diff --git a/src/scripts/bgi/archive/dsc.rs b/src/scripts/bgi/archive/dsc.rs index 663ed4b..2da651a 100644 --- a/src/scripts/bgi/archive/dsc.rs +++ b/src/scripts/bgi/archive/dsc.rs @@ -7,8 +7,7 @@ use crate::utils::bit_stream::*; use crate::utils::num_range::*; use anyhow::Result; use rand::RngExt; - -use std::io::{Seek, Write}; +use std::io::{BufWriter, Seek, Write}; #[derive(Debug)] struct HuffmanCode { @@ -409,6 +408,30 @@ fn generate_canonical_codes(depths: &[u8]) -> Vec> { huffman_codes } +#[inline(always)] +fn get_match_length(a: &[u8], b: &[u8]) -> usize { + let max = a.len().min(b.len()); + let mut len = 0; + + // 每次比较 8 个字节，极大提升长文本或连续 0 的匹配效率 + while len + 8 <= max { + let a_chunk = u64::from_le_bytes(a[len..len + 8].try_into().unwrap()); + let b_chunk = u64::from_le_bytes(b[len..len + 8].try_into().unwrap()); + if a_chunk != b_chunk { + let diff = a_chunk ^ b_chunk; + // 找出第一个不同的字节位置 + return len + (diff.trailing_zeros() / 8) as usize; + } + len += 8; + } + + // 比较剩余的尾部字节 + while len < max && a[len] == b[len] { + len += 1; + } + len +} + #[inline(always)] fn find_match( data: &[u8], @@ -417,29 +440,35 @@ fn find_match( prev: &[i32], config: &CompressConfig, ) -> (usize, usize) { + // 如果是 Store 或者连 2 个字节的匹配空间都不够了，直接返回 if config.mode == MatchMode::Store || pos + 2 > data.len() { return (0, 0); } let max_len = (data.len() - pos).min(257); - // 低等级 RLE: 仅扫描由于格式限制的最小可用距离 (pos - 2) + // 针对 RLE 路径的极速优化 if config.mode == MatchMode::Rle { - let mut best_len = 0; if pos >= 2 { - while best_len < max_len && data.get(pos + best_len) == data.get(pos - 2 + best_len) { - best_len += 1; + // 极速前置检查：由于格式要求匹配至少2个字节， + // 只有当最前面的 2 个字节完全相同时，我们才进行切片和后续耗时的 SWAR 比对。 + // 这排除了 99% 的无效调用。 + if data[pos] == data[pos - 2] && data[pos + 1] == data[pos - 1] { + let src_slice = &data[pos..pos + max_len]; + let match_slice = &data[pos - 2..pos - 2 + max_len]; + let best_len = get_match_length(src_slice, match_slice); + return (best_len, 2); } } - if best_len >= 2 { - return (best_len, 2); - } return (0, 0); } + let src_slice = &data[pos..pos + max_len]; let limit = pos.saturating_sub(4097); - let key = ((data[pos] as u16) << 8) | (data[pos + 1] as u16); - let mut match_pos_i32 = head[key as usize]; + + // Level 3~9: 基于哈希字典进行跳跃搜索 + let key = ((data[pos] as usize) << 8) | (data[pos + 1] as usize); + let mut match_pos_i32 = head[key]; let mut chain_length = config.max_chain; let mut best_len = 0; @@ -458,21 +487,19 @@ fn find_match( continue; } - // 快速剪枝优化 - if best_len < max_len { - if data.get(match_pos + best_len) != data.get(pos + best_len) { + let match_slice = &data[match_pos..match_pos + max_len]; + + // 高速剪枝：如果已知最长长度 best_len 处的那一个字节对不上， + // 说明这条链条再怎么强也不可能突破当前的 best_len，直接跳过整个字符串比对 + if best_len > 0 && best_len < max_len { + if match_slice[best_len] != src_slice[best_len] { match_pos_i32 = prev[match_pos]; chain_length -= 1; continue; } } - let mut current_len = 0; - while current_len < max_len - && data.get(pos + current_len) == data.get(match_pos + current_len) - { - current_len += 1; - } + let current_len = get_match_length(src_slice, match_slice); if current_len > best_len { best_len = current_len; @@ -498,7 +525,7 @@ fn find_match( /// Encoder for Buriko General Interpreter/Ethornell compressed files (DSC format). pub struct DscEncoder<'a, T: Write + Seek> { - stream: MsbBitWriter<'a, T>, + stream: MsbBitWriter<'a, BufWriter>, magic: u32, key: u32, dec_count: u32, @@ -507,7 +534,7 @@ pub struct DscEncoder<'a, T: Write + Seek> { impl<'a, T: Write + Seek> DscEncoder<'a, T> { /// Creates a new DscEncoder with the given writer and compression level (0-9). - pub fn new(writer: &'a mut T, level: u8) -> Self { + pub fn new(writer: &'a mut BufWriter, level: u8) -> Self { let stream = MsbBitWriter::new(writer); DscEncoder { stream, @@ -524,16 +551,10 @@ impl<'a, T: Write + Seek> DscEncoder<'a, T> { let mut pos = 0; let config = &COMPRESS_CONFIGS[self.level as usize]; - let mut head: Vec = vec![-1; 1 << 16]; - let