about summary refs log tree commit diff
path: root/tvix/cli/src/refscan.rs
diff options
context:
space:
mode:
authorVincent Ambo <mail@tazj.in>2023-02-02T12·51+0300
committertazjin <tazjin@tvl.su>2023-02-02T17·50+0000
commit9d6f29a72b3b466dd697c2eaa97f9a41b767fdff (patch)
treeb8964a57844dae703390247886b0530594c8a147 /tvix/cli/src/refscan.rs
parent2c07ff0f8c126cb475c6e100b56bbaa03303dda7 (diff)
refactor(tvix/cli): use Wu-Manber string scanning for drv references r/5825
Switch out the string-scanning algorithm used in the reference scanner.

The construction of aho-corasick automata made up the vast majority of
runtime when evaluating nixpkgs previously. While the actual scanning
with a constructed automaton is relatively fast, we almost never scan
for the same set of strings twice and the cost is not worth it.

An algorithm that better matches our needs is the Wu-Manber multiple
string match algorithm, which works efficiently on *long* and *random*
strings of the *same length*, which describes store paths (up to their
hash component).

This switches the refscanner crate to a Rust implementation[0][1] of
this algorithm.

This has several implications:

1. This crate does not provide a way to scan streams. I'm not sure if
   this is an inherent problem with the algorithm (probably not, but
   it would need buffering). Either way, related functions and
   tests (which were actually unused) have been removed.

2. All strings need to be of the same length. For this reason, we
   truncate the known paths after their hash part (they are still
   unique, of course).

3. Passing an empty set of matches, or a match that is shorter than
   the length of a store path, causes the crate to panic. We safeguard
   against this by completely skipping the refscanning if there are no
   known paths (i.e. when evaluating the first derivation of an eval),
   and by bailing out of scanning a string that is shorter than a
   store path.

On the upside, this reduces overall runtime to less 1/5 of what it was
before when evaluating `pkgs.stdenv.drvPath`.

[0]: Frankly, it's a random, research-grade MIT-licensed
     crate that I found on Github:

     https://github.com/jneem/wu-manber

[1]: We probably want to rewrite or at least fork the above crate, and
     add things like a three-byte wide scanner. Evaluating large
     portions of nixpkgs can easily lead to more than 65k derivations
     being scanned for.

Change-Id: I08926778e1e5d5a87fc9ac26e0437aed8bbd9eb0
Reviewed-on: https://cl.tvl.fyi/c/depot/+/8017
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
Diffstat (limited to 'tvix/cli/src/refscan.rs')
-rw-r--r--tvix/cli/src/refscan.rs61
1 files changed, 10 insertions, 51 deletions
diff --git a/tvix/cli/src/refscan.rs b/tvix/cli/src/refscan.rs
index 74110e108809..4314e0164429 100644
--- a/tvix/cli/src/refscan.rs
+++ b/tvix/cli/src/refscan.rs
@@ -7,15 +7,16 @@
 //! The scanner itself is an Aho-Corasick automaton, using the `aho-corasick`
 //! crate.
 
-use aho_corasick::AhoCorasick;
 use std::collections::BTreeSet;
-use std::io;
+use wu_manber::TwoByteWM;
+
+pub const STORE_PATH_LEN: usize = "/nix/store/00000000000000000000000000000000".len();
 
 /// Represents a "primed" reference scanner with an automaton that knows the set
 /// of store paths to scan for.
 pub struct ReferenceScanner {
     candidates: Vec<String>,
-    searcher: AhoCorasick,
+    searcher: TwoByteWM,
     matches: Vec<usize>,
 }
 
@@ -23,7 +24,7 @@ impl ReferenceScanner {
     /// Construct a new `ReferenceScanner` that knows how to scan for the given
     /// candidate store paths.
     pub fn new(candidates: Vec<String>) -> Self {
-        let searcher = AhoCorasick::new_auto_configured(&candidates);
+        let searcher = TwoByteWM::new(&candidates);
 
         ReferenceScanner {
             searcher,
@@ -34,25 +35,14 @@ impl ReferenceScanner {
 
     /// Scan the given string for all non-overlapping matches and collect them
     /// in the scanner.
-    pub fn scan_str<H: AsRef<[u8]>>(&mut self, haystack: H) {
-        for m in self.searcher.find_iter(&haystack) {
-            self.matches.push(m.pattern());
+    pub fn scan_str(&mut self, haystack: &str) {
+        if haystack.len() < STORE_PATH_LEN {
+            return;
         }
-    }
 
