mirror of
https://github.com/denoland/deno.git
synced 2024-12-22 07:14:47 -05:00
perf(cache): single cache file for typescript emit (#24994)
This commit is contained in:
parent
d8dfe6dc97
commit
a8ce02473a
1 changed files with 92 additions and 73 deletions
165
cli/cache/emit.rs
vendored
165
cli/cache/emit.rs
vendored
|
@ -5,33 +5,25 @@ use std::path::PathBuf;
|
|||
use deno_ast::ModuleSpecifier;
|
||||
use deno_core::anyhow::anyhow;
|
||||
use deno_core::error::AnyError;
|
||||
use deno_core::serde_json;
|
||||
use deno_core::unsync::sync::AtomicFlag;
|
||||
use serde::Deserialize;
|
||||
use serde::Serialize;
|
||||
|
||||
use super::DiskCache;
|
||||
use super::FastInsecureHasher;
|
||||
|
||||
#[derive(Debug, Deserialize, Serialize)]
|
||||
struct EmitMetadata {
|
||||
pub source_hash: u64,
|
||||
pub emit_hash: u64,
|
||||
}
|
||||
|
||||
/// The cache that stores previously emitted files.
|
||||
pub struct EmitCache {
|
||||
disk_cache: DiskCache,
|
||||
cli_version: &'static str,
|
||||
emit_failed_flag: AtomicFlag,
|
||||
file_serializer: EmitFileSerializer,
|
||||
}
|
||||
|
||||
impl EmitCache {
|
||||
pub fn new(disk_cache: DiskCache) -> Self {
|
||||
Self {
|
||||
disk_cache,
|
||||
cli_version: crate::version::DENO_VERSION_INFO.deno,
|
||||
emit_failed_flag: Default::default(),
|
||||
file_serializer: EmitFileSerializer {
|
||||
cli_version: crate::version::DENO_VERSION_INFO.deno,
|
||||
},
|
||||
}
|
||||
}
|
||||
|
||||
|
@ -48,37 +40,11 @@ impl EmitCache {
|
|||
specifier: &ModuleSpecifier,
|
||||
expected_source_hash: u64,
|
||||
) -> Option<Vec<u8>> {
|
||||
let meta_filename = self.get_meta_filename(specifier)?;
|
||||
let emit_filename = self.get_emit_filename(specifier)?;
|
||||
|
||||
// load and verify the meta data file is for this source and CLI version
|
||||
let bytes = self.disk_cache.get(&meta_filename).ok()?;
|
||||
let meta: EmitMetadata = serde_json::from_slice(&bytes).ok()?;
|
||||
if meta.source_hash != expected_source_hash {
|
||||
return None;
|
||||
}
|
||||
|
||||
// load and verify the emit is for the meta data
|
||||
let emit_bytes = self.disk_cache.get(&emit_filename).ok()?;
|
||||
if meta.emit_hash != compute_emit_hash(&emit_bytes, self.cli_version) {
|
||||
return None;
|
||||
}
|
||||
|
||||
// everything looks good, return it
|
||||
Some(emit_bytes)
|
||||
}
|
||||
|
||||
/// Gets the filepath which stores the emit.
|
||||
pub fn get_emit_filepath(
|
||||
&self,
|
||||
specifier: &ModuleSpecifier,
|
||||
) -> Option<PathBuf> {
|
||||
Some(
|
||||
self
|
||||
.disk_cache
|
||||
.location
|
||||
.join(self.get_emit_filename(specifier)?),
|
||||
)
|
||||
let bytes = self.disk_cache.get(&emit_filename).ok()?;
|
||||
self
|
||||
.file_serializer
|
||||
.deserialize(bytes, expected_source_hash)
|
||||
}
|
||||
|
||||
/// Sets the emit code in the cache.
|
||||
|
@ -107,32 +73,26 @@ impl EmitCache {
|
|||
return Ok(());
|
||||
}
|
||||
|
||||
let meta_filename = self
|
||||
.get_meta_filename(specifier)
|
||||
.ok_or_else(|| anyhow!("Could not get meta filename."))?;
|
||||
let emit_filename = self
|
||||
.get_emit_filename(specifier)
|
||||
.ok_or_else(|| anyhow!("Could not get emit filename."))?;
|
||||
|
||||
// save the metadata
|
||||
let metadata = EmitMetadata {
|
||||
source_hash,
|
||||
emit_hash: compute_emit_hash(code, self.cli_version),
|
||||
};
|
||||
self
|
||||
.disk_cache
|
||||
.set(&meta_filename, &serde_json::to_vec(&metadata)?)?;
|
||||
|
||||
// save the emit source
|
||||
self.disk_cache.set(&emit_filename, code)?;
|
||||
let cache_data = self.file_serializer.serialize(code, source_hash);
|
||||
self.disk_cache.set(&emit_filename, &cache_data)?;
|
||||
|
||||
Ok(())
|
||||
}
|
||||
|
||||
fn get_meta_filename(&self, specifier: &ModuleSpecifier) -> Option<PathBuf> {
|
||||
self
|
||||
.disk_cache
|
||||
.get_cache_filename_with_extension(specifier, "meta")
|
||||
/// Gets the filepath which stores the emit.
