inefficient_bytes_concat.md

description	Inefficient bytes concatenation — avoid repeated per-iteration concatenations that waste CPU and memory
sidebar_position	6

inefficient_bytes_concat

Default Severity	Category
warn	Compute

What it does

Flags repeated concatenation of Bytes or BytesN objects inside loop bodies using operator-based concatenation.

Why is this bad

Concatenating bytes repeatedly inside a loop allocates new host objects and copies backing buffers on every iteration, causing quadratic performance degradation and excessive CPU consumption.

Example

Bad — concatenating bytes inside a loop:

// ❌ Triggers: inefficient_bytes_concat
let mut result = Bytes::new(&env);
for item in &items {
    result = result + item;
}

Good — preallocate or accumulate using a vector or buffer before converting to host bytes:

// ✅ Fixed: collect/accumulate outside the loop
let mut buffer = Vec::new();
for item in &items {
    buffer.extend_from_slice(item.to_alloc_vec());
}
let result = Bytes::from_slice(&env, &buffer);

Suggested Fix

Replace + concatenation of Bytes values with a Vec<u8> buffer that accumulates bytes, then convert to Bytes once after the loop using Bytes::from.

Relationship to other lints

Note that this lint (inefficient_bytes_concat) specifically flags the binary + operator (b1 + b2). A related lint, soroban_inefficient_bytes_concat, flags the .push_back() and .append() method calls on Bytes. They detect genuinely different code shapes, but both enforce the same best practice: accumulate bytes in a Vec<u8> instead of using host objects in a loop.