Update wordlists (0.7.20).

finnbear · Dec 18, 2023 · 684231c · 684231c
1 parent 8d2d6ff
commit 684231c
Show file tree

Hide file tree

Showing 11 changed files with 210 additions and 341 deletions.
diff --git a/Cargo.toml b/Cargo.toml
@@ -1,7 +1,7 @@
 [package]
 name = "rustrict"
 authors = ["Finn Bear"]
-version = "0.7.19"
+version = "0.7.20"
 edition = "2021"
 license = "MIT OR Apache-2.0"
 repository = "https://github.com/finnbear/rustrict/"

diff --git a/README.md b/README.md
@@ -177,7 +177,7 @@ is used as a dataset. Positive accuracy is the percentage of profanity detected
 
 | Crate | Accuracy | Positive Accuracy | Negative Accuracy | Time |
 |-------|----------|-------------------|-------------------|------|
-| [rustrict](https://crates.io/crates/rustrict) | 79.86%   | 93.96%            | 76.34%            | 8s   |
+| [rustrict](https://crates.io/crates/rustrict) | 79.85%   | 93.99%            | 76.32%            | 8s   |
 | [censor](https://crates.io/crates/censor) | 76.16%   | 72.76%            | 77.01%            | 23s  |
 
 ## Development

diff --git a/src/censor.rs b/src/censor.rs
@@ -8,6 +8,7 @@ use crate::{is_whitespace, Replacements, Type};
 use std::iter::Filter;
 use std::mem;
 use std::ops::Deref;
+use std::ops::RangeInclusive;
 use std::str::Chars;
 use unicode_normalization::{Decompositions, Recompositions, UnicodeNormalization};
 
@@ -424,8 +425,10 @@ impl<I: Iterator<Item = char>> Iterator for Censor<I> {
                 raw_c, skippable, replacement
             );
 
+            const BLOCK_ELEMENTS : RangeInclusive<char> = '\u{2580}'..='\u{259F}';
+
             if (!self.inline.separate || self.inline.last == Some(self.options.censor_replacement))
-                && raw_c == self.options.censor_replacement
+                && (raw_c == self.options.censor_replacement || BLOCK_ELEMENTS.contains(&raw_c))
             {
                 // Censor replacement found but not beginning of word.
                 self.inline.self_censoring = self.inline.self_censoring.saturating_add(1);

diff --git a/src/dictionary_blacklist.txt b/src/dictionary_blacklist.txt
@@ -175,6 +175,9 @@ commies
 condoms
 coons
 copulates
+cotton picker
+cotton farm
+cotton farmer
 cowards
 crackers
 crapper(.*)
@@ -324,6 +327,7 @@ gyppo
 h
 hand job
 handjobs
+hang yourself
 hater
 haters
 hates
@@ -339,6 +343,9 @@ hijackers
 hoars
 hoes
 holocausts
+homophile
+homophobia
+homophobic
 homos
 homosexual
 honkeys
@@ -450,6 +457,7 @@ menstruates
 menstruations
 mi ger
 micropenis
+mike hawk
 milfs
 missionary position
 mo ron

diff --git a/src/dictionary_extra.txt b/src/dictionary_extra.txt
@@ -71,6 +71,7 @@ enola gay
 fatty acid
 fatty food
 few secs
+ffa game
 fire cracker
 fire crackers
 francoitalian
@@ -171,6 +172,8 @@ pp. 6
 pp. 7
 pp. 8
 pp. 9
+pussinboots
+puss in boots
 ref'd
 rip
 saturated fat