From ac347d788c356b5ce9c181e0eb4ac45a6b125ad3 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Cl=C3=A9ment=20Renault?= <renault.cle@gmail.com>
Date: Thu, 10 Jan 2019 15:23:52 +0100
Subject: [PATCH] feat: Make multiple document updates shadow themselves

---
 src/database/serde/indexer_serializer.rs | 10 +++-------
 src/database/update/raw_builder.rs       | 14 ++++++++++++--
 2 files changed, 15 insertions(+), 9 deletions(-)

diff --git a/src/database/serde/indexer_serializer.rs b/src/database/serde/indexer_serializer.rs
index ae3eba436..bdbfb281d 100644
--- a/src/database/serde/indexer_serializer.rs
+++ b/src/database/serde/indexer_serializer.rs
@@ -49,8 +49,8 @@ where B: TokenizerBuilder
     }
 
     fn serialize_str(self, v: &str) -> Result<Self::Ok, Self::Error> {
-        for Token { word, word_index, char_index } in self.tokenizer_builder.build(v) {
-
+        for token in self.tokenizer_builder.build(v) {
+            let Token { word, word_index, char_index } = token;
             let document_id = self.document_id;
 
             // FIXME must u32::try_from instead
@@ -61,15 +61,13 @@ where B: TokenizerBuilder
 
             // insert the exact representation
             let word_lower = word.to_lowercase();
+            let length = word.chars().count() as u16;
 
             if self.stop_words.contains(&word_lower) { continue }
 
             // and the unidecoded lowercased version
             let word_unidecoded = unidecode::unidecode(word).to_lowercase();
             if word_lower != word_unidecoded {
-
-                // FIXME must u16/u32::try_from instead
-                let length = word_unidecoded.chars().count() as u16;
                 let word_area = match WordArea::new(char_index as u32, length) {
                     Ok(word_area) => word_area,
                     Err(_) => return Ok(()),
@@ -79,8 +77,6 @@ where B: TokenizerBuilder
                 self.update.insert_doc_index(word_unidecoded.into_bytes(), doc_index);
             }
 
-            // FIXME must u16/u32::try_from instead
-            let length = word.chars().count() as u16;
             let word_area = match WordArea::new(char_index as u32, length) {
                 Ok(word_area) => word_area,
                 Err(_) => return Ok(()),
diff --git a/src/database/update/raw_builder.rs b/src/database/update/raw_builder.rs
index d116e05fe..323f55eb5 100644
--- a/src/database/update/raw_builder.rs
+++ b/src/database/update/raw_builder.rs
@@ -1,4 +1,4 @@
-use std::collections::BTreeMap;
+use std::collections::btree_map::{BTreeMap, Entry};
 use std::path::PathBuf;
 use std::error::Error;
 
@@ -39,6 +39,10 @@ impl DocumentUpdate {
 
     pub fn remove(&mut self) {
         self.cleared = true;
+        self.clear();
+    }
+
+    pub fn clear(&mut self) {
         self.words_indexes.clear();
         self.attributes.clear();
     }
@@ -61,7 +65,13 @@ impl RawUpdateBuilder {
     }
 
     pub fn document_update(&mut self, document_id: DocumentId) -> &mut DocumentUpdate {
-        self.document_updates.entry(document_id).or_insert_with(DocumentUpdate::new)
+        match self.document_updates.entry(document_id) {
+            Entry::Occupied(mut occupied) => {
+                occupied.get_mut().clear();
+                occupied.into_mut()
+            },
+            Entry::Vacant(vacant) => vacant.insert(DocumentUpdate::new()),
+        }
     }
 
     pub fn build(mut self) -> Result<Update, Box<Error>> {