2015-06-23 08:58:58 -05:00
|
|
|
// Copyright 2015 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 <LICENSE-APACHE or
|
|
|
|
// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
|
|
|
|
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
|
|
|
|
// option. This file may not be copied, modified, or distributed
|
|
|
|
// except according to those terms.
|
|
|
|
|
2016-01-10 08:12:15 -06:00
|
|
|
// Formatting and tools for comments.
|
2015-06-23 08:58:58 -05:00
|
|
|
|
2016-01-10 08:12:15 -06:00
|
|
|
use std::{self, iter};
|
|
|
|
|
|
|
|
use syntax::codemap::Span;
|
2015-08-27 07:07:15 -05:00
|
|
|
|
2015-09-06 00:39:28 -05:00
|
|
|
use Indent;
|
|
|
|
use config::Config;
|
2016-01-10 08:12:15 -06:00
|
|
|
use rewrite::RewriteContext;
|
2015-06-23 08:58:58 -05:00
|
|
|
use string::{StringFormat, rewrite_string};
|
2016-01-10 08:12:15 -06:00
|
|
|
use utils::wrap_str;
|
2015-06-23 08:58:58 -05:00
|
|
|
|
2016-09-29 14:34:46 -05:00
|
|
|
fn is_custom_comment(comment: &str) -> bool {
|
|
|
|
if !comment.starts_with("//") {
|
|
|
|
false
|
|
|
|
} else {
|
|
|
|
if let Some(c) = comment.chars().nth(2) {
|
|
|
|
!c.is_alphanumeric() && !c.is_whitespace()
|
|
|
|
} else {
|
|
|
|
false
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-09-06 00:39:28 -05:00
|
|
|
pub fn rewrite_comment(orig: &str,
|
|
|
|
block_style: bool,
|
|
|
|
width: usize,
|
|
|
|
offset: Indent,
|
|
|
|
config: &Config)
|
2015-09-25 05:53:25 -05:00
|
|
|
-> Option<String> {
|
2015-06-23 08:58:58 -05:00
|
|
|
let s = orig.trim();
|
|
|
|
|
|
|
|
// Edge case: block comments. Let's not trim their lines (for now).
|
2016-06-03 16:18:19 -05:00
|
|
|
let (opener, closer, line_start) =
|
|
|
|
if block_style {
|
2015-12-07 22:04:40 -06:00
|
|
|
("/* ", " */", " * ")
|
2016-06-03 16:18:19 -05:00
|
|
|
} else if !config.normalize_comments {
|
2016-07-09 08:41:28 -05:00
|
|
|
if orig.starts_with("/**") && !orig.starts_with("/**/") {
|
2016-06-03 16:18:19 -05:00
|
|
|
("/** ", " **/", " ** ")
|
|
|
|
} else if orig.starts_with("/*!") {
|
|
|
|
("/*! ", " */", " * ")
|
|
|
|
} else if orig.starts_with("/*") {
|
|
|
|
("/* ", " */", " * ")
|
|
|
|
} else if orig.starts_with("///") {
|
|
|
|
("/// ", "", "/// ")
|
|
|
|
} else if orig.starts_with("//!") {
|
|
|
|
("//! ", "", "//! ")
|
|
|
|
} else {
|
|
|
|
("// ", "", "// ")
|
|
|
|
}
|
2016-07-09 08:41:28 -05:00
|
|
|
} else if orig.starts_with("///") ||
|
|
|
|
(orig.starts_with("/**") && !orig.starts_with("/**/")) {
|
2015-12-07 22:04:40 -06:00
|
|
|
("/// ", "", "/// ")
|
2016-06-03 16:18:19 -05:00
|
|
|
} else if orig.starts_with("//!") || orig.starts_with("/*!") {
|
2015-12-07 22:04:40 -06:00
|
|
|
("//! ", "", "//! ")
|
2016-09-29 14:34:46 -05:00
|
|
|
} else if is_custom_comment(orig) {
|
|
|
|
if orig.chars().nth(3) == Some(' ') {
|
|
|
|
(&orig[0..4], "", &orig[0..4])
|
|
|
|
} else {
|
|
|
|
(&orig[0..3], "", &orig[0..3])
|
|
|
|
}
|
2015-12-07 22:04:40 -06:00
|
|
|
} else {
|
|
|
|
("// ", "", "// ")
|
2016-06-03 16:18:19 -05:00
|
|
|
};
|
2015-06-23 08:58:58 -05:00
|
|
|
|
2015-09-10 17:52:16 -05:00
|
|
|
let max_chars = width.checked_sub(closer.len() + opener.len()).unwrap_or(1);
|
2015-06-23 08:58:58 -05:00
|
|
|
|
2015-07-15 20:31:20 -05:00
|
|
|
let fmt = StringFormat {
|
|
|
|
opener: "",
|
|
|
|
closer: "",
|
|
|
|
line_start: line_start,
|
|
|
|
line_end: "",
|
|
|
|
width: max_chars,
|
2015-09-06 00:39:28 -05:00
|
|
|
offset: offset + (opener.len() - line_start.len()),
|
2015-07-15 20:31:20 -05:00
|
|
|
trim_end: true,
|
2015-09-06 00:39:28 -05:00
|
|
|
config: config,
|
2015-07-15 20:31:20 -05:00
|
|
|
};
|
2015-06-23 08:58:58 -05:00
|
|
|
|
2015-09-18 23:50:44 -05:00
|
|
|
let indent_str = offset.to_string(config);
|
2015-06-23 08:58:58 -05:00
|
|
|
let line_breaks = s.chars().filter(|&c| c == '\n').count();
|
|
|
|
|
2016-09-16 20:20:00 -05:00
|
|
|
let num_bare_lines = s.lines()
|
|
|
|
.enumerate()
|
|
|
|
.map(|(_, line)| line.trim())
|
|
|
|
.filter(|l| !(l.starts_with('*') || l.starts_with("//") || l.starts_with("/*")))
|
|
|
|
.count();
|
|
|
|
|
|
|
|
if num_bare_lines > 0 && !config.normalize_comments {
|
|
|
|
return Some(orig.to_owned());
|
|
|
|
}
|
|
|
|
|
2015-09-25 05:53:25 -05:00
|
|
|
let lines = s.lines()
|
2016-04-22 02:03:36 -05:00
|
|
|
.enumerate()
|
|
|
|
.map(|(i, mut line)| {
|
|
|
|
line = line.trim();
|
|
|
|
// Drop old closer.
