1 |
// |
// |
2 |
// assembly: System |
// assembly: System |
3 |
// namespace: System.Text.RegularExpressions |
// namespace: System.Text.RegularExpressions |
4 |
// file: replace.cs |
// file: replace.cs |
5 |
// |
// |
6 |
// author: Dan Lewis (dlewis@gmx.co.uk) |
// author: Dan Lewis (dlewis@gmx.co.uk) |
7 |
// (c) 2002 |
// (c) 2002 |
8 |
|
|
9 |
using System; |
// |
10 |
using System.Text; |
// Permission is hereby granted, free of charge, to any person obtaining |
11 |
using System.Collections; |
// a copy of this software and associated documentation files (the |
12 |
|
// "Software"), to deal in the Software without restriction, including |
13 |
using Parser = System.Text.RegularExpressions.Syntax.Parser; |
// without limitation the rights to use, copy, modify, merge, publish, |
14 |
|
// distribute, sublicense, and/or sell copies of the Software, and to |
15 |
namespace System.Text.RegularExpressions { |
// permit persons to whom the Software is furnished to do so, subject to |
16 |
|
// the following conditions: |
17 |
class ReplacementEvaluator { |
// |
18 |
public static string Evaluate (string replacement, Match match) { |
// The above copyright notice and this permission notice shall be |
19 |
ReplacementEvaluator ev = new ReplacementEvaluator (match.Regex, replacement); |
// included in all copies or substantial portions of the Software. |
20 |
return ev.Evaluate (match); |
// |
21 |
} |
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, |
22 |
|
// EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF |
23 |
public ReplacementEvaluator (Regex regex, string replacement) { |
// MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND |
24 |
this.regex = regex; |
// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE |
25 |
terms = new ArrayList (); |
// LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION |
26 |
Compile (replacement); |
// OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION |
27 |
} |
// WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. |
28 |
|
// |
29 |
public string Evaluate (Match match) { |
|
30 |
StringBuilder result = new StringBuilder (); |
using System; |
31 |
foreach (Term term in terms) |
using System.Text; |
32 |
result.Append (term.GetResult (match)); |
using System.Collections; |
33 |
|
|
34 |
return result.ToString (); |
using Parser = System.Text.RegularExpressions.Syntax.Parser; |
35 |
} |
|
36 |
|
namespace System.Text.RegularExpressions { |
37 |
// private |
|
38 |
|
class ReplacementEvaluator { |
39 |
private void Compile (string replacement) { |
public static string Evaluate (string replacement, Match match) { |
40 |
replacement = Parser.Unescape (replacement); |
ReplacementEvaluator ev = new ReplacementEvaluator (match.Regex, replacement); |
41 |
string literal = ""; |
return ev.Evaluate (match); |
42 |
|
} |
43 |
int ptr = 0; |
|
44 |
char c; |
public ReplacementEvaluator (Regex regex, string replacement) { |
45 |
Term term = null; |
this.regex = regex; |
46 |
while (ptr < replacement.Length) { |
this.replacement = replacement; |
47 |
c = replacement[ptr ++]; |
this.pieces = null; |
48 |
|
this.n_pieces = 0; |
49 |
if (c == '$') { |
Compile (); |
50 |
if (replacement[ptr] != '$') |
} |
51 |
term = CompileTerm (replacement, ref ptr); |
|
52 |
else |
public string Evaluate (Match match) |
53 |
++ ptr; |
{ |
54 |
} |
StringBuilder sb = new StringBuilder (); |
55 |
|
EvaluateAppend (match, sb); |
56 |
if (term != null) { |
return sb.ToString (); |
57 |
term.Literal = literal; |
} |
58 |
terms.Add (term); |
|
59 |
|
public void EvaluateAppend (Match match, StringBuilder sb) |
60 |
term = null; |
{ |
61 |
literal = ""; |
int i = 0, k, count; |
62 |
} |
|
63 |
else |
if (n_pieces == 0) { |
64 |
literal += c; |
sb.Append (replacement); |
65 |
} |
return; |
66 |
|
} |
67 |
if (term == null && literal.Length > 0) { |
|
68 |
terms.Add (new Term (literal)); |
while (i < n_pieces) { |
69 |
} |
k = pieces [i++]; |
70 |
} |
if (k >= 0) { |
71 |
|
count = pieces [i++]; |
72 |
private Term CompileTerm (string str, ref int ptr) { |
sb.Append (replacement, k, count); |
73 |
char c = str[ptr]; |
} else if (k < -3) { |
74 |
|
Group group = match.Groups [-(k + 4)]; |
75 |
if (Char.IsDigit (c)) { // numbered group |
sb.Append (group.Text, group.Index, group.Length); |
76 |
int n = Parser.ParseDecimal (str, ref ptr); |
} else if (k == -1) { |
77 |
if (n < 0 || n > regex.GroupCount) |
sb.Append (match.Text); |
78 |
throw new ArgumentException ("Bad group number."); |
} else if (k == -2) { |
79 |
|
sb.Append (match.Text, 0, match.Index); |
80 |
return new Term (TermOp.Match, n); |
} else { // k == -3 |
81 |
} |
int matchend = match.Index + match.Length; |
82 |
|
