TAP format requires all results to be reported.
Removed extraneous preferences function (handled by parent)
Incorporated fix from 3d9e7db2e0
Simplified total printing
This means that code such as
```cpp
TEST_CASE() {
SECTION("first") { SUCCEED(); }
auto _ = GENERATE(1, 2);
SECTION("second") { SUCCEED(); }
}
```
will run and report 3 assertions, 1 from section "first" and 2
from section "second". This also applies for greater and potentially
more confusing nesting, but fundamentally it is up to the user to
avoid overly complex and confusing nestings, just as with `SECTION`s.
The old behaviour of `GENERATE` as first thing in a `TEST_CASE`,
`GENERATE` not followed by a `SECTION`, etc etc should be unchanged.
Closes#1938
A test runner already has a --durations option to print durations.
However, this isn't entirely satisfactory.
When there are many tests, this produces output spam which makes it hard
to find the test failure output. Nevertheless, it is helpful to be
informed of tests which are unusually slow.
Therefore, introduce a new option --min-duration that causes all
durations above a certain threshold to be printed. This allows slow
tests to be visible without mentioning every test.
Some compilers, e.g. the Green Hills C++ compiler, react badly to the
appearance of std::exception_ptr, std::current_exception,
std::rethrow_exception and std::uncaught_exception(s). To allow usage of
Catch2 with these compilers when exceptions are disabled, hide the usage
of std::exception_ptr etc. when compiling with
CATCH_CONFIG_DISABLE_EXCEPTIONS.
* Successive executions of the same `GENERATE` macro (e.g. because
of a for loop) no longer lead to multiple nested generators.
* The same line can now contain multiple `GENERATE` macros without
issues.
Fixes#1913
This brings our output inline with GTest's. We do not handle skipped
tests properly, but that should be currently less important than
having the attribute exist with proper value for non-skipped tests.
Thanks @joda-01.
Closes#1899
The issue is caused by deleted `std::__detail::begin` declared in `bits/iterator_concepts.h`. This would be found by ADL, and because it is deleted, compilation would fail. This change makes it so that we SFINAE on `begin(std::declval<T>())` and `end(std::declval<T>())` being well-formed.
D:\vcpkg\toolsrc\include\catch2\catch.hpp(11285): warning C6330: 'char' passed as _Param_(1) when 'unsigned char' is required in call to 'isspace'.
D:\vcpkg\toolsrc\include\catch2\catch.hpp(11288): warning C6330: 'char' passed as _Param_(1) when 'unsigned char' is required in call to 'isspace'.
ISO/IEC 9899:2011:
"7.4 Character handling <ctype.h>"/1
[...] In all cases the argument is an int, the value of which shall be
representable as an unsigned char or shall equal the value of the macro
EOF. If the argument has any other value, the behavior is undefined.
This means if isspace was passed a character like ñ it could corrupt
memory without the static_cast to treat it as a positive value after
integral promotion (and C libraries commonly use the int index supplied
as a key into a table which result in out of bounds access if the
resulting int is negative).
Ideally, clang-tidy would be smart that if one alias of a warning
is suppressed, then the other one is suppressed as well, but as of
right now, it isn't. This means that for now we have to suppress
both aliases of this warning. Opened upstream issue to fix this:
https://bugs.llvm.org/show_bug.cgi?id=45859
Obviously, ideally clang-tidy would also not warn that we are calling
a vararg function when it is an unevaluated magic builtin, but that
also is not happening right now and I opened an issue for it:
https://bugs.llvm.org/show_bug.cgi?id=45860Closes#1921
Doing some benchmarking with ClangBuildAnalyzer suggests that
compiling Catch2's `SelfTest` spends 10% of the time instantiating
`std::unique_ptr` for some interface types required for registering
and running tests.
The lesser compilation overhead of `Catch::Detail::unique_ptr` should
significantly reduce that time.
