Rope indexing doesn't handle pathologically evil strings
Opened this issue · 1 comments
hawkw commented
I've disabled the QuickCheck rope indexing tests in my dev branch, because although we can now handle a lot of weird unicode graphemes, the tests like to make strings consisting entirely of null
s or entirely of control characters.
We should eventually work on more robust handling of strings like
"\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}\u{0}"
hawkw commented
We should probably be testing against the Big List of Naughty Strings at some point.