1 | use std::iter::Peekable; |
2 | use std::slice::Split; |
3 | |
4 | pub fn unindent(s: &str) -> String { |
5 | let bytes: &[u8] = s.as_bytes(); |
6 | let unindented: Vec = unindent_bytes(bytes); |
7 | String::from_utf8(vec:unindented).unwrap() |
8 | } |
9 | |
10 | // Compute the maximal number of spaces that can be removed from every line, and |
11 | // remove them. |
12 | pub fn unindent_bytes(s: &[u8]) -> Vec<u8> { |
13 | // Document may start either on the same line as opening quote or |
14 | // on the next line |
15 | let ignore_first_line = s.starts_with(b" \n" ) || s.starts_with(b" \r\n" ); |
16 | |
17 | // Largest number of spaces that can be removed from every |
18 | // non-whitespace-only line after the first |
19 | let spaces = s |
20 | .lines() |
21 | .skip(1) |
22 | .filter_map(count_spaces) |
23 | .min() |
24 | .unwrap_or(0); |
25 | |
26 | let mut result = Vec::with_capacity(s.len()); |
27 | for (i, line) in s.lines().enumerate() { |
28 | if i > 1 || (i == 1 && !ignore_first_line) { |
29 | result.push(b' \n' ); |
30 | } |
31 | if i == 0 { |
32 | // Do not un-indent anything on same line as opening quote |
33 | result.extend_from_slice(line); |
34 | } else if line.len() > spaces { |
35 | // Whitespace-only lines may have fewer than the number of spaces |
36 | // being removed |
37 | result.extend_from_slice(&line[spaces..]); |
38 | } |
39 | } |
40 | result |
41 | } |
42 | |
43 | pub trait Unindent { |
44 | type Output; |
45 | |
46 | fn unindent(&self) -> Self::Output; |
47 | } |
48 | |
49 | impl Unindent for str { |
50 | type Output = String; |
51 | |
52 | fn unindent(&self) -> Self::Output { |
53 | unindent(self) |
54 | } |
55 | } |
56 | |
57 | impl Unindent for String { |
58 | type Output = String; |
59 | |
60 | fn unindent(&self) -> Self::Output { |
61 | unindent(self) |
62 | } |
63 | } |
64 | |
65 | impl Unindent for [u8] { |
66 | type Output = Vec<u8>; |
67 | |
68 | fn unindent(&self) -> Self::Output { |
69 | unindent_bytes(self) |
70 | } |
71 | } |
72 | |
73 | impl<'a, T: ?Sized + Unindent> Unindent for &'a T { |
74 | type Output = T::Output; |
75 | |
76 | fn unindent(&self) -> Self::Output { |
77 | (**self).unindent() |
78 | } |
79 | } |
80 | |
81 | // Number of leading spaces in the line, or None if the line is entirely spaces. |
82 | fn count_spaces(line: &[u8]) -> Option<usize> { |
83 | for (i: usize, ch: &u8) in line.iter().enumerate() { |
84 | if *ch != b' ' && *ch != b' \t' { |
85 | return Some(i); |
86 | } |
87 | } |
88 | None |
89 | } |
90 | |
91 | // Based on core::str::StrExt. |
92 | trait BytesExt { |
93 | fn lines(&self) -> Lines; |
94 | } |
95 | |
96 | impl BytesExt for [u8] { |
97 | fn lines(&self) -> Lines { |
98 | fn is_newline(b: &u8) -> bool { |
99 | *b == b' \n' |
100 | } |
101 | let bytestring: &[u8] = if self.starts_with(needle:b" \r\n" ) { |
102 | &self[1..] |
103 | } else { |
104 | self |
105 | }; |
106 | Lines { |
107 | split: bytestring.split(pred:is_newline as fn(&u8) -> bool).peekable(), |
108 | } |
109 | } |
110 | } |
111 | |
112 | struct Lines<'a> { |
113 | split: Peekable<Split<'a, u8, fn(&u8) -> bool>>, |
114 | } |
115 | |
116 | impl<'a> Iterator for Lines<'a> { |
117 | type Item = &'a [u8]; |
118 | |
119 | fn next(&mut self) -> Option<Self::Item> { |
120 | match self.split.next() { |
121 | None => None, |
122 | Some(fragment: &[u8]) => { |
123 | if fragment.is_empty() && self.split.peek().is_none() { |
124 | None |
125 | } else { |
126 | Some(fragment) |
127 | } |
128 | } |
129 | } |
130 | } |
131 | } |
132 | |