v1.7.8

[1.7.8] - 08/18/2026

Features

• (runtime-core) Add peekByte to SdkBufferedSource and SdkBuffer for reading the next byte without consuming it
• (serde-cbor) Optimize CBOR serialization by writing values directly to the output buffer instead of building intermediate byte arrays
• (serde-cbor) Optimize CBOR deserialization by reading values directly from the input buffer instead of building intermediate byte arrays
• (serde-cbor) Optimize CBOR deserialization by peeking the next head byte via indexed buffer access instead of allocating a peek-source chain per value

Fixes

• (serde-cbor) Encode and decode CBOR negative bignums using the RFC 8949 unsigned magnitude representation instead of a signed two's-complement encoding, fixing interoperability for negative BigInteger values
• (serde-cbor) Fix CBOR decimal fraction encoding and decoding to use the RFC 8949 scale exponent, correcting the transmitted value for BigDecimals whose mantissa has more than one significant digit
• (serde-cbor) Encode CBOR text string lengths as a UTF-8 byte count instead of a UTF-16 code-unit count, so multibyte strings produce valid RFC 8949 output
• (serde-cbor) Decode CBOR string, byte-string, and collection lengths as a Long so large lengths are no longer silently truncated to Int
• (serde-cbor) Convert CBOR floating-point values that arrive encoded as integers numerically instead of misinterpreting the integer as raw IEEE-754 bits
• (serde-cbor) Fix decoding of CBOR timestamps encoded as negative integers, which previously decoded with the wrong sign for pre-epoch instants
• #3190 Upgrade to Smithy v1.73.0 to pick up bugfix for endpoint ruleset trimming
• TelemetryProvider and related types are no longer experimental

Miscellaneous

• (serde-cbor) Optimize CBOR text string decoding by reading UTF-8 directly from the buffer instead of copying into an intermediate byte array