The compiled implementation was also changed to use the custom impl,
to avoid having to convert between using `std::unique_ptr` and
`Catch::Detail::unique_ptr`. This will likely also improve the compile
times of the implementation, but that is less important than improving
compilation times of the user's TUs with tests.
This simplified variant supports only a subset of the functionality
in `std::unique_ptr<T>`. `Catch::Detail::unique_ptr<T>` only supports
single element pointer (no array support) with default deleter.
By removing the support for custom deleters, we also avoid requiring
significant machinery to support EBO, speeding up instantiations of
`unique_ptr<T>` significantly. Catch2 also currently does not need
to support `unique_ptr<T[]>`, so that is not supported either.
As far as I know, only a few users actually use it, but these changes
allow us to avoid including a surprising amount of code in the main
compilation path.
As part of `-Wdeprecated-copy-dtor` sweep, I noticed that Capturer's
copies are defaulted. Given that the class would likely break horribly
in the event of actual copy happening, they are now deleted.
This should improve the compilation times by decreasing the number
of TUs compiled, without making overly big TUs that would cause
problems with heavy-tailed compilation times.
There is one "combined TU" for the top level part, and each subpart,
except for Reporters, which currently do not have any trivial TUs.
Special member functions are now implicit, which should make them
both noexcept and constexpr. The `operator<<` has been made into
hidden friend as per best practices.
As Clara is no longer maintained as a separate project, the
implementation was moved to the internal subfolder of top-level
folder. This removes one folder and avoids potential user confusion.
Also simplified the convenience header checking script accordingly.
There are two reasons for this:
1) It is highly unlikely that someone has use for this header,
which has no customization points and only provides simplest
possible main, and cannot link the static library which also
provides a default main implementation.
2) It being a header was causing extra complications with
the convenience headers, and our checking script. This would either
require special handling in the checking script, or would break user's
of the main convenience header.
All in all, it is simpler and better in the long term to remove it,
than to fix its problems.
I do not think we need a safeguard against not including files in
CMake anymore, and as it is, it caused annoying false positive about
the default main implementation.
The naming scheme is simple, to include all matchers, include header
`catch2/matchers/catch_matchers_all.hpp`. To include **everything**,
include `catch2/catch_all.hpp`.
This describes the reality better, as it also links in the rest
of Catch2.
The on-disk name of the static library remains just `Catch2Main`,
as that is what it is -- single main function -- and on-disk artifacts
cannot describe link dependencies.
Catch assumes std::uncaught_exceptions is available whenever C++17 is
available, but for macOS versions older than 10.12 this is not the case.
Instead of checking the C++ version, use a macro to check whether the
feature is available.
Previously a random test ordering was obtained by applying std::shuffle
to the tests in declaration order. This has two problems:
- It depends on the declaration order, so the order in which the tests
will be run will be platform-specific.
- When trying to debug accidental inter-test dependencies, it is helpful
to be able to find a minimal subset of tests which exhibits the issue.
However, any change to the set of tests being run will completely
change the test ordering, making it difficult or impossible to reduce
the set of tests being run in any reasonably efficient manner.
Therefore, change the randomization approach to resolve both these
issues.
Generate a random value based on the user-provided RNG seed. Convert
every test case to an integer by hashing a combination of that value
with the test name. Sort the test cases by this integer.
The test names and RNG are platform-independent, so this should be
consistent across platforms. Also, removing one test does not change
the integer value associated with the remaining tests, so they remain in
the same order.
To hash, use the FNV-1a hash, except with the basis being our randomly
selected value rather than the fixed basis set in the algorithm. Cannot
use std::hash, because it is important that the result be
platform-independent.
It did not clear out all of its internal state when switching from
one pattern to another, so when it should've escaped `,`, it took
its position from its position in the original user-provided string,
rather than its position in the current pattern.
