[issue28701] Replace PyUnicode_CompareWithASCIIString with _PyUnicode_EqualToASCIIString

[Serhiy Storchaka]

Serhiy Storchaka added the comment:

> Can you please also document the behaviour if you pass two non-ASCII strings which are equal?

What mean "equal"? The left argument is a Unicode string, but the right argument is a byte string. For comparing them we should decode right argument or encode left argument. The result depends on using encoding. _PyUnicode_EqualToASCIIString() uses ASCII (as shown from its name). Non-ASCII strings can't be equal. This is documented.

If the documentation is not clear, could you provide better wording?

> Maybe the API should be more strict and require right to be ASCII: "right string must be encoded to ASCII". I expect an assertion error or a fatal error if right is non-ASCII when Python is compiled in debug mode.

I hesitated about adding an assertion error or a fatal error in a bug fix. But this can be added in develop version.

I don't know what is better -- return 0 in all builds or return 0 in release build and crash in debug build?

----------

_______________________________________
Python tracker <report at bugs.python.org>
<http://bugs.python.org/issue28701>
_______________________________________