Skip to content

remove some UB when dealing with non-ASCII data

Following on from #2502 (closed), this series deals with a few more spotted cases of insufficient non-ASCII robustness. The scenarios by which you actually run into trouble here are pretty esoteric (non-ASCII UTF-8 characters in unexpected places combined with a ctype.h implementation that misbehaves when given non-ASCII characters), but nevertheless these are technically Undefined Behavior.

Merge request reports