aboutsummaryrefslogtreecommitdiff
path: root/crates/ra_syntax/src/validation
diff options
context:
space:
mode:
Diffstat (limited to 'crates/ra_syntax/src/validation')
-rw-r--r--crates/ra_syntax/src/validation/byte.rs211
-rw-r--r--crates/ra_syntax/src/validation/byte_string.rs178
-rw-r--r--crates/ra_syntax/src/validation/char.rs192
-rw-r--r--crates/ra_syntax/src/validation/mod.rs4
4 files changed, 492 insertions, 93 deletions
diff --git a/crates/ra_syntax/src/validation/byte.rs b/crates/ra_syntax/src/validation/byte.rs
new file mode 100644
index 000000000..43c0d7edd
--- /dev/null
+++ b/crates/ra_syntax/src/validation/byte.rs
@@ -0,0 +1,211 @@
1//! Validation of byte literals
2
3use crate::{
4 ast::{self, AstNode},
5 string_lexing::{self, CharComponentKind},
6 TextRange,
7 validation::char,
8 yellow::{
9 SyntaxError,
10 SyntaxErrorKind::*,
11 },
12};
13
14pub(super) fn validate_byte_node(node: ast::Byte, errors: &mut Vec<SyntaxError>) {
15 let literal_text = node.text();
16 let literal_range = node.syntax().range();
17 let mut components = string_lexing::parse_byte_literal(literal_text);
18 let mut len = 0;
19 for component in &mut components {
20 len += 1;
21 let text = &literal_text[component.range];
22 let range = component.range + literal_range.start();
23 validate_byte_component(text, component.kind, range, errors);
24 }
25
26 if !components.has_closing_quote {
27 errors.push(SyntaxError::new(UnclosedByte, literal_range));
28 }
29
30 if len == 0 {
31 errors.push(SyntaxError::new(EmptyByte, literal_range));
32 }
33
34 if len > 1 {
35 errors.push(SyntaxError::new(OverlongByte, literal_range));
36 }
37}
38
39pub(super) fn validate_byte_component(
40 text: &str,
41 kind: CharComponentKind,
42 range: TextRange,
43 errors: &mut Vec<SyntaxError>,
44) {
45 use self::CharComponentKind::*;
46 match kind {
47 AsciiEscape => validate_byte_escape(text, range, errors),
48 AsciiCodeEscape => validate_byte_code_escape(text, range, errors),
49 UnicodeEscape => errors.push(SyntaxError::new(UnicodeEscapeForbidden, range)),
50 CodePoint => {
51 let c = text
52 .chars()
53 .next()
54 .expect("Code points should be one character long");
55
56 // These bytes must always be escaped
57 if c == '\t' || c == '\r' || c == '\n' {
58 errors.push(SyntaxError::new(UnescapedByte, range));
59 }
60
61 // Only ASCII bytes are allowed
62 if c > 0x7F as char {
63 errors.push(SyntaxError::new(ByteOutOfRange, range));
64 }
65 }
66 }
67}
68
69fn validate_byte_escape(text: &str, range: TextRange, errors: &mut Vec<SyntaxError>) {
70 if text.len() == 1 {
71 // Escape sequence consists only of leading `\`
72 errors.push(SyntaxError::new(EmptyByteEscape, range));
73 } else {
74 let escape_code = text.chars().skip(1).next().unwrap();
75 if !char::is_ascii_escape(escape_code) {
76 errors.push(SyntaxError::new(InvalidByteEscape, range));
77 }
78 }
79}
80
81fn validate_byte_code_escape(text: &str, range: TextRange, errors: &mut Vec<SyntaxError>) {
82 // A ByteCodeEscape has 4 chars, example: `\xDD`
83 if text.len() < 4 {
84 errors.push(SyntaxError::new(TooShortByteCodeEscape, range));
85 } else {
86 assert!(
