From 37489fd4958e09a8231bdc4842c40b23522ae0ac Mon Sep 17 00:00:00 2001
From: ManyTheFish <many@meilisearch.com>
Date: Wed, 1 Mar 2023 18:52:14 +0100
Subject: [PATCH] Return an internal error in the case of matching word is
 invalid

---
 benchmarks/benches/formatting.rs           |  2 +-
 milli/src/error.rs                         |  2 ++
 milli/src/search/matches/matching_words.rs | 14 +++++++++++---
 milli/src/search/matches/mod.rs            |  6 +++---
 milli/src/search/query_tree.rs             |  2 +-
 5 files changed, 18 insertions(+), 8 deletions(-)

diff --git a/benchmarks/benches/formatting.rs b/benchmarks/benches/formatting.rs
index 25d162411..2e0fa0ce7 100644
--- a/benchmarks/benches/formatting.rs
+++ b/benchmarks/benches/formatting.rs
@@ -29,7 +29,7 @@ fn bench_formatting(c: &mut criterion::Criterion) {
 	            (vec![Rc::new(MatchingWord::new("thedoord".to_string(), 1, true).unwrap())], vec![0, 1, 2]),
 	            (vec![Rc::new(MatchingWord::new("doord".to_string(), 1, true).unwrap())], vec![1, 2]),
         	]
-            ), TokenizerBuilder::default().build()),
+            ).unwrap(), TokenizerBuilder::default().build()),
 		},
     ];
 
diff --git a/milli/src/error.rs b/milli/src/error.rs
index 1685f984a..9b1ec0a87 100644
--- a/milli/src/error.rs
+++ b/milli/src/error.rs
@@ -59,6 +59,8 @@ pub enum InternalError {
     Utf8(#[from] str::Utf8Error),
     #[error("An indexation process was explicitly aborted.")]
     AbortedIndexation,
+    #[error("The matching words list contains at least one invalid member.")]
+    InvalidMatchingWords,
 }
 
 #[derive(Error, Debug)]
diff --git a/milli/src/search/matches/matching_words.rs b/milli/src/search/matches/matching_words.rs
index 22ba973b5..22241c457 100644
--- a/milli/src/search/matches/matching_words.rs
+++ b/milli/src/search/matches/matching_words.rs
@@ -7,6 +7,7 @@ use std::rc::Rc;
 use charabia::Token;
 use levenshtein_automata::{Distance, DFA};
 
+use crate::error::InternalError;
 use crate::search::build_dfa;
 use crate::MAX_WORD_LENGTH;
 
@@ -31,12 +32,19 @@ impl fmt::Debug for MatchingWords {
 }
 
 impl MatchingWords {
-    pub fn new(mut matching_words: Vec<(Vec<Rc<MatchingWord>>, Vec<PrimitiveWordId>)>) -> Self {
+    pub fn new(
+        mut matching_words: Vec<(Vec<Rc<MatchingWord>>, Vec<PrimitiveWordId>)>,
+    ) -> crate::Result<Self> {
+        // if one of the matching_words vec doesn't contain a word.
+        if matching_words.iter().any(|(mw, _)| mw.is_empty()) {
+            return Err(InternalError::InvalidMatchingWords.into());
+        }
+
         // Sort word by len in DESC order prioritizing the longuest matches,
         // in order to highlight the longuest part of the matched word.
         matching_words.sort_unstable_by_key(|(mw, _)| Reverse((mw.len(), mw[0].word.len())));
 
-        Self { inner: matching_words }
+        Ok(Self { inner: matching_words })
     }
 
     /// Returns an iterator over terms that match or partially match the given token.
@@ -360,7 +368,7 @@ mod tests {
             (vec![all[2].clone()], vec![2]),
         ];
 
-        let matching_words = MatchingWords::new(matching_words);
+        let matching_words = MatchingWords::new(matching_words).unwrap();
 
         assert_eq!(
             matching_words
diff --git a/milli/src/search/matches/mod.rs b/milli/src/search/matches/mod.rs
index 6ac5123a8..c634ae297 100644
--- a/milli/src/search/matches/mod.rs
+++ b/milli/src/search/matches/mod.rs
@@ -513,7 +513,7 @@ mod tests {
             (vec![all[2].clone()], vec![2]),
         ];
 
-        MatchingWords::new(matching_words)
+        MatchingWords::new(matching_words).unwrap()
     }
 
     impl MatcherBuilder<'_, Vec<u8>> {
@@ -600,7 +600,7 @@ mod tests {
         ];
         let matching_words = vec![(vec![all[0].clone()], vec![0]), (vec![all[1].clone()], vec![1])];
 
-        let matching_words = MatchingWords::new(matching_words);
+        let matching_words = MatchingWords::new(matching_words).unwrap();
 
         let builder = MatcherBuilder::from_matching_words(matching_words);
 
@@ -847,7 +847,7 @@ mod tests {
             (vec![all[4].clone()], vec![2]),
         ];
 
-        let matching_words = MatchingWords::new(matching_words);
+        let matching_words = MatchingWords::new(matching_words).unwrap();
 
         let mut builder = MatcherBuilder::from_matching_words(matching_words);
         builder.highlight_prefix("_".to_string());
diff --git a/milli/src/search/query_tree.rs b/milli/src/search/query_tree.rs
index 431012bdb..541dd8f7a 100755
--- a/milli/src/search/query_tree.rs
+++ b/milli/src/search/query_tree.rs
@@ -747,7 +747,7 @@ fn create_matching_words(
     let mut matching_word_cache = MatchingWordCache::default();
     let mut matching_words = Vec::new();
     ngrams(ctx, authorize_typos, query, &mut matching_words, &mut matching_word_cache, 0)?;
-    Ok(MatchingWords::new(matching_words))
+    MatchingWords::new(matching_words)
 }
 
 pub type PrimitiveQuery = Vec<PrimitiveQueryPart>;