From 03f2a590c21797db03239545f21eb99ec0b229cf Mon Sep 17 00:00:00 2001 From: =?utf8?q?David=20Davidovi=C4=87?= Date: Wed, 24 Dec 2014 04:27:07 +0100 Subject: [PATCH] Add Levenshtein distance implementation Copy lev_distance.rs verbatim from Rust since str::lev_distance has been deprecated by rust-lang/rust@9b99436 with no replacement and change code to use it instead of the deprecated function --- src/bin/cargo.rs | 4 +-- src/cargo/util/lev_distance.rs | 63 ++++++++++++++++++++++++++++++++++ src/cargo/util/mod.rs | 2 ++ 3 files changed, 67 insertions(+), 2 deletions(-) create mode 100644 src/cargo/util/lev_distance.rs diff --git a/src/bin/cargo.rs b/src/bin/cargo.rs index 9a96d6469..7d05e9ccb 100644 --- a/src/bin/cargo.rs +++ b/src/bin/cargo.rs @@ -13,7 +13,7 @@ use std::io::process::{Command,InheritFd,ExitStatus,ExitSignal}; use cargo::{execute_main_without_stdin, handle_error, shell}; use cargo::core::MultiShell; -use cargo::util::{CliError, CliResult}; +use cargo::util::{CliError, CliResult, lev_distance}; #[deriving(RustcDecodable)] struct Flags { @@ -139,7 +139,7 @@ fn find_closest(cmd: &str) -> Option { // doing it this way (instead of just .min_by(|c| c.lev_distance(cmd))) // allows us to only make suggestions that have an edit distance of // 3 or less - .map(|c| (c.lev_distance(cmd), c)) + .map(|c| (lev_distance(c.as_slice(), cmd), c)) .filter(|&(d, _): &(uint, &String)| d < 4u) .min_by(|&(d, _)| d) { Some((_, c)) => { diff --git a/src/cargo/util/lev_distance.rs b/src/cargo/util/lev_distance.rs new file mode 100644 index 000000000..24e988374 --- /dev/null +++ b/src/cargo/util/lev_distance.rs @@ -0,0 +1,63 @@ +// Copyright 2012-2014 The Rust Project Developers. See the COPYRIGHT +// file at the top-level directory of this distribution and at +// http://rust-lang.org/COPYRIGHT. +// +// Licensed under the Apache License, Version 2.0 or the MIT license +// , at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +use std::cmp; + +pub fn lev_distance(me: &str, t: &str) -> uint { + if me.is_empty() { return t.chars().count(); } + if t.is_empty() { return me.chars().count(); } + + let mut dcol = Vec::from_fn(t.len() + 1, |x| x); + let mut t_last = 0; + + for (i, sc) in me.chars().enumerate() { + + let mut current = i; + dcol[0] = current + 1; + + for (j, tc) in t.chars().enumerate() { + + let next = dcol[j + 1]; + + if sc == tc { + dcol[j + 1] = current; + } else { + dcol[j + 1] = cmp::min(current, next); + dcol[j + 1] = cmp::min(dcol[j + 1], dcol[j]) + 1; + } + + current = next; + t_last = j; + } + } + + dcol[t_last + 1] +} + +#[test] +fn test_lev_distance() { + use std::char::{ from_u32, MAX }; + // Test bytelength agnosticity + for c in range(0u32, MAX as u32) + .filter_map(|i| from_u32(i)) + .map(|i| String::from_char(1, i)) { + assert_eq!(lev_distance(c[], c[]), 0); + } + + let a = "\nMäry häd ä little lämb\n\nLittle lämb\n"; + let b = "\nMary häd ä little lämb\n\nLittle lämb\n"; + let c = "Mary häd ä little lämb\n\nLittle lämb\n"; + assert_eq!(lev_distance(a, b), 1); + assert_eq!(lev_distance(b, a), 1); + assert_eq!(lev_distance(a, c), 2); + assert_eq!(lev_distance(c, a), 2); + assert_eq!(lev_distance(b, c), 1); + assert_eq!(lev_distance(c, b), 1); +} diff --git a/src/cargo/util/mod.rs b/src/cargo/util/mod.rs index 18daf6d97..1ce80b094 100644 --- a/src/cargo/util/mod.rs +++ b/src/cargo/util/mod.rs @@ -5,6 +5,7 @@ pub use self::errors::{CargoResult, CargoError, BoxError, ChainError, CliResult} pub use self::errors::{CliError, FromError, ProcessError}; pub use self::errors::{process_error, internal_error, internal, human, caused_human}; pub use self::paths::{realpath, join_paths}; +pub use self::lev_distance::{lev_distance}; pub use self::hex::{to_hex, short_hash}; pub use self::dependency_queue::{DependencyQueue, Fresh, Dirty, Freshness}; pub use self::dependency_queue::Dependency; @@ -26,6 +27,7 @@ pub mod result; pub mod to_semver; pub mod to_url; pub mod toml; +pub mod lev_distance; mod dependency_queue; mod sha256; mod vcs; -- 2.30.2