String#dump returns incorrect character escaping for unicode codepoints > 0xffff

[postmodern]

I noticed a slight difference in how String#dump escapes unicode characters greater than 0xffff on TruffleRuby 22.3.0 vs. CRuby 3.0.4. TruffleRuby should escape unicode characters greater than 0xffff as \u{xxxx} to match CRuby; no clue why CRuby does this.

TruffleRuby 22.3.0

0xffff.chr(Encoding::UTF_8).dump
# =>  "\"\\uFFFF\""
0x10000.chr(Encoding::UTF_8).dump
# => "\"\\u10000\""

CRuby 3.0.4

0xffff.chr(Encoding::UTF_8).dump
# => "\"\\uFFFF\""
0x10000.chr(Encoding::UTF_8).dump
# => "\"\\u{10000}\""

[eregon]

Thanks for the report.
I guess it's done so to know how many digits to look at, i.e., always 4 without {}, and what's between \u{ and } otherwise.

[andrykonchin]

Fixed in 57e53f8