|
|
|
|
if i == line_breaks && line.ends_with("*/") && !line.starts_with("//") {
|
|
|
|
line = &line[..(line.len() - 2)];
|
|
|
|
}
|
|
|
|
|
|
|
|
line.trim_right()
|
|
|
|
})
|
|
|
|
.map(left_trim_comment_line)
|
|
|
|
.map(|line| {
|
2016-05-02 03:54:25 -05:00
|
|
|
if orig.starts_with("/*") && line_breaks == 0 {
|
2016-04-22 02:03:36 -05:00
|
|
|
line.trim_left()
|
|
|
|
} else {
|
|
|
|
line
|
|
|
|
}
|
|
|
|
});
|
2015-09-25 05:53:25 -05:00
|
|
|
|
|
|
|
let mut result = opener.to_owned();
|
|
|
|
for line in lines {
|
2015-12-07 22:04:40 -06:00
|
|
|
if result == opener {
|
2016-08-23 09:14:45 -05:00
|
|
|
if line.is_empty() {
|
2015-12-07 22:04:40 -06:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
} else {
|
2015-09-25 05:53:25 -05:00
|
|
|
result.push('\n');
|
|
|
|
result.push_str(&indent_str);
|
|
|
|
result.push_str(line_start);
|
|
|
|
}
|
|
|
|
|
2015-11-09 13:03:01 -06:00
|
|
|
if config.wrap_comments && line.len() > max_chars {
|
2016-04-11 13:45:47 -05:00
|
|
|
let rewrite = rewrite_string(line, &fmt).unwrap_or(line.to_owned());
|
2015-09-25 05:53:25 -05:00
|
|
|
result.push_str(&rewrite);
|
|
|
|
} else {
|
2016-08-23 09:14:45 -05:00
|
|
|
if line.is_empty() {
|
2015-10-19 14:41:47 -05:00
|
|
|
// Remove space if this is an empty comment or a doc comment.
|
|
|
|
result.pop();
|
2015-09-25 05:53:25 -05:00
|
|
|
}
|
2015-10-19 14:41:47 -05:00
|
|
|
result.push_str(line);
|
2015-09-25 05:53:25 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
result.push_str(closer);
|
2015-12-07 22:04:40 -06:00
|
|
|
if result == opener {
|
|
|
|
// Trailing space.
|
|
|
|
result.pop();
|
|
|
|
}
|
2015-09-25 05:53:25 -05:00
|
|
|
|
|
|
|
Some(result)
|
2015-06-23 08:58:58 -05:00
|
|
|
}
|
|
|
|
|
2015-09-04 16:39:33 -05:00
|
|
|
fn left_trim_comment_line(line: &str) -> &str {
|
2015-12-07 22:04:40 -06:00
|
|
|
if line.starts_with("//! ") || line.starts_with("/// ") || line.starts_with("/*! ") ||
|
|
|
|
line.starts_with("/** ") {
|
2015-10-20 04:12:52 -05:00
|
|
|
&line[4..]
|
2016-09-29 14:34:46 -05:00
|
|
|
} else if is_custom_comment(line) {
|
|
|
|
if line.len() > 3 && line.chars().nth(3) == Some(' ') {
|
|
|
|
&line[4..]
|
|
|
|
} else {
|
|
|
|
&line[3..]
|
|
|
|
}
|
2015-10-20 04:12:52 -05:00
|
|
|
} else if line.starts_with("/* ") || line.starts_with("// ") || line.starts_with("//!") ||
|
2016-06-03 16:18:19 -05:00
|
|
|
line.starts_with("///") ||
|
|
|
|
line.starts_with("** ") || line.starts_with("/*!") ||
|
2016-07-09 08:41:28 -05:00
|
|
|
(line.starts_with("/**") && !line.starts_with("/**/")) {
|
2015-06-23 08:58:58 -05:00
|
|
|
&line[3..]
|
2015-12-07 22:04:40 -06:00
|
|
|
} else if line.starts_with("/*") || line.starts_with("* ") || line.starts_with("//") ||
|
2016-06-03 16:18:19 -05:00
|
|
|
line.starts_with("**") {
|
2015-06-23 08:58:58 -05:00
|
|
|
&line[2..]
|
2016-08-23 09:14:45 -05:00
|
|
|
} else if line.starts_with('*') {
|
2015-06-23 08:58:58 -05:00
|
|
|
&line[1..]
