1 | use crate::{ast, hir, Error}; |
2 | |
3 | /// A convenience routine for parsing a regex using default options. |
4 | /// |
5 | /// This is equivalent to `Parser::new().parse(pattern)`. |
6 | /// |
7 | /// If you need to set non-default options, then use a [`ParserBuilder`]. |
8 | /// |
9 | /// This routine returns an [`Hir`](hir::Hir) value. Namely, it automatically |
10 | /// parses the pattern as an [`Ast`](ast::Ast) and then invokes the translator |
11 | /// to convert the `Ast` into an `Hir`. If you need access to the `Ast`, then |
12 | /// you should use a [`ast::parse::Parser`]. |
13 | pub fn parse(pattern: &str) -> Result<hir::Hir, Error> { |
14 | Parser::new().parse(pattern) |
15 | } |
16 | |
17 | /// A builder for a regular expression parser. |
18 | /// |
19 | /// This builder permits modifying configuration options for the parser. |
20 | /// |
21 | /// This type combines the builder options for both the [AST |
22 | /// `ParserBuilder`](ast::parse::ParserBuilder) and the [HIR |
23 | /// `TranslatorBuilder`](hir::translate::TranslatorBuilder). |
24 | #[derive (Clone, Debug, Default)] |
25 | pub struct ParserBuilder { |
26 | ast: ast::parse::ParserBuilder, |
27 | hir: hir::translate::TranslatorBuilder, |
28 | } |
29 | |
30 | impl ParserBuilder { |
31 | /// Create a new parser builder with a default configuration. |
32 | pub fn new() -> ParserBuilder { |
33 | ParserBuilder::default() |
34 | } |
35 | |
36 | /// Build a parser from this configuration with the given pattern. |
37 | pub fn build(&self) -> Parser { |
38 | Parser { ast: self.ast.build(), hir: self.hir.build() } |
39 | } |
40 | |
41 | /// Set the nesting limit for this parser. |
42 | /// |
43 | /// The nesting limit controls how deep the abstract syntax tree is allowed |
44 | /// to be. If the AST exceeds the given limit (e.g., with too many nested |
45 | /// groups), then an error is returned by the parser. |
46 | /// |
47 | /// The purpose of this limit is to act as a heuristic to prevent stack |
48 | /// overflow for consumers that do structural induction on an `Ast` using |
49 | /// explicit recursion. While this crate never does this (instead using |
50 | /// constant stack space and moving the call stack to the heap), other |
51 | /// crates may. |
52 | /// |
53 | /// This limit is not checked until the entire Ast is parsed. Therefore, |
54 | /// if callers want to put a limit on the amount of heap space used, then |
55 | /// they should impose a limit on the length, in bytes, of the concrete |
56 | /// pattern string. In particular, this is viable since this parser |
57 | /// implementation will limit itself to heap space proportional to the |
58 | /// length of the pattern string. |
59 | /// |
60 | /// Note that a nest limit of `0` will return a nest limit error for most |
61 | /// patterns but not all. For example, a nest limit of `0` permits `a` but |
62 | /// not `ab`, since `ab` requires a concatenation, which results in a nest |
63 | /// depth of `1`. In general, a nest limit is not something that manifests |
64 | /// in an obvious way in the concrete syntax, therefore, it should not be |
65 | /// used in a granular way. |
66 | pub fn nest_limit(&mut self, limit: u32) -> &mut ParserBuilder { |
67 | self.ast.nest_limit(limit); |
68 | self |
69 | } |
70 | |
71 | /// Whether to support octal syntax or not. |
72 | /// |
73 | /// Octal syntax is a little-known way of uttering Unicode codepoints in |
74 | /// a regular expression. For example, `a`, `\x61`, `\u0061` and |
75 | /// `\141` are all equivalent regular expressions, where the last example |
76 | /// shows octal syntax. |
77 | /// |
78 | /// While supporting octal syntax isn't in and of itself a problem, it does |
79 | /// make good error messages harder. That is, in PCRE based regex engines, |
80 | /// syntax like `\0` invokes a backreference, which is explicitly |
81 | /// unsupported in Rust's regex engine. However, many users expect it to |