Fixes#1905
CATCH_INTERNAL_IGNORE_BUT_WARN() introduced with b7b346c triggers
clang-tidy warning 'cppcoreguidelines-pro-type-vararg' for every usage
of assertion macros like CHECK() and REQUIRE(). Silence it via NOLINT
in the '#if defined(__clang__)' block only, as clang-tidy honors those.
The old code caused warnings to fire under MSVC, and Clang <3.8.
I could not find a GCC version where it worked, but I assume that it
did at some point.
This new code causes all of MSVC, GCC, Clang, in current versions,
to emit signed/unsigned comparison warning in test like this:
```cpp
TEST_CASE() {
int32_t i = -1;
uint32_t j = 1;
REQUIRE(i != j);
}
```
Where previously only MSVC would emit the warning.
Fixes#1880
These files are not included by the default
`#include <catch2/catch_test_macros.hpp>` path, so that users do
not have to pay for them if they do not use them. Follow up is to
split out the small part of `catch_preprocessor.hpp` used by the
default test macros (AFAIK, it is just `INTERNAL_CATCH_REMOVE_PARENS`
macro), so that it is not included by the default path either.
Also fixes#1892 by providing the missing macros.
This is both a really big and a really small commit. It is small in
that it only contains renaming, moving and modification of include
directives caused by this.
It is really big in the obvious way of touching something like 200
files.
The new rules for naming files is simple: headers use the `.hpp`
extension. The rules for physical file layout is still kinda in
progress, but the basics are also simple:
* Significant parts of functionality get their own subfolder
* Benchmarking is in `catch2/benchmark`
* Matchers are in `catch2/matchers`
* Generators are in `catch2/generators`
* Reporters are in `catch2/reporters`
* Baseline testing facilities are in `catch2/`
* Various top level folders also contain `internal` subfolder,
with files that users probably do not want to include directly,
at least not until they have to write something like their own
reporter.
* The exact files in these subfolders is likely to change later
on
Note that while some includes were cleaned up in this commit, it
is only the low hanging fruit and further cleanup using automatic
tooling will happen later.
Also note that various include guards, copyright notices and file
headers will also be standardized later, rather than in this commit.
It used to be a file that would collect interfaces we always wanted
to provide to users, so that the single header stitching script
would place them in the common part of the single header version.
As v3 is moving to separate headers model, the file is no longer
useful.
This was an old "include all" header, that we no longer want to be
usable, to make the include differences in new versions explicit.
We will introduce new "include all" headers later, in the form of
`catch_all.hpp`, `catch_matchers_all.hpp` and so on...
Removed nested `StdString` namespace and added Doxygen comments.
Also renamed some matchers to avoid colisions now that there are
less separate namespaces for matchers to go to. Since this is a
breaking release anyway, it shouldn't matter, and the factory
functions that the users should use remain the same anyway.
Removed the `generic` nested namespace, so PredicateMatcher now
lives in `Catch::Matchers` namespace, just like other matchers.
Also cleaned up and doxygenized comments on the `Predicate` factory
function for `PredicateMatcher`.
The two changes are
`catch_matchers_templates` -> `catch_matchers_templated` and
`catch_matchers_generic` -> `catch_matchers_predicate`. The former
is mostly cosmetic, but the second was previously significantly
misleading, and as the library is now to be consumed by including
specific headers, this needed to be fixed.
`SizeIs` can accept both `size_t` and a matcher. In the first case,
it checks whether the size of the range is equal to specified size.
In the second case, it checks whether the provided matcher accepts
the size of the range.
Outside of `MatcherBase` and `GenericMatcherBase`, matchers are not
designed to be overriden. This means that doing so can easily lead
to errors, and matchers are generally fairly simple functionality-wise.
so there is not much code reuse to be gained anyway.
Thus, Catch2-provided concrete matchers are now final.
In general, for Catch2 v3 we are making virtual types `final`,
unless they were explicitly designed to be derived-from.