|
|
|
|
} else {
|
|
|
|
line
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
pub trait FindUncommented {
|
|
|
|
fn find_uncommented(&self, pat: &str) -> Option<usize>;
|
|
|
|
}
|
|
|
|
|
|
|
|
impl FindUncommented for str {
|
|
|
|
fn find_uncommented(&self, pat: &str) -> Option<usize> {
|
|
|
|
let mut needle_iter = pat.chars();
|
2015-08-27 07:07:15 -05:00
|
|
|
for (kind, (i, b)) in CharClasses::new(self.char_indices()) {
|
2015-06-23 08:58:58 -05:00
|
|
|
match needle_iter.next() {
|
2015-08-27 07:07:15 -05:00
|
|
|
None => {
|
|
|
|
return Some(i - pat.len());
|
|
|
|
}
|
2015-11-20 14:05:10 -06:00
|
|
|
Some(c) => {
|
|
|
|
match kind {
|
2016-01-10 08:12:15 -06:00
|
|
|
FullCodeCharKind::Normal if b == c => {}
|
2015-11-20 14:05:10 -06:00
|
|
|
_ => {
|
|
|
|
needle_iter = pat.chars();
|
|
|
|
}
|
2015-06-23 08:58:58 -05:00
|
|
|
}
|
2015-11-20 14:05:10 -06:00
|
|
|
}
|
2015-06-23 08:58:58 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// Handle case where the pattern is a suffix of the search string
|
|
|
|
match needle_iter.next() {
|
|
|
|
Some(_) => None,
|
2015-08-15 22:58:17 -05:00
|
|
|
None => Some(self.len() - pat.len()),
|
2015-06-23 08:58:58 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// Returns the first byte position after the first comment. The given string
|
|
|
|
// is expected to be prefixed by a comment, including delimiters.
|
|
|
|
// Good: "/* /* inner */ outer */ code();"
|
|
|
|
// Bad: "code(); // hello\n world!"
|
|
|
|
pub fn find_comment_end(s: &str) -> Option<usize> {
|
2015-08-27 07:07:15 -05:00
|
|
|
let mut iter = CharClasses::new(s.char_indices());
|
|
|
|
for (kind, (i, _c)) in &mut iter {
|
2016-01-10 08:12:15 -06:00
|
|
|
if kind == FullCodeCharKind::Normal {
|
2015-08-27 07:07:15 -05:00
|
|
|
return Some(i);
|
2015-06-23 08:58:58 -05:00
|
|
|
}
|
2015-08-27 07:07:15 -05:00
|
|
|
}
|
2015-06-23 08:58:58 -05:00
|
|
|
|
2015-08-27 07:07:15 -05:00
|
|
|
// Handle case where the comment ends at the end of s.
|
|
|
|
if iter.status == CharClassesStatus::Normal {
|
|
|
|
Some(s.len())
|
|
|
|
} else {
|
2015-06-23 08:58:58 -05:00
|
|
|
None
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-08-27 07:07:15 -05:00
|
|
|
/// Returns true if text contains any comment.
|
|
|
|
pub fn contains_comment(text: &str) -> bool {
|
2016-01-10 08:12:15 -06:00
|
|
|
CharClasses::new(text.chars()).any(|(kind, _)| kind.is_comment())
|
2015-08-27 07:07:15 -05:00
|
|
|
}
|
|
|
|
|
2015-10-19 14:41:47 -05:00
|
|
|
struct CharClasses<T>
|
2015-08-27 07:07:15 -05:00
|
|
|
where T: Iterator,
|
|
|
|
T::Item: RichChar
|
|
|
|
{
|
|
|
|
base: iter::Peekable<T>,
|
|
|
|
status: CharClassesStatus,
|
|
|
|
}
|
|
|
|
|
2015-10-19 14:41:47 -05:00
|
|
|
trait RichChar {
|
2015-08-27 07:07:15 -05:00
|
|
|
fn get_char(&self) -> char;
|
|
|
|
}
|
|
|
|
|
|
|
|
impl RichChar for char {
|
|
|
|
fn get_char(&self) -> char {
|
|
|
|
*self
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
impl RichChar for (usize, char) {
|
|
|
|
fn get_char(&self) -> char {
|
|
|
|
self.1
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
#[derive(PartialEq, Eq, Debug, Clone, Copy)]
|
|
|
|
enum CharClassesStatus {
|
|
|
|
Normal,
|
|
|
|
LitString,
|
|
|
|
LitStringEscape,
|
|
|
|
LitChar,
|
|
|
|
LitCharEscape,
|
|
|
|
// The u32 is the nesting deepness of the comment
|
|
|
|
BlockComment(u32),
|
2015-10-19 14:40:00 -05:00
|
|
|
// Status when the '/' has been consumed, but not yet the '*', deepness is
|
|
|
|
// the new deepness (after the comment opening).
|
2015-08-27 07:07:15 -05:00
|
|
|
BlockCommentOpening(u32),
|
2015-10-19 14:40:00 -05:00
|
|
|
// Status when the '*' has been consumed, but not yet the '/', deepness is
|
|
|
|
// the new deepness (after the comment closing).
|
2015-08-27 07:07:15 -05:00
|
|
|
BlockCommentClosing(u32),
|
|
|
|
LineComment,
|
|
|
|
}
|
|
|
|
|
2016-01-10 15:04:30 -06:00
|
|
|
/// Distinguish between functionnal part of code and comments
|
2015-08-27 07:07:15 -05:00
|
|
|
#[derive(PartialEq, Eq, Debug, Clone, Copy)]
|
2015-10-19 14:40:00 -05:00
|
|
|
pub enum CodeCharKind {
|
2015-08-27 07:07:15 -05:00
|
|
|
Normal,
|
|
|
|
Comment,
|
|
|
|
}
|
|
|
|
|
2016-01-10 15:04:30 -06:00
|
|
|
/// Distinguish between functionnal part of code and comments,
|
|
|
|
/// describing opening and closing of comments for ease when chunking
|
|
|
|
/// code from tagged characters
|
2016-01-10 08:12:15 -06:00
|
|
|
#[derive(PartialEq, Eq, Debug, Clone, Copy)]
|
|
|
|
enum FullCodeCharKind {
|
|
|
|
Normal,
|
2016-01-10 15:04:30 -06:00
|
|
|
/// The first character of a comment, there is only one for a comment (always '/')
|
2016-01-10 08:12:15 -06:00
|
|
|
StartComment,
|
2016-01-10 15:04:30 -06:00
|
|
|
/// Any character inside a comment including the second character of comment
|
|
|
|
/// marks ("//", "/*")
|
2016-01-10 08:12:15 -06:00
|
|
|
InComment,
|
2016-01-10 15:04:30 -06:00
|
|
|
/// Last character of a comment, '\n' for a line comment, '/' for a block comment.