87 text.chars().count() == 4,
88 "ByteCodeEscape cannot be longer than 4 chars"
89 );
90
91 if u8::from_str_radix(&text[2..], 16).is_err() {
92 errors.push(SyntaxError::new(MalformedByteCodeEscape, range));
93 }
94 }
95}
96
97#[cfg(test)]
98mod test {
99 use crate::SourceFileNode;
100
101 fn build_file(literal: &str) -> SourceFileNode {
102 let src = format!("const C: u8 = b'{}';", literal);
103 SourceFileNode::parse(&src)
104 }
105
106 fn assert_valid_byte(literal: &str) {
107 let file = build_file(literal);
108 assert!(
109 file.errors().len() == 0,
110 "Errors for literal '{}': {:?}",
111 literal,
112 file.errors()
113 );
114 }
115
116 fn assert_invalid_byte(literal: &str) {
117 let file = build_file(literal);
118 assert!(file.errors().len() > 0);
119 }
120
121 #[test]
122 fn test_ansi_codepoints() {
123 for byte in 0..128 {
124 match byte {
125 b'\n' | b'\r' | b'\t' => assert_invalid_byte(&(byte as char).to_string()),
126 b'\'' | b'\\' => { /* Ignore character close and backslash */ }
127 _ => assert_valid_byte(&(byte as char).to_string()),
128 }
129 }
130
131 for byte in 128..=255u8 {
132 assert_invalid_byte(&(byte as char).to_string());
133 }
134 }
135
136 #[test]
137 fn test_unicode_codepoints() {
138 let invalid = ["Ƒ", "バ", "メ", "﷽"];
139 for c in &invalid {
140 assert_invalid_byte(c);
141 }
142 }
143
144 #[test]
145 fn test_unicode_multiple_codepoints() {
146 let invalid = ["नी", "👨‍👨‍"];
147 for c in &invalid {
148 assert_invalid_byte(c);
149 }
150 }
151
152 #[test]
153 fn test_valid_byte_escape() {
154 let valid = [r"\'", "\"", "\\\\", "\\\"", r"\n", r"\r", r"\t", r"\0"];
155 for c in &valid {
156 assert_valid_byte(c);
157 }
158 }
159
160 #[test]
161 fn test_invalid_byte_escape() {
162 let invalid = [r"\a", r"\?", r"\"];
163 for c in &invalid {
164 assert_invalid_byte(c);
165 }
166 }
167
168 #[test]
169 fn test_valid_byte_code_escape() {
170 let valid = [r"\x00", r"\x7F", r"\x55", r"\xF0"];
171 for c in &valid {
172 assert_valid_byte(c);
173 }
174 }
175
176 #[test]
177 fn test_invalid_byte_code_escape() {
178 let invalid = [r"\x", r"\x7"];
179 for c in &invalid {
180 assert_invalid_byte(c);
181 }
182 }
183
184 #[test]
185 fn test_invalid_unicode_escape() {
186 let well_formed = [
187 r"\u{FF}",
188 r"\u{0}",
189 r"\u{F}",
190 r"\u{10FFFF}",
191 r"\u{1_0__FF___FF_____}",
192 ];
193 for c in &well_formed {
194 assert_invalid_byte(c);
195 }
196
197 let invalid = [
198 r"\u",
199 r"\u{}",
200 r"\u{",
201 r"\u{FF",
202 r"\u{FFFFFF}",
203 r"\u{_F}",
204 r"\u{00FFFFF}",
205 r"\u{110000}",
206 ];
207 for c in &invalid {
208 assert_invalid_byte(c);
209 }
210 }
211}
diff --git a/crates/ra_syntax/src/validation/byte_string.rs b/crates/ra_syntax/src/validation/byte_string.rs
new file mode 100644
index 000000000..7b830e97c
--- /dev/null
+++ b/crates/ra_syntax/src/validation/byte_string.rs
@@ -0,0 +1,178 @@
1use crate::{
2 ast::{self, AstNode},
3 string_lexing::{self, StringComponentKind},
4 yellow::{
5 SyntaxError,
6 SyntaxErrorKind::*,
7 },
8};
9
10use super::byte;
11
12pub(crate) fn validate_byte_string_node(node: ast::ByteString, errors: &mut Vec<SyntaxError>) {
