02Sheet 02.03.112 · Entries
A 28-byte text memo that is not valid UTF-8, mapped into the transfer event
MEMO_TEXT holds 28 raw bytes that are not UTF-8 (the maximum length).
FIG. · Where it sits · inside LedgerCloseMetaCAP-0067 · H-B5 · H-A6
MEMO_TEXT bytes that are not UTF-8
MEMO_TEXT carried into the transfer event as to_muxed_id
string
The shape
MEMO_TEXT holds 28 raw bytes that are not UTF-8 (the maximum length). CAP-0067 copies it into the payment's transfer event as to_muxed_id (a string). Readers that decode either as UTF-8 lose or mangle it.
Variants
| Variant | Protocol | Network | Source | Size | sha256 |
|---|---|---|---|---|---|
p30-mainnet-generated | P30 | mainnet passphrase | generated by stellar-core v30.0.0 | 3,708 B | bca68e8dd94e… |
p30-testnet-generated | P30 | testnet passphrase | generated by stellar-core v30.0.0 | 3,708 B | 14208217f729… |
What a correct reader returns
| A correct reader finds | Value | Backed by |
|---|---|---|
| The memo text is 28 bytes, first ff fe 00. | \xff\xfe\0\xc3(\xed\xa0\x80xdrproof-not-utf8-20 | constructionsnapshotcorewalker |
| The transfer event carries the same bytes as to_muxed_id. | some change where 2 fields match | snapshotcorewalker |
Check it yourself
make verify # schema, hashes, the byte-for-byte round trip and every fact above
make oracle # the facts' values, re-read by stellar-core and the schema walkerEntry id memo.text-not-utf8 · compiled from catalog/memo.text-not-utf8.json · entry.json
MEMO_RETURN mapped to to_muxed_id bytes, the same as MEMO_HASH
CAP-0067 renders MEMO_RETURN in the transfer event as bytes, exactly as MEMO_HASH: from the event alone the two cannot be told apart.
A data entry whose value is 64 bytes that are not UTF-8
MANAGE_DATA with a 64-byte value (the maximum) alternating 0x00 and 0xff: not text.