|
2016-01-10 08:12:15 -06:00
|
|
|
EndComment,
|
|
|
|
}
|
|
|
|
|
|
|
|
impl FullCodeCharKind {
|
|
|
|
fn is_comment(&self) -> bool {
|
|
|
|
match *self {
|
|
|
|
FullCodeCharKind::Normal => false,
|
|
|
|
FullCodeCharKind::StartComment |
|
|
|
|
FullCodeCharKind::InComment |
|
|
|
|
FullCodeCharKind::EndComment => true,
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fn to_codecharkind(&self) -> CodeCharKind {
|
|
|
|
if self.is_comment() {
|
|
|
|
CodeCharKind::Comment
|
|
|
|
} else {
|
|
|
|
CodeCharKind::Normal
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-11-22 17:02:54 -06:00
|
|
|
impl<T> CharClasses<T>
|
|
|
|
where T: Iterator,
|
|
|
|
T::Item: RichChar
|
|
|
|
{
|
2015-10-19 14:41:47 -05:00
|
|
|
fn new(base: T) -> CharClasses<T> {
|
2015-09-26 01:29:48 -05:00
|
|
|
CharClasses {
|
|
|
|
base: base.peekable(),
|
|
|
|
status: CharClassesStatus::Normal,
|
|
|
|
}
|
2015-08-27 07:07:15 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-11-22 17:02:54 -06:00
|
|
|
impl<T> Iterator for CharClasses<T>
|
|
|
|
where T: Iterator,
|
|
|
|
T::Item: RichChar
|
|
|
|
{
|
2016-01-10 08:12:15 -06:00
|
|
|
type Item = (FullCodeCharKind, T::Item);
|
2015-08-27 07:07:15 -05:00
|
|
|
|
2016-01-10 08:12:15 -06:00
|
|
|
fn next(&mut self) -> Option<(FullCodeCharKind, T::Item)> {
|
2015-08-27 07:07:15 -05:00
|
|
|
let item = try_opt!(self.base.next());
|
|
|
|
let chr = item.get_char();
|
|
|
|
self.status = match self.status {
|
2015-11-20 14:05:10 -06:00
|
|
|
CharClassesStatus::LitString => {
|
|
|
|
match chr {
|
|
|
|
'"' => CharClassesStatus::Normal,
|
|
|
|
'\\' => CharClassesStatus::LitStringEscape,
|
|
|
|
_ => CharClassesStatus::LitString,
|
|
|
|
}
|
|
|
|
}
|
2015-08-27 07:07:15 -05:00
|
|
|
CharClassesStatus::LitStringEscape => CharClassesStatus::LitString,
|
2015-11-20 14:05:10 -06:00
|
|
|
CharClassesStatus::LitChar => {
|
|
|
|
match chr {
|
|
|
|
'\\' => CharClassesStatus::LitCharEscape,
|
|
|
|
'\'' => CharClassesStatus::Normal,
|
|
|
|
_ => CharClassesStatus::LitChar,
|
|
|
|
}
|
|
|
|
}
|
2015-08-27 07:07:15 -05:00
|
|
|
CharClassesStatus::LitCharEscape => CharClassesStatus::LitChar,
|
|
|
|
CharClassesStatus::Normal => {
|
|
|
|
match chr {
|
|
|
|
'"' => CharClassesStatus::LitString,
|
|
|
|
'\'' => CharClassesStatus::LitChar,
|
2015-11-20 14:05:10 -06:00
|
|
|
'/' => {
|
|
|
|
match self.base.peek() {
|
|
|
|
Some(next) if next.get_char() == '*' => {
|
|
|
|
self.status = CharClassesStatus::BlockCommentOpening(1);
|
2016-01-10 08:12:15 -06:00
|
|
|
return Some((FullCodeCharKind::StartComment, item));
|
2015-11-20 14:05:10 -06:00
|
|
|
}
|
|
|
|
Some(next) if next.get_char() == '/' => {
|
|
|
|
self.status = CharClassesStatus::LineComment;
|
2016-01-10 08:12:15 -06:00
|
|
|
return Some((FullCodeCharKind::StartComment, item));
|
2015-11-20 14:05:10 -06:00
|
|
|
}
|
|
|
|
_ => CharClassesStatus::Normal,
|
2015-08-27 07:07:15 -05:00
|
|
|
}
|
2015-11-20 14:05:10 -06:00
|
|
|
}
|
2015-08-27 07:07:15 -05:00
|
|
|
_ => CharClassesStatus::Normal,
|
|
|
|
}
|
|
|
|
}
|
|
|
|
CharClassesStatus::BlockComment(deepness) => {
|
2016-01-10 08:12:15 -06:00
|
|
|
assert!(deepness != 0);
|
2015-08-27 07:07:15 -05:00
|
|
|
self.status = match self.base.peek() {
|
2015-11-20 14:05:10 -06:00
|
|
|
Some(next) if next.get_char() == '/' && chr == '*' => {
|
|
|
|
CharClassesStatus::BlockCommentClosing(deepness - 1)
|
|
|
|
}
|
|
|
|
Some(next) if next.get_char() == '*' && chr == '/' => {
|
|
|
|
CharClassesStatus::BlockCommentOpening(deepness + 1)
|
|
|
|
}
|
2015-08-27 07:07:15 -05:00
|
|
|
_ => CharClassesStatus::BlockComment(deepness),
|
|
|
|
};
|
2016-01-10 08:12:15 -06:00
|
|
|