mut prev: Vec = vec![-1; data.len()]; - - let insert_dict = |p: usize, head: &mut [i32], prev: &mut [i32]| { - if config.mode != MatchMode::Rle && p + 1 < data.len() { - let key = ((data[p] as u16) << 8) | (data[p + 1] as u16); - prev[p] = head[key as usize]; - head[key as usize] = p as i32; - } - }; + // 预分配哈希表，65536 对应 2 bytes 的所有可能 + let mut head = vec![-1i32; 1 << 16]; + let mut prev = vec![-1i32; data.len()]; + let insert_limit = data.len().saturating_sub(1); // 防止 data[p + 1] 越界 while pos < data.len() { if config.mode == MatchMode::Store { @@ -547,12 +568,17 @@ impl<'a, T: Write + Seek> DscEncoder<'a, T> { if match_len >= 2 { let mut lazy_match = false; - // 延迟匹配逻辑 + // 延迟匹配逻辑 (Lazy Evaluation) if config.mode == MatchMode::Lazy && match_len <= config.max_lazy && pos + 1 < data.len() { - insert_dict(pos, &mut head, &mut prev); + // 为下一次尝试预先将当前 pos 插入字典 + if pos < insert_limit { + let key = ((data[pos] as usize) << 8) | (data[pos + 1] as usize); + prev[pos] = head[key]; + head[key] = pos as i32; + } let (next_len, _) = find_match(data, pos + 1, &head, &prev, config); @@ -576,18 +602,30 @@ impl<'a, T: Write + Seek> DscEncoder<'a, T> { && match_len <= config.max_lazy && pos + 1 < data.len() { - 1 // 如果进行了延迟检查，pos 已被插入 + 1 // 如果进行了延迟检查，pos 已被插入，从 1 开始 } else { 0 }; - for i in start_insert..match_len { - insert_dict(pos + i, &mut head, &mut prev); + // 批量插入字典，使用 usize 强制类型，移除闭包产生的隐式开销 + if config.mode != MatchMode::Rle { + for i in start_insert..match_len { + let p = pos + i; + if p < insert_limit { + let key = ((data[p] as usize) << 8) | (data[p + 1] as usize); + prev[p] = head[key]; + head[key] = p as i32; + } + } } pos += match_len; } else { ops.push(LzssOp::Literal(data[pos])); - insert_dict(pos, &mut head, &mut prev); + if config.mode != MatchMode::Rle && pos < insert_limit { + let key = ((data[pos] as usize) << 8) | (data[pos + 1] as usize); + prev[pos] = head[key]; + head[key] = pos as i32; + } pos += 1; } } @@ -713,6 +751,7 @@ impl ScriptBuilder for DscBuilder { _file_encoding: Encoding, config: &ExtraConfig, ) -> Result<()> { + let mut writer = std::io::BufWriter::new(&mut writer); let encoder = DscEncoder::new(&mut writer, config.bgi_compress_level); let data = crate::utils::files::read_file(filename)?; encoder.pack(&data)?; @@ -775,6 +814,7 @@ impl Script for Dsc { _encoding: Encoding, _output_encoding: Encoding, ) -> Result<()> { + let mut file = std::io::BufWriter::new(&mut file); let encoder = DscEncoder::new(&mut file, self.level); let data = crate::utils::files::read_file(custom_filename)?; encoder.pack(&data)?; diff --git a/src/scripts/bgi/archive/v1.rs b/src/scripts/bgi/archive/v1.rs index 44f525c..c085daa 100644 --- a/src/scripts/bgi/archive/v1.rs +++ b/src/scripts/bgi/archive/v1.rs @@ -720,7 +720,8 @@ impl<'a, T: Write + Seek> Seek for BgiArchiveFileWithDsc<'a, T> { impl<'a, T: Write + Seek> Drop for BgiArchiveFileWithDsc<'a, T> { fn drop(&mut self) { let buf = self.buf.as_slice(); - let encoder = DscEncoder::new(&mut self.writer, self.compress_level); + let mut writer = std::io::BufWriter::new(&mut self.writer); + let encoder = DscEncoder::new(&mut writer, self.compress_level); if let Err(e) = encoder.pack(&buf) { eprintln!("Failed to write DSC data: {}", e); crate::COUNTER.inc_error(); diff --git a/src/scripts/bgi/archive/v2.rs b/src/scripts/bgi/archive/v2.rs index ca5467c..4649131 100644 --- a/src/scripts/bgi/archive/v2.rs +++ b/src/scripts/bgi/archive/v2.rs @@ -723,7 +723,8 @@ impl<'a, T: Write + Seek> Seek for BgiArchiveFileWithDsc<'a, T> { impl<'a, T: Write + Seek> Drop for BgiArchiveFileWithDsc<'a, T> { fn drop(&mut self) { let buf = self.buf.as_slice(); - let encoder = DscEncoder::new(&mut self.writer, self.compress_level); + let mut writer = std::io::BufWriter::new(&mut self.writer); + let encoder = DscEncoder::new(&mut writer, self.compress_level); if let Err(e) = encoder.pack(&buf) { eprintln!("Failed to write DSC data: {}", e); crate::COUNTER.inc_error();