82 | /// be supported. Therefore, when octal support is disabled, the error |
83 | /// message will explicitly mention that backreferences aren't supported. |
84 | /// |
85 | /// Octal syntax is disabled by default. |
86 | pub fn octal(&mut self, yes: bool) -> &mut ParserBuilder { |
87 | self.ast.octal(yes); |
88 | self |
89 | } |
90 | |
91 | /// When disabled, translation will permit the construction of a regular |
92 | /// expression that may match invalid UTF-8. |
93 | /// |
94 | /// When enabled (the default), the translator is guaranteed to produce an |
95 | /// expression that, for non-empty matches, will only ever produce spans |
96 | /// that are entirely valid UTF-8 (otherwise, the translator will return an |
97 | /// error). |
98 | /// |
99 | /// Perhaps surprisingly, when UTF-8 is enabled, an empty regex or even |
100 | /// a negated ASCII word boundary (uttered as `(?-u:\B)` in the concrete |
101 | /// syntax) will be allowed even though they can produce matches that split |
102 | /// a UTF-8 encoded codepoint. This only applies to zero-width or "empty" |
103 | /// matches, and it is expected that the regex engine itself must handle |
104 | /// these cases if necessary (perhaps by suppressing any zero-width matches |
105 | /// that split a codepoint). |
106 | pub fn utf8(&mut self, yes: bool) -> &mut ParserBuilder { |
107 | self.hir.utf8(yes); |
108 | self |
109 | } |
110 | |
111 | /// Enable verbose mode in the regular expression. |
112 | /// |
113 | /// When enabled, verbose mode permits insignificant whitespace in many |
114 | /// places in the regular expression, as well as comments. Comments are |
115 | /// started using `#` and continue until the end of the line. |
116 | /// |
117 | /// By default, this is disabled. It may be selectively enabled in the |
118 | /// regular expression by using the `x` flag regardless of this setting. |
119 | pub fn ignore_whitespace(&mut self, yes: bool) -> &mut ParserBuilder { |
120 | self.ast.ignore_whitespace(yes); |
121 | self |
122 | } |
123 | |
124 | /// Enable or disable the case insensitive flag by default. |
125 | /// |
126 | /// By default this is disabled. It may alternatively be selectively |
127 | /// enabled in the regular expression itself via the `i` flag. |
128 | pub fn case_insensitive(&mut self, yes: bool) -> &mut ParserBuilder { |
129 | self.hir.case_insensitive(yes); |
130 | self |
131 | } |
132 | |
133 | /// Enable or disable the multi-line matching flag by default. |
134 | /// |
135 | /// By default this is disabled. It may alternatively be selectively |
136 | /// enabled in the regular expression itself via the `m` flag. |
137 | pub fn multi_line(&mut self, yes: bool) -> &mut ParserBuilder { |
138 | self.hir.multi_line(yes); |
139 | self |
140 | } |
141 | |
142 | /// Enable or disable the "dot matches any character" flag by default. |
143 | /// |
144 | /// By default this is disabled. It may alternatively be selectively |
145 | /// enabled in the regular expression itself via the `s` flag. |
146 | pub fn dot_matches_new_line(&mut self, yes: bool) -> &mut ParserBuilder { |
147 | self.hir.dot_matches_new_line(yes); |
148 | self |
149 | } |
150 | |
151 | /// Enable or disable the CRLF mode flag by default. |
152 | /// |
153 | /// By default this is disabled. It may alternatively be selectively |
154 | /// enabled in the regular expression itself via the `R` flag. |
155 | /// |
156 | /// When CRLF mode is enabled, the following happens: |
157 | /// |
158 | /// * Unless `dot_matches_new_line` is enabled, `.` will match any character |
159 | /// except for `\r` and `\n`. |
160 | /// * When `multi_line` mode is enabled, `^` and `$` will treat `\r\n`, |
161 | /// `\r` and `\n` as line terminators. And in particular, neither will |
162 | /// match between a `\r` and a `\n`. |
163 | pub fn crlf(&mut self, yes: bool) -> &mut ParserBuilder { |
164 | self.hir.crlf(yes); |
165 | self |
166 | } |
167 | |
168 | /// Sets the line terminator for use with `(?u-s:.)` and `(?-us:.)`. |