return Some((FullCodeCharKind::InComment, item));
|
2015-08-27 07:07:15 -05:00
|
|
|
}
|
|
|
|
CharClassesStatus::BlockCommentOpening(deepness) => {
|
|
|
|
assert_eq!(chr, '*');
|
|
|
|
self.status = CharClassesStatus::BlockComment(deepness);
|
2016-01-10 08:12:15 -06:00
|
|
|
return Some((FullCodeCharKind::InComment, item));
|
2015-08-27 07:07:15 -05:00
|
|
|
}
|
|
|
|
CharClassesStatus::BlockCommentClosing(deepness) => {
|
|
|
|
assert_eq!(chr, '/');
|
2016-01-10 08:12:15 -06:00
|
|
|
if deepness == 0 {
|
|
|
|
self.status = CharClassesStatus::Normal;
|
|
|
|
return Some((FullCodeCharKind::EndComment, item));
|
2015-08-27 07:07:15 -05:00
|
|
|
} else {
|
2016-01-10 08:12:15 -06:00
|
|
|
self.status = CharClassesStatus::BlockComment(deepness);
|
|
|
|
return Some((FullCodeCharKind::InComment, item));
|
|
|
|
}
|
2015-08-27 07:07:15 -05:00
|
|
|
}
|
|
|
|
CharClassesStatus::LineComment => {
|
2016-01-10 08:12:15 -06:00
|
|
|
match chr {
|
|
|
|
'\n' => {
|
|
|
|
self.status = CharClassesStatus::Normal;
|
|
|
|
return Some((FullCodeCharKind::EndComment, item));
|
|
|
|
}
|
|
|
|
_ => {
|
|
|
|
self.status = CharClassesStatus::LineComment;
|
|
|
|
return Some((FullCodeCharKind::InComment, item));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
};
|
|
|
|
Some((FullCodeCharKind::Normal, item))
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/// Iterator over functional and commented parts of a string. Any part of a string is either
|
|
|
|
/// functional code, either *one* block comment, either *one* line comment. Whitespace between
|
|
|
|
/// comments is functional code. Line comments contain their ending newlines.
|
|
|
|
struct UngroupedCommentCodeSlices<'a> {
|
|
|
|
slice: &'a str,
|
|
|
|
iter: iter::Peekable<CharClasses<std::str::CharIndices<'a>>>,
|
|
|
|
}
|
|
|
|
|
|
|
|
impl<'a> UngroupedCommentCodeSlices<'a> {
|
|
|
|
fn new(code: &'a str) -> UngroupedCommentCodeSlices<'a> {
|
|
|
|
UngroupedCommentCodeSlices {
|
|
|
|
slice: code,
|
|
|
|
iter: CharClasses::new(code.char_indices()).peekable(),
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
impl<'a> Iterator for UngroupedCommentCodeSlices<'a> {
|
|
|
|
type Item = (CodeCharKind, usize, &'a str);
|
|
|
|
|
|
|
|
fn next(&mut self) -> Option<Self::Item> {
|
|
|
|
let (kind, (start_idx, _)) = try_opt!(self.iter.next());
|
|
|
|
match kind {
|
|
|
|
FullCodeCharKind::Normal => {
|
|
|
|
// Consume all the Normal code
|
|
|
|
while let Some(&(FullCodeCharKind::Normal, (_, _))) = self.iter.peek() {
|
|
|
|
let _ = self.iter.next();
|
|
|
|
}
|
2015-08-27 07:07:15 -05:00
|
|
|
}
|
2016-01-10 08:12:15 -06:00
|
|
|
FullCodeCharKind::StartComment => {
|
|
|
|
// Consume the whole comment
|
|
|
|
while let Some((FullCodeCharKind::InComment, (_, _))) = self.iter.next() {}
|
|
|
|
}
|
|
|
|
_ => panic!(),
|
|
|
|
}
|
|
|
|
let slice = match self.iter.peek() {
|
|
|
|
Some(&(_, (end_idx, _))) => &self.slice[start_idx..end_idx],
|
|
|
|
None => &self.slice[start_idx..],
|
2015-08-27 07:07:15 -05:00
|
|
|
};
|
2016-01-10 08:12:15 -06:00
|
|
|
Some((if kind.is_comment() {
|
2016-09-10 00:02:05 -05:00
|
|
|
CodeCharKind::Comment
|
|
|
|
} else {
|
|
|
|
CodeCharKind::Normal
|
|
|
|
},
|
2016-01-10 08:12:15 -06:00
|
|
|
start_idx,
|
|
|
|
slice))
|
2015-08-27 07:07:15 -05:00
|
|
|
}
|
|
|
|
}
|
2015-08-31 22:39:37 -05:00
|
|
|
|
2016-01-10 08:12:15 -06:00
|
|
|
|
|
|
|
|
|
|
|
|
2015-10-19 14:41:47 -05:00
|
|
|
/// Iterator over an alternating sequence of functional and commented parts of
|
|
|
|
/// a string. The first item is always a, possibly zero length, subslice of
|
|
|
|
/// functional text. Line style comments contain their ending newlines.