`ListeningReporter` is definitely not designed to be derived-from.
This was previously used to avoid `dynamic_cast` inside our code,
when we were creating more than one reporter, or a reporter
together with listeners. However, since then the offending code
was refactored to be smarter instead, and this query member function
is no longer needed nor used.
The only way to stream those is to use the `bool` overload of `op<<`.
However, to convert a function to bool, GCC creates AST equivalent
of `A? true : false`. Then, because `A` is a function, it warns that
it will never be `false`. 🤦
As a bonus, newer GCC versions issue _two_ different warnings about
this, but older GCC versions do not know both of them, so we also
have to suppress warning about unknown warning suppression.
The variable initialization has test registration as a side-effect,
but as far as GCC is concerned, the variable itself is unused.
Because the macro substitution always happens at global scope, we
cannot use cast to `void` as is usually done.
Sadly most versions still cannot properly handle the suppression
via `_Pragma`, so it has to leak to the users when they use older
GCC versions to compile their code
In the future we can expect many more matchers, so let's give them
a place to live.
Also moved matcher-related internal files to `internal` subfolder.
Ideally we should sort out all of our source code, but that will
have to come later.
This commit also forbids composing lvalues of composed matchers, as
per previous deprecation notice. I do not expect this to be contentious
in practice, because there was a bug in that usage for years, and
nobody complained.
Given that in the 2 or so years that matchers are thing nobody complained,
it seems that people do not actually write this sort of code, and the
possibility will be removed in v3. However, to avoid correctness bugs,
we will have to support this weird code in v2.
C++11 math requires _GLIBCXX_USE_C99_MATH_TR1 to be true with gcc/clang.
Also fixes an issue with uClibc-ng where __UCLIBC__ is defined in features.h but
that is not included here and is thus no-op.
- Overrides added
- usages of push_back() replaced with emplace_back()
- Loop variable made const-refernce
- NULL replaced with nullptr
- Names used in the declaration and definition unified
- size() replaced with empty
- Identical cases merged
This commit extends the Matchers feature with the ability to have type-independent (e.g. templated) matchers. This is done by adding a new base type that Matchers can extend, `MatcherGenericBase`, and overloads of operators `!`, `&&` and `||` that handle matchers extending `MatcherGenericBase` in a special manner.
These new matchers can also take their arguments as values and non-const references.
Closes#1307Closes#1553Closes#1554
Co-authored-by: Martin Hořeňovský <martin.horenovsky@gmail.com>
b77cec05c0 fixed this problem for tagging tests, so that a test
case tagged with `[.foo]` would be parsed as tagged with `[.][foo]`.
This does the same for the test spec parsing.
Fixes#1798
Copying a `ReusableStringStream` would lead to "double free" of
the stream, and thus it could be used in multiple places at the
same time, breaking the output.
On systems where std::chrono::steady_clock::period is not std::nano, benchmark tests fail to compile due to trying to convert analysis.samples from a vector of duration<double, clock::period> to a vector of std::chrono::duration<double, std::nano>.
Its intent was to show which headers are expected to be useable by
Catch2's users, and to enforce their inclusion in the single header
distribution at the right place.
Given the new library model, the second use case is not needed and
the first one is better served with documentation and physical file
layout.
Now that Catch2 is a proper library, we can always build the full
library (comparatively minor slowdown) and the user can avoid
including benchmarking headers to avoid the compilation slowdown.
The use we previously used the polyfill or naked new is that we
supported C++11, which did not yet have `std::make_unique`. However,
with the move to C++14 as the minimum, `std::make_unique` can be
expected to be always available.
Because some of the tooling used by Catch2 does not properly support
version postfixes, such as `preview-1`, we will report the
in-development version is `v3.0.0`, and the first real release will
have to be `v3.0.1`.
Closes#1824
Now that the recommended distribution and usage method is proper
library, users can just avoid including the matcher headers to get
basically the same effect.
