sycl-cts/test_queue fails after 46e49ecb70f84ee9f6880d6e89d40657787b3189

[KornevNikita]

Describe the bug

We didn't run Nightly for the last week because it's broken, so I noticed it only now after running the testing manually on my machine.
Output:

-------------------------------------------------------------------------------
queue shortcuts unified shared memory core
-------------------------------------------------------------------------------
../tests/queue/queue_shortcuts_usm_core.cpp:30
...............................................................................

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

../tests/queue/queue_shortcuts_usm.h:149: FAILED:
  CHECK( (t_test == h_actual[i]) )
with expansion:
  false
with message:
  for type "custom struct": 

===============================================================================
test cases:   36 |   34 passed |  1 failed | 1 skipped
assertions: 2803 | 2788 passed | 15 failed

To reproduce

No response

Environment

No response

Additional context

No response

[KornevNikita]

@konradkusiak97 do you know possible reason?

[KornevNikita]

This "custom target" is:
https://github.com/KhronosGroup/SYCL-CTS/blob/SYCL-2020/tests/common/type_list.h#L61

namespace user_def_types {
// A user-defined struct with several scalar member variables, no constructor,
//  destructor or non-special member functions.
struct no_cnstr {
  float a;
  int b;
  char c;

  constexpr auto &operator=(const int &v) {
    a = v;
    b = v;
    c = v;
    return *this;
  }

  friend bool operator==(const no_cnstr &lhs, const no_cnstr &rhs) {
    return ((lhs.a == rhs.a) && (lhs.b == rhs.b) && (lhs.c == rhs.c));
  }

  friend bool operator!=(const no_cnstr &lhs, const no_cnstr &rhs) {
    return !(lhs == rhs);
  }

  friend no_cnstr operator+(const no_cnstr &lhs, int i) {
    return {lhs.a + static_cast<float>(i), lhs.b + i,
            static_cast<char>(lhs.c + i)};
  }
};

[keyradical]

Thanks for reporting @KornevNikita, the patch that possibly broke this will be reverted with #13675

[KornevNikita]

@konradkusiak97 thanks for the quick feedback. Could you please ping me when you will push the new PR so we can verify, that this test also passes?

[keyradical]

@KornevNikita The revert PR was merged, can you verify if that fixes the test?

[KornevNikita]

@KornevNikita The revert PR was merged, can you verify if that fixes the test?

Yes, that fixes the issue.