On Wed, Dec 3, 2025, at 23:23, Rowan Tommins [IMSoP] wrote: > On 3 December 2025 22:01:10 GMT, Rob Landers <[email protected]> wrote: > >On Wed, Dec 3, 2025, at 22:56, Morgan wrote: > >> Well, provided you stick to ASCII-only identifiers... > >> > >> ``` > >> > >> class Thïng {} > >> > >> class THÏNG {} > >> > >> class thïng {} > >> ``` > > > >Classes are not case-sensitive: https://3v4l.org/PLb4J#v8.5.0 > > > Morgan is right that the case folding is only applied to ASCII letters. In > the example, "Thïng" and "thïng" are the same, because they differ only by > "T" vs "t", but "THÏNG" is different, because "ï" and "Ï" are not case > folded. <https://3v4l.org/PLb4J#v8.5.0> > > In fact, they're not treated as letters at all. If I remember correctly, > they're not even analysed as code points, each individual byte above 0x7F is > treated as completely opaque. > > > Rowan Tommins > [IMSoP]
Ah, thanks! I missed the little dots on the i. It’s worth pointing out that the behaviour here isn’t changed, just my understanding of what the current behaviour actually is. :) It's using the same stuff everywhere else is using, so there isn't any new behaviour here, just the same behaviour in new places. — Rob
