In this track: Dynamic ABI and nested layouts
advanced · 5 MIN READ

Count UTF-8 bytes correctly

Visible characters and encoded bytes are different counts.

LET’S PICTURE IT

One printed symbol can need several storage boxes when written in a byte alphabet.

SEE IT FOR YOURSELF
Follow a dynamic string from head to tailStep 1 / 4
Follow a dynamic string from head to tailArgument headbyte 0 · offset 32hex word ends in 20String tailstarts at byte 32length + contentPositions are measured from argument byte zero, excluding any selectorOffset destinationargument byte 32read a length wordLength = 5hello is five UTF-8 bytescontent starts at byte 64A length word and an offset word have different jobsContent word · argument byte 6468 65 6c 6c 6f5 payload bytes27 padding bytesBlocks are schematic; their drawn widths do not represent the exact byte ratioHeadoffset: 32argument byte 0Tail lengthlength: 5argument byte 32Tail contenthello + paddingargument byte 6432 + 32 + 32 = 96 bytesWith a four-byte selector, total calldata length is 100 bytes

This example encodes one string parameter with value hello. A dynamic argument uses an offset in its head, even when it is the only argument.

Animated local teaching model. No wallet, RPC or live blockchain data. Some labels and hashes are abbreviated for readability.

Choose the exact text

A string is encoded as UTF-8 bytes. The text é uses two bytes, c3 a9, in its precomposed form. A displayed symbol is not automatically one byte.

Measure encoded bytes

The length word for this example is two. JavaScript string length counts UTF-16 code units, so it is not a general substitute for a UTF-8 byte measurement.

Watch equivalent appearances

An accented letter can also be expressed with a base letter and a combining mark. Similar-looking text may have different bytes unless an application deliberately normalizes it.

Verify before adding padding

First encode the text, then count those bytes, then pad the content block. Do not count the hex prefix, printed hex digits or trailing zeros as original content.

Trace a local example

types: string
values: ["é"]
no function selector prefix

word 0 · argument offset 0 bytes
0x0000000000000000000000000000000000000000000000000000000000000020

word 1 · argument offset 32 bytes
0x0000000000000000000000000000000000000000000000000000000000000002

word 2 · argument offset 64 bytes
0xc3a9000000000000000000000000000000000000000000000000000000000000

The length word is 2 even though the displayed text looks like one letter.

TURN UNDERSTANDING INTO PRACTICEOpen the ABI encoder →
QUICK CHECK · NO PRESSURE

How many UTF-8 bytes encode precomposed é?

Your progress stays in this browser.
Keep exploring the original referenceSolidity documentation ↗

Original explanations by EtherTools. Examples are educational and source-linked; this site is independent of the Ethereum Foundation. Our approach.