diff --git a/example/jsonfindptrs/jsonfindptrs.cc b/example/jsonfindptrs/jsonfindptrs.cc index 2e3ae872..2bf74f98 100644 --- a/example/jsonfindptrs/jsonfindptrs.cc +++ b/example/jsonfindptrs/jsonfindptrs.cc @@ -624,6 +624,10 @@ JsonThing::parse_string(TokenStream& ts, TokenStream::Result tsr) { // program doesn't enable that by default, but if you're copy/pasting // this JsonThing code and your program does enable that option, // here's how to handle it. + // + // As per the quirk documentation, there are two options for how to + // interpret a backslash-x: as a byte or as a Unicode code point. + // This implementation chooses as a byte. wuffs_base__slice_u8 encoded = tsr.src_data; if (encoded.len & 3) { return Result( diff --git a/std/json/decode_quirks.wuffs b/std/json/decode_quirks.wuffs index 7feb5afb..69ca90bc 100644 --- a/std/json/decode_quirks.wuffs +++ b/std/json/decode_quirks.wuffs @@ -75,14 +75,22 @@ pub const quirk_allow_backslash_single_quote base.u32 = 0x4909_9400 | 0x06 pub const quirk_allow_backslash_v base.u32 = 0x4909_9400 | 0x07 // When this quirk is enabled, e.g. "abc\xeFz" is accepted as a JSON string, -// decoding to 5 bytes: 0x61, 0x62, 0x63, 0xEF and 0x7A. There are exactly 2 -// encoded bytes after each "\x". +// There are exactly 2 encoded bytes after each "\x". // +// "\x", "\x9", "\x9$" and "\X99" are all still rejected. +// +// It is up to the caller how to interpret the two hexadecimal digits. There +// are two sensible choices: as a byte or as a Unicode code point. +// +// As a byte means decoding to 5 bytes: 0x61, 0x62, 0x63, 0xEF and 0x7A. // Decoded strings are byte strings, no longer guaranteed to be valid UTF-8 and // even if the overall byte string is valid UTF-8, tokens are also no longer // guaranteed to split on UTF-8 boundaries. // -// "\x", "\x9", "\x9$" and "\X99" are all still rejected. +// As a Unicode code point means decoding to 6 bytes: 0x61, 0x62, 0x63, 0xC3, +// 0xAF and 0x7A. The UTF-8 encoding of U+00EF LATIN SMALL LETTER I WITH +// DIAERESIS is the two byte sequence (0xC3, 0xAF). Decoded strings are still +// valid UTF-8 and tokens still split on UTF-8 boundaries. pub const quirk_allow_backslash_x base.u32 = 0x4909_9400 | 0x08 // When this quirk is enabled, e.g. "abc\0z" is accepted as a JSON string,