-    /// Scan the given reader for all non-overlapping matches, and collect them
-    /// in the scanner. On read failures, this method aborts and returns an
-    /// error to the caller.
-    ///
-    /// Please note that the internal machinery has its own buffering mechanism,
-    /// and where possible the given reader should be unbuffered. See
-    /// [`AhoCorasick::stream_find_iter`] for details on this.
-    pub fn scan_stream<R: io::Read>(&mut self, stream: R) -> io::Result<()> {
-        for m in self.searcher.stream_find_iter(stream) {
-            self.matches.push(m?.pattern());
+        for m in self.searcher.find(&haystack) {
+            self.matches.push(m.pat_idx);
         }
-
-        Ok(())
     }
 
     /// Finalise the reference scanner and return the resulting matches.
@@ -72,13 +62,6 @@ mod tests {
     const HELLO_DRV: &'static str = r#"Derive([("out","/nix/store/33l4p0pn0mybmqzaxfkpppyh7vx1c74p-hello-2.12.1","","")],[("/nix/store/6z1jfnqqgyqr221zgbpm30v91yfj3r45-bash-5.1-p16.drv",["out"]),("/nix/store/ap9g09fxbicj836zm88d56dn3ff4clxl-stdenv-linux.drv",["out"]),("/nix/store/pf80kikyxr63wrw56k00i1kw6ba76qik-hello-2.12.1.tar.gz.drv",["out"])],["/nix/store/9krlzvny65gdc8s7kpb6lkx8cd02c25b-default-builder.sh"],"x86_64-linux","/nix/store/4xw8n979xpivdc46a9ndcvyhwgif00hz-bash-5.1-p16/bin/bash",["-e","/nix/store/9krlzvny65gdc8s7kpb6lkx8cd02c25b-default-builder.sh"],[("buildInputs",""),("builder","/nix/store/4xw8n979xpivdc46a9ndcvyhwgif00hz-bash-5.1-p16/bin/bash"),("cmakeFlags",""),("configureFlags",""),("depsBuildBuild",""),("depsBuildBuildPropagated",""),("depsBuildTarget",""),("depsBuildTargetPropagated",""),("depsHostHost",""),("depsHostHostPropagated",""),("depsTargetTarget",""),("depsTargetTargetPropagated",""),("doCheck","1"),("doInstallCheck",""),("mesonFlags",""),("name","hello-2.12.1"),("nativeBuildInputs",""),("out","/nix/store/33l4p0pn0mybmqzaxfkpppyh7vx1c74p-hello-2.12.1"),("outputs","out"),("patches",""),("pname","hello"),("propagatedBuildInputs",""),("propagatedNativeBuildInputs",""),("src","/nix/store/pa10z4ngm0g83kx9mssrqzz30s84vq7k-hello-2.12.1.tar.gz"),("stdenv","/nix/store/cp65c8nk29qq5cl1wyy5qyw103cwmax7-stdenv-linux"),("strictDeps",""),("system","x86_64-linux"),("version","2.12.1")])"#;
 
     #[test]
-    fn test_empty() {
-        let mut scanner = ReferenceScanner::new(vec![]);
-        scanner.scan_str("hello world");
-        assert!(scanner.finalise().is_empty());
-    }
-
-    #[test]
     fn test_single_match() {
         let mut scanner = ReferenceScanner::new(vec![
             "/nix/store/4xw8n979xpivdc46a9ndcvyhwgif00hz-bash-5.1-p16".into(),
@@ -112,28 +95,4 @@ mod tests {
             assert!(result.contains(c));
         }
     }
-
-    #[test]
-    fn test_multiple_stream() {
-        let candidates = vec![
-            // these exist in the drv:
-            "/nix/store/33l4p0pn0mybmqzaxfkpppyh7vx1c74p-hello-2.12.1".into(),
-            "/nix/store/pf80kikyxr63wrw56k00i1kw6ba76qik-hello-2.12.1.tar.gz.drv".into(),
-            "/nix/store/cp65c8nk29qq5cl1wyy5qyw103cwmax7-stdenv-linux".into(),
-            // this doesn't:
-            "/nix/store/fn7zvafq26f0c8b17brs7s95s10ibfzs-emacs-28.2.drv".into(),
-        ];
-
-        let mut scanner = ReferenceScanner::new(candidates.clone());
-        scanner
-            .scan_stream(HELLO_DRV.as_bytes())
-            .expect("scanning should succeed");
-
-        let result = scanner.finalise();
-        assert_eq!(result.len(), 3);
-
-        for c in candidates[..3].iter() {
-            assert!(result.contains(c));
-        }
-    }
 }