|
jpayne@69
|
1 // Copyright (c) 2005, Google Inc.
|
|
jpayne@69
|
2 // All rights reserved.
|
|
jpayne@69
|
3 //
|
|
jpayne@69
|
4 // Redistribution and use in source and binary forms, with or without
|
|
jpayne@69
|
5 // modification, are permitted provided that the following conditions are
|
|
jpayne@69
|
6 // met:
|
|
jpayne@69
|
7 //
|
|
jpayne@69
|
8 // * Redistributions of source code must retain the above copyright
|
|
jpayne@69
|
9 // notice, this list of conditions and the following disclaimer.
|
|
jpayne@69
|
10 // * Redistributions in binary form must reproduce the above
|
|
jpayne@69
|
11 // copyright notice, this list of conditions and the following disclaimer
|
|
jpayne@69
|
12 // in the documentation and/or other materials provided with the
|
|
jpayne@69
|
13 // distribution.
|
|
jpayne@69
|
14 // * Neither the name of Google Inc. nor the names of its
|
|
jpayne@69
|
15 // contributors may be used to endorse or promote products derived from
|
|
jpayne@69
|
16 // this software without specific prior written permission.
|
|
jpayne@69
|
17 //
|
|
jpayne@69
|
18 // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
|
|
jpayne@69
|
19 // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
|
|
jpayne@69
|
20 // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
|
|
jpayne@69
|
21 // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
|
|
jpayne@69
|
22 // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
|
|
jpayne@69
|
23 // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
|
|
jpayne@69
|
24 // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
|
|
jpayne@69
|
25 // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
|
|
jpayne@69
|
26 // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
|
|
jpayne@69
|
27 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
|
|
jpayne@69
|
28 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
jpayne@69
|
29 //
|
|
jpayne@69
|
30 // Author: Sanjay Ghemawat
|
|
jpayne@69
|
31 //
|
|
jpayne@69
|
32 // Regular-expression based scanner for parsing an input stream.
|
|
jpayne@69
|
33 //
|
|
jpayne@69
|
34 // Example 1: parse a sequence of "var = number" entries from input:
|
|
jpayne@69
|
35 //
|
|
jpayne@69
|
36 // Scanner scanner(input);
|
|
jpayne@69
|
37 // string var;
|
|
jpayne@69
|
38 // int number;
|
|
jpayne@69
|
39 // scanner.SetSkipExpression("\\s+"); // Skip any white space we encounter
|
|
jpayne@69
|
40 // while (scanner.Consume("(\\w+) = (\\d+)", &var, &number)) {
|
|
jpayne@69
|
41 // ...;
|
|
jpayne@69
|
42 // }
|
|
jpayne@69
|
43
|
|
jpayne@69
|
44 #ifndef _PCRE_SCANNER_H
|
|
jpayne@69
|
45 #define _PCRE_SCANNER_H
|
|
jpayne@69
|
46
|
|
jpayne@69
|
47 #include <assert.h>
|
|
jpayne@69
|
48 #include <string>
|
|
jpayne@69
|
49 #include <vector>
|
|
jpayne@69
|
50
|
|
jpayne@69
|
51 #include <pcrecpp.h>
|
|
jpayne@69
|
52 #include <pcre_stringpiece.h>
|
|
jpayne@69
|
53
|
|
jpayne@69
|
54 namespace pcrecpp {
|
|
jpayne@69
|
55
|
|
jpayne@69
|
56 class PCRECPP_EXP_DEFN Scanner {
|
|
jpayne@69
|
57 public:
|
|
jpayne@69
|
58 Scanner();
|
|
jpayne@69
|
59 explicit Scanner(const std::string& input);
|
|
jpayne@69
|
60 ~Scanner();
|
|
jpayne@69
|
61
|
|
jpayne@69
|
62 // Return current line number. The returned line-number is
|
|
jpayne@69
|
63 // one-based. I.e. it returns 1 + the number of consumed newlines.
