Debugging

The first question about a CBOR document is usually what is in it. A decoded tree answers that directly:

// The whole shape
var_dump($object);

// Entry by entry, for a map or a list
foreach ($map as $item) {
    echo sprintf(
        "Key: %s, Value: %s\n",
        $item->getKey()->normalize(),
        $item->getValue()->normalize()
    );
}

// What it encodes to
$encoded = (string) $object;
echo 'Hex: ' . bin2hex($encoded) . "\n";
echo 'Length: ' . strlen($encoded) . " bytes\n";

Printing the class of each object is often faster than printing its value: get_class($object) tells you whether that number came in as a tagged instant or a plain integer, and whether a string arrived in definite or indefinite-length form.

Reading the bytes

A CBOR head is one byte: the three high bits are the major type, the five low bits the additional information. bin2hex() and the table of major types are enough to read the start of a document by hand — a2 is a map of two entries, d9 d9 f7 the self-described CBOR tag.

Elsewhere

• cbor.me shows a document in its diagnostic notation, both ways round.
• cbor.io lists the specification, the registry and the implementations in other languages.

Both are third-party sites: paste nothing into them that you would not publish.