|
2015-10-19 14:41:18 -05:00
|
|
|
pub struct CommentCodeSlices<'a> {
|
2015-10-19 14:40:00 -05:00
|
|
|
slice: &'a str,
|
2015-10-19 14:41:47 -05:00
|
|
|
last_slice_kind: CodeCharKind,
|
2015-10-19 14:40:00 -05:00
|
|
|
last_slice_end: usize,
|
|
|
|
}
|
|
|
|
|
|
|
|
impl<'a> CommentCodeSlices<'a> {
|
2015-10-19 14:41:18 -05:00
|
|
|
pub fn new(slice: &'a str) -> CommentCodeSlices<'a> {
|
2015-10-19 14:40:00 -05:00
|
|
|
CommentCodeSlices {
|
|
|
|
slice: slice,
|
2015-10-19 14:41:47 -05:00
|
|
|
last_slice_kind: CodeCharKind::Comment,
|
2015-10-19 14:40:00 -05:00
|
|
|
last_slice_end: 0,
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
impl<'a> Iterator for CommentCodeSlices<'a> {
|
2015-10-19 14:41:18 -05:00
|
|
|
type Item = (CodeCharKind, usize, &'a str);
|
2015-10-19 14:40:00 -05:00
|
|
|
|
|
|
|
fn next(&mut self) -> Option<Self::Item> {
|
|
|
|
if self.last_slice_end == self.slice.len() {
|
|
|
|
return None;
|
|
|
|
}
|
|
|
|
|
|
|
|
let mut sub_slice_end = self.last_slice_end;
|
2015-10-19 14:41:47 -05:00
|
|
|
let mut first_whitespace = None;
|
|
|
|
let subslice = &self.slice[self.last_slice_end..];
|
|
|
|
let mut iter = CharClasses::new(subslice.char_indices());
|
|
|
|
|
|
|
|
for (kind, (i, c)) in &mut iter {
|
|
|
|
let is_comment_connector = self.last_slice_kind == CodeCharKind::Normal &&
|
|
|
|
&subslice[..2] == "//" &&
|
|
|
|
[' ', '\t'].contains(&c);
|
|
|
|
|
|
|
|
if is_comment_connector && first_whitespace.is_none() {
|
|
|
|
first_whitespace = Some(i);
|
|
|
|
}
|
|
|
|
|
2016-01-10 08:12:15 -06:00
|
|
|
if kind.to_codecharkind() == self.last_slice_kind && !is_comment_connector {
|
2015-10-19 14:41:47 -05:00
|
|
|
let last_index = match first_whitespace {
|
|
|
|
Some(j) => j,
|
|
|
|
None => i,
|
|
|
|
};
|
|
|
|
sub_slice_end = self.last_slice_end + last_index;
|
2015-10-19 14:40:00 -05:00
|
|
|
break;
|
|
|
|
}
|
2015-10-19 14:41:47 -05:00
|
|
|
|
|
|
|
if !is_comment_connector {
|
|
|
|
first_whitespace = None;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if let (None, true) = (iter.next(), sub_slice_end == self.last_slice_end) {
|
|
|
|
// This was the last subslice.
|
|
|
|
sub_slice_end = match first_whitespace {
|
|
|
|
Some(i) => self.last_slice_end + i,
|
|
|
|
None => self.slice.len(),
|
|
|
|
};
|
2015-10-19 14:40:00 -05:00
|
|
|
}
|
|
|
|
|
2015-10-19 14:41:47 -05:00
|
|
|
let kind = match self.last_slice_kind {
|
2015-10-19 14:40:00 -05:00
|
|
|
CodeCharKind::Comment => CodeCharKind::Normal,
|
|
|
|
CodeCharKind::Normal => CodeCharKind::Comment,
|
|
|
|
};
|
2016-03-14 02:51:06 -05:00
|
|
|
let res = (kind, self.last_slice_end, &self.slice[self.last_slice_end..sub_slice_end]);
|
2015-10-19 14:41:47 -05:00
|
|
|
self.last_slice_end = sub_slice_end;
|
|
|
|
self.last_slice_kind = kind;
|
2015-10-19 14:40:00 -05:00
|
|
|
|
2015-10-19 14:41:47 -05:00
|
|
|
Some(res)
|
2015-10-19 14:40:00 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-01-10 08:12:15 -06:00
|
|
|
/// Checks is `new` didn't miss any comment from `span`, if it removed any, return previous text
|
|
|
|
/// (if it fits in the width/offset, else return None), else return `new`
|
|
|
|
pub fn recover_comment_removed(new: String,
|
|
|
|
span: Span,
|
|
|
|
context: &RewriteContext,
|
|
|
|
width: usize,
|
|
|
|
offset: Indent)
|
|
|
|
-> Option<String> {
|
|
|
|
let snippet = context.snippet(span);
|
|
|
|
if changed_comment_content(&snippet, &new) {
|
|
|
|
// We missed some comments
|
|
|
|
// Keep previous formatting if it satisfies the constrains
|
2016-08-23 09:14:45 -05:00
|
|
|
wrap_str(snippet, context.config.max_width, width, offset)
|
2016-01-10 08:12:15 -06:00
|
|
|
} else {
|
|
|
|
Some(new)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/// Return true if the two strings of code have the same payload of comments.
|
|
|
|
/// The payload of comments is everything in the string except:
|
2016-05-02 03:54:25 -05:00
|
|
|
/// - actual code (not comments)
|
|
|
|
/// - comment start/end marks
|
|
|
|
/// - whitespace
|
|
|
|
/// - '*' at the beginning of lines in block comments
|
2016-01-10 08:12:15 -06:00
|
|
|
fn changed_comment_content(orig: &str, new: &str) -> bool {
|
|
|
|
// Cannot write this as a fn since we cannot return types containing closures
|
|
|
|
let code_comment_content = |code| {
|
|
|
|
let slices = UngroupedCommentCodeSlices::new(code);
|
|
|
|
slices.filter(|&(ref kind, _, _)| *kind == CodeCharKind::Comment)
|
2016-04-22 02:03:36 -05:00
|
|
|
.flat_map(|(_, _, s)| CommentReducer::new(s))
|
2016-01-10 08:12:15 -06:00
|
|
|
};
|
|
|
|
let res = code_comment_content(orig).ne(code_comment_content(new));
|
|
|
|
debug!("comment::changed_comment_content: {}\norig: '{}'\nnew: '{}'\nraw_old: {}\nraw_new: {}",
|
|
|
|
res,
|
|
|
|
orig,
|
|
|
|
new,
|
|
|
|
code_comment_content(orig).collect::<String>(),
|
|
|
|
code_comment_content(new).collect::<String>());
|
|
|
|
res
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/// Iterator over the 'payload' characters of a comment.
|
|
|
|
/// It skips whitespace, comment start/end marks, and '*' at the beginning of lines.
