switch to diceware platform independent word list (EffLong)

Cargo.lock

@@ -29,6 +29,12 @@ dependencies = [
  "cfg-if",
 ]
 
+[[package]]
+name = "diceware_wordlists"
+version = "1.2.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5f9b52b69d268c7a2bc582e3aec5cdfa43ac91cef4fe6b6751b02da2b43d6166"
+
 [[package]]
 name = "getrandom"
 version = "0.2.17"
@@ -89,6 +95,7 @@ name = "password-generator"
 version = "0.1.0"
 dependencies = [
  "bitflags",
+ "diceware_wordlists",
  "rand",
  "webster",
 ]

Cargo.toml

@@ -7,3 +7,4 @@ edition = "2024"
 rand = "0.8"
 webster = "0.3.0"
 bitflags = "2"
+diceware_wordlists = "1.2.3"

src/main.rs

@@ -3,10 +3,12 @@ use rand::rngs::OsRng;
 use rand::SeedableRng;
 use rand::rngs::StdRng;
 
-use std::fs;
+// use std::fs;
 use rand::prelude::SliceRandom;
 use bitflags::bitflags;
 
+use diceware_wordlists::Wordlist;
+
 const UPPERCASE_CHARS: &[u8] = b"ABCDEFGHIJKLMNOPQRSTUVWXYZ";
 const LOWERCASE_CHARS: &[u8] = b"abcdefghijklmnopqrstuvwxyz";
 const NUMBER_CHARS: &[u8] = b"0123456789";
@@ -196,14 +198,13 @@ fn generate_password(config: &PasswordConfig, rng: &mut impl Rng) -> Result<Stri
     Ok(res)
 }
 
-// todo: haven't updated bits of entropy since adding capitalization, symbols and numbers
+// todo: haven't updated bits of entropy comments since adding capitalization,
-// it will add a few, but not many since it's fairly predictable
+//       symbols, numbers and changing to diceware wordlist
 //
 // How many bits of entropy?
 // log2(104078) -> 16.6 bits per word derived from word_list size
 // log2(18) -> 4.1 bits for separator assuming seporator is one of our SYMBOLs
 // total 70.5 bits of entropy
-// note: dictionary is provided by linux, so entropy is platform dependent ATM
 //
 // we can take a slice of "words" because we don't care about:
 // capacity, mutation (push/pop), allocation strategy
@@ -240,16 +241,22 @@ fn random_word<'a>(word_list: &[&'a str], rng: &mut impl Rng) -> &'a str {
     word_list[rng.gen_range(0..word_list.len())]
 }
 
-fn main() {
+fn get_word_list() -> &'static [&'static str] {
-    // todo: not cross-platform
+    // platform dependent word list
-    // todo: abstract this a bit
+    // let content = fs::read_to_string("/usr/share/dict/words")
-    let content = fs::read_to_string("/usr/share/dict/words")
+    //     .expect("could not read dictionary");
-        .expect("could not read dictionary");
-
     // filter words here for only ascii
     // word_list length is 104078
-    let word_list: Vec<&str> = content.lines().filter(|w|
+    // let word_list: Vec<&str> = content.lines().filter(|w|
-        w.is_ascii()).collect();
+    //     w.is_ascii()).collect();
+
+    // platform independent word list
+    // this list doesn't seem to contain non-ascii or even punctuation
+    Wordlist::get_list(&Wordlist::EffLong)
+}
+
+fn main() {
+    let word_list = get_word_list();
 
     // we definitely want to use the OS for randomness
     // doesn't require seeding because it is backed by CSPRNG, backed by OS