|
There is some fishy error handling in some of the test methods. Let's handle those in a separate issue/PR. An example is test_k_code(), where tuple leak in multiple branches. |
Sorry, something went wrong.
|
Leaks in failing tests? The only situation I see where that's is a problem is when tests are used as API usage examples. So, it's good to fix those, but not too important for older tests. Undefining PyArg_ParseTupleAndKeywords and defining again it is also a bit fishy – IMO it would better to havs the non-Size_T definition in the public header file, albeit shadowed by a macro, so only #undef would suffice. But that's an unrelated issue. |
Sorry, something went wrong.
|
Leaks in failing tests? Yes, only in error handling paths, and it's not a problem in practice. Undefining PyArg_ParseTupleAndKeywords and defining again it is also a bit fishy – IMO it would better to havs the non-Size_T definition in the public header file, albeit shadowed by a macro, so only #undef would suffice. But that's an unrelated issue. +1 |
Sorry, something went wrong.
Uh oh!
There was an error while loading. Please reload this page.
Sorry, something went wrong.
Uh oh!
There was an error while loading. Please reload this page.