STY: use pytest.raises context syntax #24676
Conversation
gfyoung
added
Refactor
| pytest.raises(ValueError, Unpacker, read_size=5, max_buffer_size=3) | ||
| msg = "read_size should be less or equal to max_buffer_size" | ||
| with pytest.raises(ValueError, match=msg): | ||
| Unpacker(read_size=5, max_buffer_size=3) |
There was a problem hiding this comment.
I see at least two distinct tests in this one. In the interest of smaller, more modular, testing, let's at least break it up here where I've commented.
|
@simonjayhawkins : Even though it's just a handful of files, I have to commend the acrobatics you had to do to make this all work. 👍 |
| @@ -326,7 +346,8 @@ def test_next(self, mmap_file): | |||
| next_line = next(wrapper) | |||
| assert next_line.strip() == line.strip() | |||
|
|
|||
| pytest.raises(StopIteration, next, wrapper) | |||
| with pytest.raises(StopIteration, match=r'$^'): | |||
| next(wrapper) | |||
There was a problem hiding this comment.
I'm surprised r'$^' works, but it seems it does. In any case, it's a bit confusing to read, because normally the caret (^) checks for the beginning of a line and the dollar for the end. I'd switch those characters around.
There was a problem hiding this comment.
@h-vetinari well spotted. thanks. i'll get that sorted.
Thanks! this could be because either the error messages should be more consistent, or maybe it would have been better to put the expected error messages in the parametrization. The 'acrobatics' would then be scoped to just the variations due to platform/interpreter/compiler. |
|
thanks @simonjayhawkins |
xref #24332