|
|
|
|
/// The comment must be one comment, ie not more than one start mark (no multiple line comments,
|
|
|
|
/// for example).
|
|
|
|
struct CommentReducer<'a> {
|
|
|
|
is_block: bool,
|
|
|
|
at_start_line: bool,
|
|
|
|
iter: std::str::Chars<'a>,
|
|
|
|
}
|
|
|
|
|
|
|
|
impl<'a> CommentReducer<'a> {
|
|
|
|
fn new(comment: &'a str) -> CommentReducer<'a> {
|
|
|
|
let is_block = comment.starts_with("/*");
|
|
|
|
let comment = remove_comment_header(comment);
|
|
|
|
CommentReducer {
|
|
|
|
is_block: is_block,
|
|
|
|
at_start_line: false, // There are no supplementary '*' on the first line
|
|
|
|
iter: comment.chars(),
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
impl<'a> Iterator for CommentReducer<'a> {
|
|
|
|
type Item = char;
|
|
|
|
fn next(&mut self) -> Option<Self::Item> {
|
|
|
|
loop {
|
|
|
|
let mut c = try_opt!(self.iter.next());
|
|
|
|
if self.is_block && self.at_start_line {
|
|
|
|
while c.is_whitespace() {
|
|
|
|
c = try_opt!(self.iter.next());
|
|
|
|
}
|
|
|
|
// Ignore leading '*'
|
|
|
|
if c == '*' {
|
|
|
|
c = try_opt!(self.iter.next());
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
if c == '\n' {
|
|
|
|
self.at_start_line = true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if !c.is_whitespace() {
|
|
|
|
return Some(c);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
fn remove_comment_header(comment: &str) -> &str {
|
|
|
|
if comment.starts_with("///") || comment.starts_with("//!") {
|
|
|
|
&comment[3..]
|
|
|
|
} else if comment.starts_with("//") {
|
|
|
|
&comment[2..]
|
2016-07-09 08:41:28 -05:00
|
|
|
} else if (comment.starts_with("/**") && !comment.starts_with("/**/")) ||
|
|
|
|
comment.starts_with("/*!") {
|
2016-01-10 08:12:15 -06:00
|
|
|
&comment[3..comment.len() - 2]
|
|
|
|
} else {
|
|
|
|
assert!(comment.starts_with("/*"),
|
|
|
|
format!("string '{}' is not a comment", comment));
|
|
|
|
&comment[2..comment.len() - 2]
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-08-31 22:39:37 -05:00
|
|
|
#[cfg(test)]
|
|
|
|
mod test {
|
2016-01-10 08:12:15 -06:00
|
|
|
use super::{CharClasses, CodeCharKind, FullCodeCharKind, contains_comment, rewrite_comment,
|
|
|
|
FindUncommented, CommentCodeSlices};
|
2015-09-06 00:39:28 -05:00
|
|
|
use Indent;
|
2015-09-25 05:53:25 -05:00
|
|
|
|
2015-10-19 14:41:47 -05:00
|
|
|
#[test]
|
|
|
|
fn char_classes() {
|
|
|
|
let mut iter = CharClasses::new("//\n\n".chars());
|
|
|
|
|
2016-01-10 08:12:15 -06:00
|
|
|
assert_eq!((FullCodeCharKind::StartComment, '/'), iter.next().unwrap());
|
|
|
|
assert_eq!((FullCodeCharKind::InComment, '/'), iter.next().unwrap());
|
|
|
|
assert_eq!((FullCodeCharKind::EndComment, '\n'), iter.next().unwrap());
|
|
|
|
assert_eq!((FullCodeCharKind::Normal, '\n'), iter.next().unwrap());
|
2015-10-19 14:41:47 -05:00
|
|
|
assert_eq!(None, iter.next());
|
|
|
|
}
|
|
|
|
|
2015-10-19 14:40:00 -05:00
|
|
|
#[test]
|
|
|
|
fn comment_code_slices() {
|
|
|
|
let input = "code(); /* test */ 1 + 1";
|
|
|
|
let mut iter = CommentCodeSlices::new(input);
|
|
|
|
|
2015-10-19 14:41:18 -05:00
|
|
|
assert_eq!((CodeCharKind::Normal, 0, "code(); "), iter.next().unwrap());
|
|
|
|
assert_eq!((CodeCharKind::Comment, 8, "/* test */"),
|
|
|
|
iter.next().unwrap());
|
|
|
|
assert_eq!((CodeCharKind::Normal, 18, " 1 + 1"), iter.next().unwrap());
|
2015-10-19 14:40:00 -05:00
|
|
|
assert_eq!(None, iter.next());
|
|
|
|
}
|
|
|
|
|
2015-10-19 14:41:47 -05:00
|
|
|
#[test]
|
|
|
|
fn comment_code_slices_two() {
|
|
|
|
let input = "// comment\n test();";
|
|
|
|
let mut iter = CommentCodeSlices::new(input);
|
|
|
|
|
|
|
|
assert_eq!((CodeCharKind::Normal, 0, ""), iter.next().unwrap());
|
|
|
|
assert_eq!((CodeCharKind::Comment, 0, "// comment\n"),
|
|
|
|
iter.next().unwrap());
|
|
|
|
assert_eq!((CodeCharKind::Normal, 11, " test();"),
|
|
|
|
iter.next().unwrap());
|
|
|
|
assert_eq!(None, iter.next());
|
|
|
|
}
|
|
|
|
|
|
|
|
#[test]
|
|
|
|
fn comment_code_slices_three() {
|
|
|
|
let input = "1 // comment\n // comment2\n\n";
|
|
|
|
let mut iter = CommentCodeSlices::new(input);
|
|
|
|
|
|
|
|