169 | /// |
170 | /// Namely, instead of `.` (by default) matching everything except for `\n`, |
171 | /// this will cause `.` to match everything except for the byte given. |
172 | /// |
173 | /// If `.` is used in a context where Unicode mode is enabled and this byte |
174 | /// isn't ASCII, then an error will be returned. When Unicode mode is |
175 | /// disabled, then any byte is permitted, but will return an error if UTF-8 |
176 | /// mode is enabled and it is a non-ASCII byte. |
177 | /// |
178 | /// In short, any ASCII value for a line terminator is always okay. But a |
179 | /// non-ASCII byte might result in an error depending on whether Unicode |
180 | /// mode or UTF-8 mode are enabled. |
181 | /// |
182 | /// Note that if `R` mode is enabled then it always takes precedence and |
183 | /// the line terminator will be treated as `\r` and `\n` simultaneously. |
184 | /// |
185 | /// Note also that this *doesn't* impact the look-around assertions |
186 | /// `(?m:^)` and `(?m:$)`. That's usually controlled by additional |
187 | /// configuration in the regex engine itself. |
188 | pub fn line_terminator(&mut self, byte: u8) -> &mut ParserBuilder { |
189 | self.hir.line_terminator(byte); |
190 | self |
191 | } |
192 | |
193 | /// Enable or disable the "swap greed" flag by default. |
194 | /// |
195 | /// By default this is disabled. It may alternatively be selectively |
196 | /// enabled in the regular expression itself via the `U` flag. |
197 | pub fn swap_greed(&mut self, yes: bool) -> &mut ParserBuilder { |
198 | self.hir.swap_greed(yes); |
199 | self |
200 | } |
201 | |
202 | /// Enable or disable the Unicode flag (`u`) by default. |
203 | /// |
204 | /// By default this is **enabled**. It may alternatively be selectively |
205 | /// disabled in the regular expression itself via the `u` flag. |
206 | /// |
207 | /// Note that unless `utf8` is disabled (it's enabled by default), a |
208 | /// regular expression will fail to parse if Unicode mode is disabled and a |
209 | /// sub-expression could possibly match invalid UTF-8. |
210 | pub fn unicode(&mut self, yes: bool) -> &mut ParserBuilder { |
211 | self.hir.unicode(yes); |
212 | self |
213 | } |
214 | } |
215 | |
216 | /// A convenience parser for regular expressions. |
217 | /// |
218 | /// This parser takes as input a regular expression pattern string (the |
219 | /// "concrete syntax") and returns a high-level intermediate representation |
220 | /// (the HIR) suitable for most types of analysis. In particular, this parser |
221 | /// hides the intermediate state of producing an AST (the "abstract syntax"). |
222 | /// The AST is itself far more complex than the HIR, so this parser serves as a |
223 | /// convenience for never having to deal with it at all. |
224 | /// |
225 | /// If callers have more fine grained use cases that need an AST, then please |
226 | /// see the [`ast::parse`] module. |
227 | /// |
228 | /// A `Parser` can be configured in more detail via a [`ParserBuilder`]. |
229 | #[derive (Clone, Debug)] |
230 | pub struct Parser { |
231 | ast: ast::parse::Parser, |
232 | hir: hir::translate::Translator, |
233 | } |
234 | |
235 | impl Parser { |
236 | /// Create a new parser with a default configuration. |
237 | /// |
238 | /// The parser can be run with `parse` method. The parse method returns |
239 | /// a high level intermediate representation of the given regular |
240 | /// expression. |
241 | /// |
242 | /// To set configuration options on the parser, use [`ParserBuilder`]. |
243 | pub fn new() -> Parser { |
244 | ParserBuilder::new().build() |
245 | } |
246 | |
247 | /// Parse the regular expression into a high level intermediate |
248 | /// representation. |
249 | pub fn parse(&mut self, pattern: &str) -> Result<hir::Hir, Error> { |
250 | let ast: Ast = self.ast.parse(pattern)?; |
251 | let hir: Hir = self.hir.translate(pattern, &ast)?; |
252 | Ok(hir) |
253 | } |
254 | } |
255 | |