|
|
jpayne@69
|
64 //
|
|
jpayne@69
|
65 // Note: this method may be slow. It may take time proportional to
|
|
jpayne@69
|
66 // the size of the input.
|
|
jpayne@69
|
67 int LineNumber() const;
|
|
jpayne@69
|
68
|
|
jpayne@69
|
69 // Return the byte-offset that the scanner is looking in the
|
|
jpayne@69
|
70 // input data;
|
|
jpayne@69
|
71 int Offset() const;
|
|
jpayne@69
|
72
|
|
jpayne@69
|
73 // Return true iff the start of the remaining input matches "re"
|
|
jpayne@69
|
74 bool LookingAt(const RE& re) const;
|
|
jpayne@69
|
75
|
|
jpayne@69
|
76 // Return true iff all of the following are true
|
|
jpayne@69
|
77 // a. the start of the remaining input matches "re",
|
|
jpayne@69
|
78 // b. if any arguments are supplied, matched sub-patterns can be
|
|
jpayne@69
|
79 // parsed and stored into the arguments.
|
|
jpayne@69
|
80 // If it returns true, it skips over the matched input and any
|
|
jpayne@69
|
81 // following input that matches the "skip" regular expression.
|
|
jpayne@69
|
82 bool Consume(const RE& re,
|
|
jpayne@69
|
83 const Arg& arg0 = RE::no_arg,
|
|
jpayne@69
|
84 const Arg& arg1 = RE::no_arg,
|
|
jpayne@69
|
85 const Arg& arg2 = RE::no_arg
|
|
jpayne@69
|
86 // TODO: Allow more arguments?
|
|
jpayne@69
|
87 );
|
|
jpayne@69
|
88
|
|
jpayne@69
|
89 // Set the "skip" regular expression. If after consuming some data,
|
|
jpayne@69
|
90 // a prefix of the input matches this RE, it is automatically
|
|
jpayne@69
|
91 // skipped. For example, a programming language scanner would use
|
|
jpayne@69
|
92 // a skip RE that matches white space and comments.
|
|
jpayne@69
|
93 //
|
|
jpayne@69
|
94 // scanner.SetSkipExpression("\\s+|//.*|/[*](.|\n)*?[*]/");
|
|
jpayne@69
|
95 //
|
|
jpayne@69
|
96 // Skipping repeats as long as it succeeds. We used to let people do
|
|
jpayne@69
|
97 // this by writing "(...)*" in the regular expression, but that added
|
|
jpayne@69
|
98 // up to lots of recursive calls within the pcre library, so now we
|
|
jpayne@69
|
99 // control repetition explicitly via the function call API.
|
|
jpayne@69
|
100 //
|
|
jpayne@69
|
101 // You can pass NULL for "re" if you do not want any data to be skipped.
|
|
jpayne@69
|
102 void Skip(const char* re); // DEPRECATED; does *not* repeat
|
|
jpayne@69
|
103 void SetSkipExpression(const char* re);
|
|
jpayne@69
|
104
|
|
jpayne@69
|
105 // Temporarily pause "skip"ing. This
|
|
jpayne@69
|
106 // Skip("Foo"); code ; DisableSkip(); code; EnableSkip()
|
|
jpayne@69
|
107 // is similar to
|
|
jpayne@69
|
108 // Skip("Foo"); code ; Skip(NULL); code ; Skip("Foo");
|
|
jpayne@69
|
109 // but avoids creating/deleting new RE objects.
|
|
jpayne@69
|
110 void DisableSkip();
|
|
jpayne@69
|
111
|
|
jpayne@69
|
112 // Reenable previously paused skipping. Any prefix of the input
|
|
jpayne@69
|
113 // that matches the skip pattern is immediately dropped.