13 let literal_text = node.text();
14 let literal_range = node.syntax().range();
15 let mut components = string_lexing::parse_byte_string_literal(literal_text);
16 for component in &mut components {
17 let range = component.range + literal_range.start();
18
19 match component.kind {
20 StringComponentKind::Char(kind) => {
21 // Chars must escape \t, \n and \r codepoints, but strings don't
22 let text = &literal_text[component.range];
23 match text {
24 "\t" | "\n" | "\r" => { /* always valid */ }
25 _ => byte::validate_byte_component(text, kind, range, errors),
26 }
27 }
28 StringComponentKind::IgnoreNewline => { /* always valid */ }
29 }
30 }
31
32 if !components.has_closing_quote {
33 errors.push(SyntaxError::new(UnclosedString, literal_range));
34 }
35}
36
37#[cfg(test)]
38mod test {
39 use crate::SourceFileNode;
40
41 fn build_file(literal: &str) -> SourceFileNode {
42 let src = format!(r#"const S: &'static [u8] = b"{}";"#, literal);
43 println!("Source: {}", src);
44 SourceFileNode::parse(&src)
45 }
46
47 fn assert_valid_str(literal: &str) {
48 let file = build_file(literal);
49 assert!(
50 file.errors().len() == 0,
51 "Errors for literal '{}': {:?}",
52 literal,
53 file.errors()
54 );
55 }
56
57 fn assert_invalid_str(literal: &str) {
58 let file = build_file(literal);
59 assert!(file.errors().len() > 0);
60 }
61
62 #[test]
63 fn test_ansi_codepoints() {
64 for byte in 0..128 {
65 match byte {
66 b'\"' | b'\\' => { /* Ignore string close and backslash */ }
67 _ => assert_valid_str(&(byte as char).to_string()),
68 }
69 }
70
71 for byte in 128..=255u8 {
72 assert_invalid_str(&(byte as char).to_string());
73 }
74 }
75
76 #[test]
77 fn test_unicode_codepoints() {
78 let invalid = ["Ƒ", "バ", "メ", "﷽"];
79 for c in &invalid {
80 assert_invalid_str(c);
81 }
82 }
83
84 #[test]
85 fn test_unicode_multiple_codepoints() {
86 let invalid = ["नी", "👨‍👨‍"];
87 for c in &invalid {
88 assert_invalid_str(c);
89 }
90 }
91
92 #[test]
93 fn test_valid_ascii_escape() {
94 let valid = [r"\'", r#"\""#, r"\\", r"\n", r"\r", r"\t", r"\0", "a", "b"];
95 for c in &valid {
96 assert_valid_str(c);
97 }
98 }
99
100 #[test]
101 fn test_invalid_ascii_escape() {
102 let invalid = [r"\a", r"\?", r"\"];
103 for c in &invalid {
104 assert_invalid_str(c);
105 }
106 }
107
108 #[test]
109 fn test_valid_ascii_code_escape() {
110 let valid = [r"\x00", r"\x7F", r"\x55", r"\xF0"];
111 for c in &valid {
112 assert_valid_str(c);
113 }
114 }
115
116 #[test]
117 fn test_invalid_ascii_code_escape() {
118 let invalid = [r"\x", r"\x7"];
119 for c in &invalid {
120 assert_invalid_str(c);
121 }
122 }
123
124 #[test]
125 fn test_invalid_unicode_escape() {
126 let well_formed = [
127 r"\u{FF}",
128 r"\u{0}",
129 r"\u{F}",
130 r"\u{10FFFF}",
131 r"\u{1_0__FF___FF_____}",
132 ];
133 for c in &well_formed {
134 assert_invalid_str(c);
135 }
136
137 let invalid = [
138 r"\u",
139 r"\u{}",
140 r"\u{",
141 r"\u{FF",
142 r"\u{FFFFFF}",
143 r"\u{_F}",
144 r"\u{00FFFFF}",
145 r"\u{110000}",
146 ];
147 for c in &invalid {
148 assert_invalid_str(c);
149 }
150 }
151
152 #[test]
153 fn test_mixed_invalid() {
154 assert_invalid_str(