|
||||
pub fn get_emit_filepath(
|
||||
&self,
|
||||
specifier: &ModuleSpecifier,
|
||||
) -> Option<PathBuf> {
|
||||
Some(
|
||||
self
|
||||
.disk_cache
|
||||
.location
|
||||
.join(self.get_emit_filename(specifier)?),
|
||||
)
|
||||
}
|
||||
|
||||
fn get_emit_filename(&self, specifier: &ModuleSpecifier) -> Option<PathBuf> {
|
||||
|
@ -142,15 +102,68 @@ impl EmitCache {
|
|||
}
|
||||
}
|
||||
|
||||
fn compute_emit_hash(bytes: &[u8], cli_version: &str) -> u64 {
|
||||
// it's ok to use an insecure hash here because
|
||||
// if someone can change the emit source then they
|
||||
// can also change the version hash
|
||||
FastInsecureHasher::new_without_deno_version() // use cli_version param instead
|
||||
.write(bytes)
|
||||
// emit should not be re-used between cli versions
|
||||
.write_str(cli_version)
|
||||
.finish()
|
||||
const LAST_LINE_PREFIX: &str = "\n// denoCacheMetadata=";
|
||||
|
||||
struct EmitFileSerializer {
|
||||
cli_version: &'static str,
|
||||
}
|
||||
|
||||
impl EmitFileSerializer {
|
||||
pub fn deserialize(
|
||||
&self,
|
||||
mut bytes: Vec<u8>,
|
||||
expected_source_hash: u64,
|
||||
) -> Option<Vec<u8>> {
|
||||
let last_newline_index = bytes.iter().rposition(|&b| b == b'\n')?;
|
||||
let (content, last_line) = bytes.split_at(last_newline_index);
|
||||
let hashes = last_line.strip_prefix(LAST_LINE_PREFIX.as_bytes())?;
|
||||
let hashes = String::from_utf8_lossy(hashes);
|
||||
let (source_hash, emit_hash) = hashes.split_once(',')?;
|
||||
|
||||
// verify the meta data file is for this source and CLI version
|
||||
let source_hash = source_hash.parse::<u64>().ok()?;
|
||||
if source_hash != expected_source_hash {
|
||||
return None;
|
||||
}
|
||||
let emit_hash = emit_hash.parse::<u64>().ok()?;
|
||||
// prevent using an emit from a different cli version or emits that were tampered with
|
||||
if emit_hash != self.compute_emit_hash(content) {
|
||||
return None;
|
||||
}
|
||||
|
||||
// everything looks good, truncate and return it
|
||||
bytes.truncate(content.len());
|
||||
Some(bytes)
|
||||
}
|
||||
|
||||
pub fn serialize(&self, code: &[u8], source_hash: u64) -> Vec<u8> {
|
||||
let source_hash = source_hash.to_string();
|
||||
let emit_hash = self.compute_emit_hash(code).to_string();
|
||||
let capacity = code.len()
|
||||
+ LAST_LINE_PREFIX.len()
|
||||
+ source_hash.len()
|
||||
+ 1
|
||||
+ emit_hash.len();
|
||||
let mut cache_data = Vec::with_capacity(capacity);
|
||||
cache_data.extend(code);
|
||||
cache_data.extend(LAST_LINE_PREFIX.as_bytes());
|
||||
cache_data.extend(source_hash.as_bytes());
|
||||
cache_data.push(b',');
|
||||
cache_data.extend(emit_hash.as_bytes());
|
||||
debug_assert_eq!(cache_data.len(), capacity);
|
||||
cache_data
|
||||
}
|
||||
|
||||
fn compute_emit_hash(&self, bytes: &[u8]) -> u64 {
|
||||
// it's ok to use an insecure hash here because
|
||||
// if someone can change the emit source then they
|
||||
// can also change the version hash
|
||||
crate::cache::FastInsecureHasher::new_without_deno_version() // use cli_version property instead
|
||||
.write(bytes)
|
||||
// emit should not be re-used between cli versions
|
||||
.write_str(self.cli_version)
|
||||
.finish()
|
||||
}
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
|
@ -165,7 +178,9 @@ mod test {
|
|||
let disk_cache = DiskCache::new(temp_dir.path().as_path());
|
||||
let cache = EmitCache {
|
||||
disk_cache: disk_cache.clone(),
|
||||
cli_version: "1.0.0",
|
||||
file_serializer: EmitFileSerializer {
|
||||
cli_version: "1.0.0",
|
||||
},
|
||||
emit_failed_flag: Default::default(),
|
||||
};
|
||||
let to_string =
|
||||
|
@ -197,7 +212,9 @@ mod test {
|
|||
// try changing the cli version (should not load previous ones)
|
||||
let cache = EmitCache {
|
||||
disk_cache: disk_cache.clone(),
|
||||
cli_version: "2.0.0",
|
||||
file_serializer: EmitFileSerializer {
|
||||
cli_version: "2.0.0",
|
||||
},
|
||||
emit_failed_flag: Default::default(),
|
||||
};
|
||||
assert_eq!(cache.get_emit_code(&specifier1, 10), None);
|
||||
|
@ -206,7 +223,9 @@ mod test {
|
|||
// recreating the cache should still load the data because the CLI version is the same
|
||||
let cache = EmitCache {
|
||||
disk_cache,
|
||||
cli_version: "2.0.0",
|
||||
file_serializer: EmitFileSerializer {
|
||||
cli_version: "2.0.0",
|
||||
},
|
||||
emit_failed_flag: Default::default(),
|
||||
};
|
||||
assert_eq!(
|
||||
|
|
Loading…
Reference in a new issue