Struct TokenizeScratch 

pub struct TokenizeScratch {
    pub xml_buf: Vec<u8>,
    pub text_buf: String,
    /* private fields */
}

Scratch buffer pool for tokenization to minimize allocations.

Pre-allocated buffers that can be reused across tokenization operations to avoid repeated allocations in hot paths. This is critical for embedded environments where allocation overhead must be minimized.

Fields

xml_buf: Vec<u8>

Buffer for XML parsing

text_buf: String

Buffer for text accumulation and normalization

Implementations

impl TokenizeScratch

pub fn new(xml_capacity: usize, text_capacity: usize) -> Self

Create scratch buffers with specified capacities.

Arguments

• xml_capacity - Initial capacity for XML parsing buffer
• text_capacity - Initial capacity for text accumulation buffer

Example

use epub_stream::tokenizer::TokenizeScratch;

let scratch = TokenizeScratch::new(4096, 8192);

pub fn embedded() -> Self

Create buffers suitable for embedded use (small, bounded).

Uses conservative buffer sizes suitable for constrained environments:

• XML buffer: 4KB
• Text buffer: 8KB
• Element stack: 64 elements

pub fn desktop() -> Self

Create buffers for desktop use (larger, more performant).

Uses larger buffer sizes for better performance on desktop:

• XML buffer: 32KB
• Text buffer: 64KB
• Element stack: 64 elements

pub fn clear(&mut self)

Clear all buffers without deallocating.

This preserves the allocated capacity while resetting the length to zero, allowing the buffers to be reused for subsequent tokenization operations without requiring new allocations.

pub fn ensure_text_capacity(&mut self, min_cap: usize)

Ensure text buffer has at least the given capacity.

Arguments

• min_cap - Minimum capacity required

If the current capacity is less than min_cap, the buffer will be expanded to at least that capacity.

Trait Implementations

impl Debug for TokenizeScratch

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.