This is not thread safe, but I think that was already true of Catch.
The construction/destruction of the std::ostringstream is where the
vast majority of time is spent per assertion. A simple test of
100000000 CHECK()s is reduced from around 60s to 7.4s
By using char const * instead of std::string we avoid significant
copying per assertion. In a simple loop with 10000000 CHECKS on
my system, this reduces the run time from 9.8s to 6s.
This prevents Catch from disabling `Wpadded` for Clang inside test files
(files that do not define either `CATCH_CONFIG_MAIN` or
`CATCH_CONFIG_RUNNER`).
catch_suppress_warnings.h and catch_reenable_warnings.h should be
included only once*, so that the stitching script includes them as the
first and last header respectively, since it only includes each header
once. This caused a bug, where the first one was included properly, but
the second one was included prematurely, from catch_xmlwriter.hpp, and
thus was guarded by `CATCH_IMPL`.
* At least until the stitching script is changed to accomodate common
warning disabling header.
Fixes#871
Types which are truthy, but have more information than the truthiness in their string conversion were showing up as 'true' or 'false' instead of showing the underlying type's string value.
Previously, this would not print out any messages for the last CHECK
```cpp
TEST_CASE("Foo") {
INFO("Test case start");
for (int i = 0; i < 2; ++i) {
INFO("The number is " << i);
CHECK(i == 0);
}
CHECK(false);
}
```
now it does.
- typedefs long long for MSVC
- typedefs uint64_t otherwise
Should probably do finer grained compiler checking - but this should at least be better than what was there before
When using C++11, comparison operators are already templated to take
anything that can be explicitly converted to double, but constructor
took only doubles. This lead to warnings when an `Approx` was
constructed from floats, which was problematic for some users.
Since just adding float constructor would be a large breaking change, as
suddenly `Approx( 1 )` would become ambiguous, I added a templated
constructor that will take anything that is explicitly convertible to
double. This has the added benefit of allowing constructing `Approx`
instances from instances of strong typedefs, ie allowing
`calculated_temp == Approx( known_temp)`.
Closes#873
Unexpected exceptions no longer cause abort and there should be no more
potential for false negatives.
The trade-off now is that exceptions are no longer translated.
This is another warning that follows test macros, making it painful to
suppress without leaking outside. Luckily clang's `_Pragma`
implementation works.
Should fix#308
Effectively a revert of previous commit, fixing #542, where this was
added to stop linters complaining about `REQUIRE_THROWS_AS` used like
`REQUIRE_THROWS_AS(expr, std::exception);`, which would be slicing the
caught exception. Now it is user's responsibility to pass us proper
exception type.
Closes#833 which wanted to add `typename`, so that the construct works
in a template, but that would not work with MSVC and older GCC's, as
having `typename` outside of a template is allowed only from C++11
onward.
This seems to give about 15% speedup when compiling tests using GCC.
The tradeoff is that under certain circumstances, there is a chance for
false negative result, when the expression under test throws exception
and the test code catches it before it gets to the test runner.
Example:
``` cpp
TEST_CASE("False negative") {
try {
REQUIRE(throws() == "");
} catch (...) {}
}
```
This test case will succeed, reporting no assertions checked, instead of
failing as it would with `CATCH_CONFIG_FAST_COMPILE` disabled. However,
just removing the try-catch block inside client's code will fix this, so
it is worthwhile.
This change does not apply to CHECK* macros, because these are currently
specified as continuing on exception and thus need the local try-catch
to work as intended.