|
|
jpayne@69
|
114 void EnableSkip();
|
|
jpayne@69
|
115
|
|
jpayne@69
|
116 /***** Special wrappers around SetSkip() for some common idioms *****/
|
|
jpayne@69
|
117
|
|
jpayne@69
|
118 // Arranges to skip whitespace, C comments, C++ comments.
|
|
jpayne@69
|
119 // The overall RE is a disjunction of the following REs:
|
|
jpayne@69
|
120 // \\s whitespace
|
|
jpayne@69
|
121 // //.*\n C++ comment
|
|
jpayne@69
|
122 // /[*](.|\n)*?[*]/ C comment (x*? means minimal repetitions of x)
|
|
jpayne@69
|
123 // We get repetition via the semantics of SetSkipExpression, not by using *
|
|
jpayne@69
|
124 void SkipCXXComments() {
|
|
jpayne@69
|
125 SetSkipExpression("\\s|//.*\n|/[*](?:\n|.)*?[*]/");
|
|
jpayne@69
|
126 }
|
|
jpayne@69
|
127
|
|
jpayne@69
|
128 void set_save_comments(bool comments) {
|
|
jpayne@69
|
129 save_comments_ = comments;
|
|
jpayne@69
|
130 }
|
|
jpayne@69
|
131
|
|
jpayne@69
|
132 bool save_comments() {
|
|
jpayne@69
|
133 return save_comments_;
|
|
jpayne@69
|
134 }
|
|
jpayne@69
|
135
|
|
jpayne@69
|
136 // Append to vector ranges the comments found in the
|
|
jpayne@69
|
137 // byte range [start,end] (inclusive) of the input data.
|
|
jpayne@69
|
138 // Only comments that were extracted entirely within that
|
|
jpayne@69
|
139 // range are returned: no range splitting of atomically-extracted
|
|
jpayne@69
|
140 // comments is performed.
|
|
jpayne@69
|
141 void GetComments(int start, int end, std::vector<StringPiece> *ranges);
|
|
jpayne@69
|
142
|
|
jpayne@69
|
143 // Append to vector ranges the comments added
|
|
jpayne@69
|
144 // since the last time this was called. This
|
|
jpayne@69
|
145 // functionality is provided for efficiency when
|
|
jpayne@69
|
146 // interleaving scanning with parsing.
|
|
jpayne@69
|
147 void GetNextComments(std::vector<StringPiece> *ranges);
|
|
jpayne@69
|
148
|
|
jpayne@69
|
149 private:
|
|
jpayne@69
|
150 std::string data_; // All the input data
|
|
jpayne@69
|
151 StringPiece input_; // Unprocessed input
|
|
jpayne@69
|
152 RE* skip_; // If non-NULL, RE for skipping input
|
|
jpayne@69
|
153 bool should_skip_; // If true, use skip_
|
|
jpayne@69
|
154 bool skip_repeat_; // If true, repeat skip_ as long as it works
|
|
jpayne@69
|
155 bool save_comments_; // If true, aggregate the skip expression
|
|
jpayne@69
|
156
|
|
jpayne@69
|
157 // the skipped comments
|
|
jpayne@69
|
158 // TODO: later consider requiring that the StringPieces be added
|
|
jpayne@69
|
159 // in order by their start position
|
|
jpayne@69
|
160 std::vector<StringPiece> *comments_;
|
|
jpayne@69
|
161
|
|
jpayne@69
|
162 // the offset into comments_ that has been returned by GetNextComments
|
|
jpayne@69
|
163 int comments_offset_;
|
|
jpayne@69
|
164
|
|
jpayne@69
|
165 // helper function to consume *skip_ and honour
|
|
jpayne@69
|
166 // save_comments_
|
|
jpayne@69
|
167 void ConsumeSkip();
|
|
jpayne@69
|
168 };
|
|
jpayne@69
|
169
|
|
jpayne@69
|
170 } // namespace pcrecpp
|
|
jpayne@69
|
171
|
|
jpayne@69
|
172 #endif /* _PCRE_SCANNER_H */
|