assert_eq!((CodeCharKind::Normal, 0, "1 "), iter.next().unwrap());
|
|
|
|
assert_eq!((CodeCharKind::Comment, 2, "// comment\n // comment2\n"),
|
|
|
|
iter.next().unwrap());
|
|
|
|
assert_eq!((CodeCharKind::Normal, 29, "\n"), iter.next().unwrap());
|
|
|
|
assert_eq!(None, iter.next());
|
|
|
|
}
|
|
|
|
|
2015-08-31 22:39:37 -05:00
|
|
|
#[test]
|
2015-10-22 16:35:42 -05:00
|
|
|
#[cfg_attr(rustfmt, rustfmt_skip)]
|
2015-08-31 22:39:37 -05:00
|
|
|
fn format_comments() {
|
2015-11-09 13:03:01 -06:00
|
|
|
let mut config: ::config::Config = Default::default();
|
|
|
|
config.wrap_comments = true;
|
2016-04-11 13:45:47 -05:00
|
|
|
|
|
|
|
let comment = rewrite_comment(" //test", true, 100, Indent::new(0, 100), &config).unwrap();
|
|
|
|
assert_eq!("/* test */", comment);
|
|
|
|
|
|
|
|
let comment = rewrite_comment("// comment on a",
|
|
|
|
false,
|
|
|
|
10,
|
|
|
|
Indent::empty(),
|
|
|
|
&config).unwrap();
|
|
|
|
assert_eq!("// comment\n// on a", comment);
|
|
|
|
|
|
|
|
let comment = rewrite_comment("// A multi line comment\n // between args.",
|
|
|
|
false,
|
|
|
|
60,
|
|
|
|
Indent::new(0, 12),
|
|
|
|
&config).unwrap();
|
|
|
|
assert_eq!("// A multi line comment\n // between args.", comment);
|
2015-08-31 22:39:37 -05:00
|
|
|
|
|
|
|
let input = "// comment";
|
2015-09-17 13:21:06 -05:00
|
|
|
let expected =
|
2016-04-11 13:45:47 -05:00
|
|
|
"/* comment */";
|
|
|
|
let comment = rewrite_comment(input, true, 9, Indent::new(0, 69), &config).unwrap();
|
|
|
|
assert_eq!(expected, comment);
|
|
|
|
|
|
|
|
let comment = rewrite_comment("/* trimmed */",
|
|
|
|
true,
|
|
|
|
100,
|
|
|
|
Indent::new(0, 100),
|
|
|
|
&config).unwrap();
|
|
|
|
assert_eq!("/* trimmed */", comment);
|
2015-08-31 22:39:37 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
// This is probably intended to be a non-test fn, but it is not used. I'm
|
|
|
|
// keeping it around unless it helps us test stuff.
|
|
|
|
fn uncommented(text: &str) -> String {
|
2015-09-09 16:14:09 -05:00
|
|
|
CharClasses::new(text.chars())
|
|
|
|
.filter_map(|(s, c)| {
|
|
|
|
match s {
|
2016-01-10 08:12:15 -06:00
|
|
|
FullCodeCharKind::Normal => Some(c),
|
|
|
|
_ => None,
|
2015-09-09 16:14:09 -05:00
|
|
|
}
|
|
|
|
})
|
|
|
|
.collect()
|
2015-08-31 22:39:37 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
#[test]
|
|
|
|
fn test_uncommented() {
|
|
|
|
assert_eq!(&uncommented("abc/*...*/"), "abc");
|
2015-09-23 18:01:01 -05:00
|
|
|
assert_eq!(&uncommented("// .... /* \n../* /* *** / */ */a/* // */c\n"),
|
|
|
|
"..ac\n");
|
2015-10-06 15:13:14 -05:00
|
|
|
assert_eq!(&uncommented("abc \" /* */\" qsdf"), "abc \" /* */\" qsdf");
|
2015-08-31 22:39:37 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
#[test]
|
|
|
|
fn test_contains_comment() {
|
|
|
|
assert_eq!(contains_comment("abc"), false);
|
|
|
|
assert_eq!(contains_comment("abc // qsdf"), true);
|
|
|
|
assert_eq!(contains_comment("abc /* kqsdf"), true);
|
|
|
|
assert_eq!(contains_comment("abc \" /* */\" qsdf"), false);
|
|
|
|
}
|
|
|
|
|
|
|
|
#[test]
|
|
|
|
fn test_find_uncommented() {
|
|
|
|
fn check(haystack: &str, needle: &str, expected: Option<usize>) {
|
|
|
|
assert_eq!(expected, haystack.find_uncommented(needle));
|
|
|
|
}
|
|
|
|
|
|
|
|
check("/*/ */test", "test", Some(6));
|
|
|
|
check("//test\ntest", "test", Some(7));
|
|
|
|
check("/* comment only */", "whatever", None);
|
2015-09-23 18:01:01 -05:00
|
|
|
check("/* comment */ some text /* more commentary */ result",
|
|
|
|
"result",
|
|
|
|
Some(46));
|
2015-08-31 22:39:37 -05:00
|
|
|
check("sup // sup", "p", Some(2));
|
|
|
|
check("sup", "x", None);
|
2015-09-01 13:42:07 -05:00
|
|
|
check(r#"π? /**/ π is nice!"#, r#"π is nice"#, Some(9));
|
2015-08-31 22:39:37 -05:00
|
|
|
check("/*sup yo? \n sup*/ sup", "p", Some(20));
|
|
|
|
check("hel/*lohello*/lo", "hello", None);
|
|
|
|
check("acb", "ab", None);
|
|
|
|
check(",/*A*/ ", ",", Some(0));
|
|
|
|
check("abc", "abc", Some(0));
|
|
|
|
check("/* abc */", "abc", None);
|
|
|
|
check("/**/abc/* */", "abc", Some(4));
|
|
|
|
check("\"/* abc */\"", "abc", Some(4));
|
|
|
|
check("\"/* abc", "abc", Some(4));
|
|
|
|
}
|
|
|
|
}
|