the benchmark name is longer than the SSO buffer.







Fix references to license file
2022-10-28T09:30:15+00:00

Lars Toenning

2022-10-28T09:22:53+00:00

f1084fb30990fe40c8f5b70034970c7eb4e1da72

The license file was renamed with 6a502cc2f5fedd59b3495b58708f0d6d987ed9e1





The license file was renamed with 6a502cc2f5fedd59b3495b58708f0d6d987ed9e1







Added --skip-benchmarks command-line option. (#2408)
2022-04-23T21:14:49+00:00

Daniel Feist

2022-04-23T21:14:49+00:00

291b35b389b8e6d7b560cf7e43b6c72c4b73e522

Closes #2392

Co-authored-by: Martin Hořeňovský <martin.horenovsky@gmail.com>




Closes #2392

Co-authored-by: Martin Hořeňovský <martin.horenovsky@gmail.com>






Prefer _t over ::type form for std type traits
2021-08-18T22:21:13+00:00

Martin Hořeňovský

2021-08-16T15:01:00+00:00

edad4d0af7a91693448bbf884236ea6ff2a4bd15












Replace uses of std::move and std::forward with macros
2021-08-16T14:22:19+00:00

Martin Hořeňovský

2021-08-13T12:16:30+00:00

d2ee7100d26128b3956571cc802f3a52c9b9aea8

This improves the SelfTest build times by about 3% (measured
with Clang 10 on random Linux box I had lying around).





This improves the SelfTest build times by about 3% (measured
with Clang 10 on random Linux box I had lying around).







Remove superfluous catch_enforce include from complete_invoke
2021-08-11T20:52:04+00:00

Martin Hořeňovský

2021-08-09T22:38:23+00:00

2b54f1e7a63f80098f354b6ff186f47a8e6da6fa

Also added missing transitive include to `catch_benchmark.hpp`.





Also added missing transitive include to `catch_benchmark.hpp`.







Fix inconsistencies in reporting benchmarking failures
2021-08-07T13:16:11+00:00

Martin Hořeňovský

2021-06-09T18:48:58+00:00

0a3f511cfed071c9c4ad512341ae7569383922b7

With these changes, all these benchmarks
```cpp
BENCHMARK("Empty benchmark") {};
BENCHMARK("Throwing benchmark") {
    throw "just a plain literal, bleh";
};
BENCHMARK("Asserting benchmark") {
    REQUIRE(1 == 2);
};
BENCHMARK("FAIL'd benchmark") {
    FAIL("This benchmark only fails, nothing else");
};
```

report the respective failure and mark the outer `TEST_CASE` as
failed. Previously, the first two would not fail the `TEST_CASE`,
and the latter two would break xml reporter's formatting, because
`benchmarkFailed`, `benchmarkEnded` etc would not be be called
properly in failure cases.





With these changes, all these benchmarks
```cpp
BENCHMARK("Empty benchmark") {};
BENCHMARK("Throwing benchmark") {
    throw "just a plain literal, bleh";
};
BENCHMARK("Asserting benchmark") {
    REQUIRE(1 == 2);
};
BENCHMARK("FAIL'd benchmark") {
    FAIL("This benchmark only fails, nothing else");
};
```

report the respective failure and mark the outer `TEST_CASE` as
failed. Previously, the first two would not fail the `TEST_CASE`,
and the latter two would break xml reporter's formatting, because
`benchmarkFailed`, `benchmarkEnded` etc would not be be called
properly in failure cases.