155 r"This is the tale of a string
156with a newline in between, some emoji (👨‍👨‍) here and there,
157unicode escapes like this: \u{1FFBB} and weird stuff like
158this ﷽",
159 );
160 }
161
162 #[test]
163 fn test_mixed_valid() {
164 assert_valid_str(
165 r"This is the tale of a string
166with a newline in between, no emoji at all,
167nor unicode escapes or weird stuff",
168 );
169 }
170
171 #[test]
172 fn test_ignore_newline() {
173 assert_valid_str(
174 "Hello \
175 World",
176 );
177 }
178}
diff --git a/crates/ra_syntax/src/validation/char.rs b/crates/ra_syntax/src/validation/char.rs
index 63f9bad24..4728c85e6 100644
--- a/crates/ra_syntax/src/validation/char.rs
+++ b/crates/ra_syntax/src/validation/char.rs
@@ -1,3 +1,5 @@
1//! Validation of char literals
2
1use std::u32; 3use std::u32;
2 4
3use arrayvec::ArrayString; 5use arrayvec::ArrayString;
@@ -12,7 +14,7 @@ use crate::{
12 }, 14 },
13}; 15};
14 16
15pub(crate) fn validate_char_node(node: ast::Char, errors: &mut Vec<SyntaxError>) { 17pub(super) fn validate_char_node(node: ast::Char, errors: &mut Vec<SyntaxError>) {
16 let literal_text = node.text(); 18 let literal_text = node.text();
17 let literal_range = node.syntax().range(); 19 let literal_range = node.syntax().range();
18 let mut components = string_lexing::parse_char_literal(literal_text); 20 let mut components = string_lexing::parse_char_literal(literal_text);
@@ -37,7 +39,7 @@ pub(crate) fn validate_char_node(node: ast::Char, errors: &mut Vec<SyntaxError>)
37 } 39 }
38} 40}
39 41
40pub(crate) fn validate_char_component( 42pub(super) fn validate_char_component(
41 text: &str, 43 text: &str,
42 kind: CharComponentKind, 44 kind: CharComponentKind,
43 range: TextRange, 45 range: TextRange,
@@ -46,109 +48,115 @@ pub(crate) fn validate_char_component(
46 // Validate escapes 48 // Validate escapes
47 use self::CharComponentKind::*; 49 use self::CharComponentKind::*;
48 match kind { 50 match kind {
49 AsciiEscape => { 51 AsciiEscape => validate_ascii_escape(text, range, errors),
50 if text.len() == 1 { 52 AsciiCodeEscape => validate_ascii_code_escape(text, range, errors),
51 // Escape sequence consists only of leading `\` 53 UnicodeEscape => validate_unicode_escape(text, range, errors),
52 errors.push(SyntaxError::new(EmptyAsciiEscape, range)); 54 CodePoint => {
53 } else { 55 // These code points must always be escaped
54 let escape_code = text.chars().skip(1).next().unwrap(); 56 if text == "\t" || text == "\r" || text == "\n" {
55 if !is_ascii_escape(escape_code) { 57 errors.push(SyntaxError::new(UnescapedCodepoint, range));
56 errors.push(SyntaxError::new(InvalidAsciiEscape, range));
57 }
58 } 58 }
59 } 59 }
60 AsciiCodeEscape => { 60 }
61 // An AsciiCodeEscape has 4 chars, example: `\xDD` 61}
62 if text.len() < 4 { 62
63 errors.push(SyntaxError::new(TooShortAsciiCodeEscape, range)); 63fn validate_ascii_escape(text: &str, range: TextRange, errors: &mut Vec<SyntaxError>) {
64 } else { 64 if text.len() == 1 {
65 assert!( 65 // Escape sequence consists only of leading `\`
66 text.chars().count() == 4, 66 errors.push(SyntaxError::new(EmptyAsciiEscape, range));
67 "AsciiCodeEscape cannot be longer than 4 chars" 67 } else {
68 ); 68 let escape_code = text.chars().skip(1).next().unwrap();
69 69 if !is_ascii_escape(escape_code) {
70 match u8::from_str_radix(&text[2..], 16) { 70 errors.push(SyntaxError::new(InvalidAsciiEscape, range));
71 Ok(code) if code < 128 => { /* Escape code is valid */ }
72 Ok(_) => errors.push(SyntaxError::new(AsciiCodeEscapeOutOfRange, range)),
73 Err(_) => errors.push(SyntaxError::new(MalformedAsciiCodeEscape, range)),
74 }
75 }
76 } 71 }
77 UnicodeEscape => { 72 }
78 assert!(&text[..2] == "\\u", "UnicodeEscape always starts with \\u"); 73}
79 74
80 if text.len() == 2 { 75pub(super) fn is_ascii_escape(code: char) -> bool {
81 // No starting `{` 76 match code {
82 errors.push(SyntaxError::new(MalformedUnicodeEscape, range)); 77 '\\' | '\'' | '"' | 'n' | 'r' | 't' | '0' => true,
83 return; 78 _ => false,
84 } 79 }
80}
85 81
86 if text.len() == 3 { 82fn validate_ascii_code_escape(text: &str, range: TextRange, errors: &mut Vec<SyntaxError>) {
87 // Only starting `{` 83 // An AsciiCodeEscape has 4 chars, example: `\xDD`
88 errors.push(SyntaxError::new(UnclosedUnicodeEscape, range)); 84 if text.len() < 4 {
89 return; 85 errors.push(SyntaxError::new(TooShortAsciiCodeEscape, range));
90 } 86 } else {
87 assert!(
88 text.chars().count() == 4,
89 "AsciiCodeEscape cannot be longer than 4 chars"
90 );
91 91
92 let mut code = ArrayString::<[_; 6]>::new(); 92 match u8::from_str_radix(&text[2..], 16) {
93 let mut closed = false; 93 Ok(code) if code < 128 => { /* Escape code is valid */ }
94 for c in text[3..].chars() { 94 Ok(_) => errors.push(SyntaxError::new(AsciiCodeEscapeOutOfRange, range)),
95 assert!(!closed, "no characters after escape is closed"); 95 Err(_) => errors.push(SyntaxError::new(MalformedAsciiCodeEscape, range)),
96 96 }
97 if c.is_digit(16) { 97 }
98 if code.len() == 6 { 98}
99 errors.push(SyntaxError::new(OverlongUnicodeEscape, range));
100 return;
101 }
102
103 code.push(c);
104 } else if c == '_' {
105 // Reject leading _
106 if code.len() == 0 {
107 errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
108 return;
109 }
110 } else if c == '}' {
111 closed = true;
112 } else {
113 errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
114 return;
115 }
116 }
117 99
118 if !closed { 100fn validate_unicode_escape(text: &str, range: TextRange, errors: &mut Vec<SyntaxError>) {
119 errors.push(SyntaxError::new(UnclosedUnicodeEscape, range)) 101 assert!(&text[..2] == "\\u", "UnicodeEscape always starts with \\u");
120 }
121 102
122 if code.len() == 0 { 103 if text.len() == 2 {
123 errors.push(SyntaxError::new(EmptyUnicodeEcape, range)); 104 // No starting `{`
105 errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
106 return;
107 }
108
109 if text.len() == 3 {
110 // Only starting `{`
111 errors.push(SyntaxError::new(UnclosedUnicodeEscape, range));
112 return;
113 }
114
115 let mut code = ArrayString::<[_; 6]>::new();
116 let mut closed = false;
117 for c in text[3..].chars() {
118 assert!(!closed, "no characters after escape is closed");
119
120 if c.is_digit(16) {
