You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Mostly just an annoyance, but we should clean it up eventually. When you run python runtests.py it will spam a bunch of printf 50 times after the test suite summary output, with this line showing up over and over:
Testing printf: 7
The stdout spam comes from this test: tests/test_AST_translator.py::TestASTTranslator::test_call. I tried a number of ways to suppress or intercept the stream with pytest, but it was hopeless, I think because it is coming from compiled Kokkos/C++ stuff in other threads. Normal Python prints are easy to intercept and assert on with pytest using i.e., capsys fixtures.
In the absence of that option, perhaps we could support stream redirection. This is a bit of a hack below, but an improved version may work? Even just the hack version reduces the number of visible prints from 50 to 1, though I don't yet know how to assert on them b/c they don't show up in that file locally. Probably thread-related or something. Perhaps that's enough of a hint to give a C++/Kokkos guru a better idea.
Mostly just an annoyance, but we should clean it up eventually. When you run
python runtests.py
it will spam a bunch ofprintf
50 times after the test suite summary output, with this line showing up over and over:Testing printf: 7
The stdout spam comes from this test:
tests/test_AST_translator.py::TestASTTranslator::test_call
. I tried a number of ways to suppress or intercept the stream withpytest
, but it was hopeless, I think because it is coming from compiled Kokkos/C++ stuff in other threads. Normal Python prints are easy to intercept andassert
on withpytest
using i.e.,capsys
fixtures.In the absence of that option, perhaps we could support stream redirection. This is a bit of a hack below, but an improved version may work? Even just the hack version reduces the number of visible prints from
50
to1
, though I don't yet know how to assert on them b/c they don't show up in that file locally. Probably thread-related or something. Perhaps that's enough of a hint to give a C++/Kokkos guru a better idea.The text was updated successfully, but these errors were encountered: