Fix coderange calculation in String#b

Leave the new coderange unknown if the original encoding is not
ASCII-compatible. Non-ASCII-compatible encoding strings with valid or
broken coderange can end up as ascii-only.

Fixes 9a8f6e392fbd ("Cheaply derive code range for String#b return
value", 2022-07-25).
This commit is contained in:
Kazuki Yamaguchi 2022-09-26 16:39:53 +09:00
parent 1c14e406d3
commit 5b0396473b

View File

@ -10771,6 +10771,7 @@ rb_str_b(VALUE str)
}
str_replace_shared_without_enc(str2, str);
if (rb_enc_asciicompat(STR_ENC_GET(str))) {
// BINARY strings can never be broken; they're either 7-bit ASCII or VALID.
// If we know the receiver's code range then we know the result's code range.
int cr = ENC_CODERANGE(str);
@ -10786,6 +10787,7 @@ rb_str_b(VALUE str)
ENC_CODERANGE_CLEAR(str2);
break;
}
}
return str2;
}