121 if code.len() == 6 {
122 errors.push(SyntaxError::new(OverlongUnicodeEscape, range));
124 return; 123 return;
125 } 124 }
126 125
127 match u32::from_str_radix(&code, 16) { 126 code.push(c);
128 Ok(code_u32) if code_u32 > 0x10FFFF => { 127 } else if c == '_' {
129 errors.push(SyntaxError::new(UnicodeEscapeOutOfRange, range)); 128 // Reject leading _
130 } 129 if code.len() == 0 {
131 Ok(_) => { 130 errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
132 // Valid escape code 131 return;
133 }
134 Err(_) => {
135 errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
136 }
137 }
138 }
139 CodePoint => {
140 // These code points must always be escaped
141 if text == "\t" || text == "\r" {
142 errors.push(SyntaxError::new(UnescapedCodepoint, range));
143 } 132 }
133 } else if c == '}' {
134 closed = true;
135 } else {
136 errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
137 return;
144 } 138 }
145 } 139 }
146}
147 140
148fn is_ascii_escape(code: char) -> bool { 141 if !closed {
149 match code { 142 errors.push(SyntaxError::new(UnclosedUnicodeEscape, range))
150 '\\' | '\'' | '"' | 'n' | 'r' | 't' | '0' => true, 143 }
151 _ => false, 144
145 if code.len() == 0 {
146 errors.push(SyntaxError::new(EmptyUnicodeEcape, range));
147 return;
148 }
149
150 match u32::from_str_radix(&code, 16) {
151 Ok(code_u32) if code_u32 > 0x10FFFF => {
152 errors.push(SyntaxError::new(UnicodeEscapeOutOfRange, range));
153 }
154 Ok(_) => {
155 // Valid escape code
156 }
157 Err(_) => {
158 errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
159 }
152 } 160 }
153} 161}
154 162
@@ -205,9 +213,7 @@ mod test {
205 213
206 #[test] 214 #[test]
207 fn test_valid_ascii_escape() { 215 fn test_valid_ascii_escape() {
208 let valid = [ 216 let valid = [r"\'", "\"", "\\\\", "\\\"", r"\n", r"\r", r"\t", r"\0"];
209 r"\'", "\"", "\\\\", "\\\"", r"\n", r"\r", r"\t", r"\0", "a", "b",
210 ];
211 for c in &valid { 217 for c in &valid {
212 assert_valid_char(c); 218 assert_valid_char(c);
213 } 219 }
diff --git a/crates/ra_syntax/src/validation/mod.rs b/crates/ra_syntax/src/validation/mod.rs
index 2ff0bc26d..bdee8120c 100644
--- a/crates/ra_syntax/src/validation/mod.rs
+++ b/crates/ra_syntax/src/validation/mod.rs
@@ -5,6 +5,8 @@ use crate::{
5 yellow::SyntaxError, 5 yellow::SyntaxError,
6}; 6};
7 7
8mod byte;
9mod byte_string;
8mod char; 10mod char;
9mod string; 11mod string;
10 12
@@ -12,6 +14,8 @@ pub(crate) fn validate(file: &SourceFileNode) -> Vec<SyntaxError> {
12 let mut errors = Vec::new(); 14 let mut errors = Vec::new();
13 for node in file.syntax().descendants() { 15 for node in file.syntax().descendants() {
14 let _ = visitor_ctx(&mut errors) 16 let _ = visitor_ctx(&mut errors)
17 .visit::<ast::Byte, _>(self::byte::validate_byte_node)
18 .visit::<ast::ByteString, _>(self::byte_string::validate_byte_string_node)
15 .visit::<ast::Char, _>(self::char::validate_char_node) 19 .visit::<ast::Char, _>(self::char::validate_char_node)
16 .visit::<ast::String, _>(self::string::validate_string_node) 20 .visit::<ast::String, _>(self::string::validate_string_node)
17 .accept(node); 21 .accept(node);