Unfortunately, I don't have much interesting to talk about this week. Lately I've been working on fixing the parts of the built-in modules that were broken by moving to UTF-8. Most of the issues stem from the existing code expecting each unicode character to map cleanly onto one wchar_t. Also, since the RPython library for interacting with native code, known as "rffi", assumes that you want to use RPython unicode objects when manipulating arrays of wchar_t, some extra massaging has to be done to cast to the appropriate int type. Working on this, I finally had to set a Windows VM to test 16-bit wchar_t, which I managed to avoid until now somehow.
When I finish the modules, I'll be able to start working on making everything translate.
No comments:
Post a Comment