type_name_as_string_test fails

[yurivict]

Does this problem persist on the current main?

• I have verified the issue on the current main

Is there an existing issue for this?

• I have searched the existing issues

Current Behavior

$ ./work/seqan3-3.4.2/.test/utility/detail/type_name_as_string_test
Running main() from /usr/ports/biology/seqan3/work/seqan3-3.4.2/.test/_deps/gtest-src/googletest/src/gtest_main.cc
[==========] Running 7 tests from 7 test suites.
[----------] Global test environment set-up.
[----------] 1 test from type_inspection/0, where TypeParam = char
[ RUN      ] type_inspection/0.type_name_as_string
[       OK ] type_inspection/0.type_name_as_string (0 ms)
[----------] 1 test from type_inspection/0 (0 ms total)

[----------] 1 test from type_inspection/1, where TypeParam = char16_t
[ RUN      ] type_inspection/1.type_name_as_string
[       OK ] type_inspection/1.type_name_as_string (0 ms)
[----------] 1 test from type_inspection/1 (0 ms total)

[----------] 1 test from type_inspection/2, where TypeParam = char32_t
[ RUN      ] type_inspection/2.type_name_as_string
[       OK ] type_inspection/2.type_name_as_string (0 ms)
[----------] 1 test from type_inspection/2 (0 ms total)

[----------] 1 test from type_inspection/3, where TypeParam = short*
[ RUN      ] type_inspection/3.type_name_as_string
[       OK ] type_inspection/3.type_name_as_string (0 ms)
[----------] 1 test from type_inspection/3 (0 ms total)

[----------] 1 test from type_inspection/4, where TypeParam = double const*
[ RUN      ] type_inspection/4.type_name_as_string
[       OK ] type_inspection/4.type_name_as_string (0 ms)
[----------] 1 test from type_inspection/4 (0 ms total)

[----------] 1 test from type_inspection/5, where TypeParam = N3foo3barIJcEEE
[ RUN      ] type_inspection/5.type_name_as_string
/usr/ports/biology/seqan3/work/seqan3-3.4.2/test/unit/utility/detail/type_name_as_string_test.cpp:62: Failure
Expected equality of these values:
  seqan3::detail::type_name_as_string<TypeParam>
    Which is: "N3foo3barIJcEEE (abi::__cxa_demangle error status (-2): mangled_name is not a valid name under the C++ ABI mangling rules.)"
  this->expected_name()
    Which is: "foo::bar<char> const &"

[  FAILED  ] type_inspection/5.type_name_as_string, where TypeParam = N3foo3barIJcEEE (0 ms)
[----------] 1 test from type_inspection/5 (0 ms total)

[----------] 1 test from type_inspection/6, where TypeParam = N3foo3barIJNS0_IJcdEEEEEE
[ RUN      ] type_inspection/6.type_name_as_string
/usr/ports/biology/seqan3/work/seqan3-3.4.2/test/unit/utility/detail/type_name_as_string_test.cpp:62: Failure
Expected equality of these values:
  seqan3::detail::type_name_as_string<TypeParam>
    Which is: "N3foo3barIJNS0_IJcdEEEEEE (abi::__cxa_demangle error status (-2): mangled_name is not a valid name under the C++ ABI mangling rules.)"
  this->expected_name()
    Which is: "foo::bar<foo::bar<char, double>>"

[  FAILED  ] type_inspection/6.type_name_as_string, where TypeParam = N3foo3barIJNS0_IJcdEEEEEE (0 ms)
[----------] 1 test from type_inspection/6 (0 ms total)

[----------] Global test environment tear-down
[==========] 7 tests from 7 test suites ran. (0 ms total)
[  PASSED  ] 5 tests.
[  FAILED  ] 2 tests, listed below:
[  FAILED  ] type_inspection/5.type_name_as_string, where TypeParam = N3foo3barIJcEEE
[  FAILED  ] type_inspection/6.type_name_as_string, where TypeParam = N3foo3barIJNS0_IJcdEEEEEE

 2 FAILED TESTS

Expected Behavior

n/a

Steps To Reproduce

build + test

Environment

- Operating system: FreeBSD 15 STABLE
- SeqAn version: 3.4.2
- Compiler: clang-19

Anything else?

No response

[eseiler]

Confirmed.
I'll check if it can be fixed somehow. If not, we probably just disable the tests for FreeBSD.

[ RUN      ] type_inspection/6.type_name_as_string
/root/seqan3/test/unit/utility/detail/type_name_as_string_test.cpp:62: Failure
Expected equality of these values:
  seqan3::detail::type_name_as_string<TypeParam>
    Which is: "N3foo3barIJNS0_IJcdEEEEEE (abi::__cxa_demangle error status (-2): mangled_name is not a valid name under the C++ ABI mangling rules.)"
  this->expected_name()
    Which is: "foo::bar<foo::bar<char, double>>"

[  FAILED  ] type_inspection/6.type_name_as_string, where TypeParam = N3foo3barIJNS0_IJcdEEEEEE (0 ms)

GoogleTest apparently also has trouble demangling (TypeParam = N3foo3barIJNS0_IJcdEEEEEE).

[eseiler]

Ok, so FreeBSD's compiler-rt is different, and its __cxa_demangle doesn't handle variadic template parameters.

Nested templates for

namespace foo
{
template <typename t>
struct bar
{};
}

demangle just fine.

Nested templates for

namespace foo
{
template <typename... t>
struct bar
{};
}

do not.

I'll adjust the test cases accordingly.