sb.Append (match.Text, matchend, match.Text.Length - matchend); |
83 |
++ ptr; |
} |
84 |
|
} |
85 |
switch (c) { |
} |
86 |
case '{': { // named group |
|
87 |
string name = Parser.ParseName (str, ref ptr); |
void Ensure (int size) |
88 |
if (str[ptr ++] != '}' || name == null) |
{ |
89 |
throw new ArgumentException ("Bad group name."); |
int new_size; |
90 |
|
if (pieces == null) { |
91 |
int n = regex.GroupNumberFromName (name); |
new_size = 4; |
92 |
|
if (new_size < size) |
93 |
if (n < 0) |
new_size = size; |
94 |
throw new ArgumentException ("Bad group name."); |
pieces = new int [new_size]; |
95 |
|
} else if (size >= pieces.Length) { |
96 |
return new Term (TermOp.Match, n); |
new_size = pieces.Length + (pieces.Length >> 1); |
97 |
} |
if (new_size < size) |
98 |
|
new_size = size; |
99 |
case '&': // entire match |
int [] new_pieces = new int [new_size]; |
100 |
return new Term (TermOp.Match, 0); |
Array.Copy (pieces, new_pieces, n_pieces); |
101 |
|
pieces = new_pieces; |
102 |
case '`': // text before match |
} |
103 |
return new Term (TermOp.PreMatch, 0); |
} |
104 |
|
|
105 |
case '\'': // text after match |
void AddFromReplacement (int start, int end) |
106 |
return new Term (TermOp.PostMatch, 0); |
{ |
107 |
|
if (start == end) |
108 |
case '+': // last group |
return; |
109 |
return new Term (TermOp.Match, regex.GroupCount - 1); |
Ensure (n_pieces + 2); |
110 |
|
pieces [n_pieces++] = start; |
111 |
case '_': // entire text |
pieces [n_pieces++] = end - start; |
112 |
return new Term (TermOp.All, 0); |
} |
113 |
|
|
114 |
default: |
void AddInt (int i) |
115 |
throw new ArgumentException ("Bad replacement pattern."); |
{ |
116 |
} |
Ensure (n_pieces + 1); |
117 |
} |
pieces [n_pieces++] = i; |
118 |
|
} |
119 |
private Regex regex; |
|
120 |
private ArrayList terms; |
// private |
121 |
|
private void Compile () { |
122 |
private enum TermOp { |
replacement = Parser.Unescape (replacement); |
123 |
None, // no action |
|
124 |
Match, // input within group |
int anchor = 0, ptr = 0, saveptr; |
125 |
PreMatch, // input before group |
char c; |
126 |
PostMatch, // input after group |
while (ptr < replacement.Length) { |
127 |
All // entire input |
c = replacement [ptr++]; |
128 |
} |
|
129 |
|
if (c != '$') |
130 |
private class Term { |
continue; |
131 |
public Term (TermOp op, int arg) { |
|
132 |
this.op = op; |
// If the '$' was the last character, just emit it as is |
133 |
this.arg = arg; |
if (ptr == replacement.Length) |
134 |
this.literal = ""; |
break; |
135 |
} |
|
136 |
|
// If we saw a '$$' |
137 |
public Term (string literal) { |
if (replacement [ptr] == '$') { |
138 |
this.op = TermOp.None; |
// Everthing from 'anchor' upto and including the first '$' is copied from the replacement string |
139 |
this.arg = 0; |
AddFromReplacement (anchor, ptr); |
140 |
this.literal = literal; |
// skip over the second '$'. |
141 |
} |
anchor = ++ptr; |
142 |
|
continue; |
143 |
public string Literal { |
} |
144 |
set { literal = value; } |
|
145 |
} |
saveptr = ptr - 1; |
146 |
|
|
147 |
public string GetResult (Match match) { |
int from_match = CompileTerm (ref ptr); |
148 |
Group group = match.Groups[arg]; |
|
149 |
|
// We couldn't recognize the term following the '$'. Just treat it as a literal. |
150 |
switch (op) { |
// 'ptr' has already been advanced, no need to rewind it back |
151 |
case TermOp.None: |
if (from_match >= 0) |
152 |
return literal; |
continue; |
153 |
|
|
154 |
case TermOp.Match: |
AddFromReplacement (anchor, saveptr); |
155 |
return literal + group.Value; |
AddInt (from_match); |
156 |
|
anchor = ptr; |
157 |
case TermOp.PreMatch: |
} |
158 |
return literal + group.Text.Substring (0, group.Index); |
|
159 |
|
// If we never needed to advance anchor, it means the result is the whole replacement string. |
160 |
case TermOp.PostMatch: |
// We optimize that case by never allocating the pieces array. |
161 |
return literal + group.Text.Substring (group.Index + group.Length); |
if (anchor != 0) |
162 |
|
AddFromReplacement (anchor, ptr); |
163 |
case TermOp.All: |
} |
164 |
return literal + group.Text; |
|
165 |
} |
private int CompileTerm (ref int ptr) { |
166 |
|
char c = replacement [ptr]; |
167 |
return ""; |
|
168 |
} |
if (Char.IsDigit (c)) { // numbered group |
169 |
|
int n = Parser.ParseDecimal (replacement, ref ptr); |
170 |
public TermOp op; // term type |
if (n < 0 || n > regex.GroupCount) |
171 |
public int arg; // group argument |
return 0; |
172 |
public string literal; // literal to prepend |
|
173 |
|
return -n - 4; |
174 |
public override string ToString () { |
} |
175 |
return op.ToString () + "(" + arg + ") " + literal; |
|
176 |
} |
++ ptr; |
177 |
} |
|
178 |
} |
switch (c) { |
179 |
} |
case '{': { // named group |
180 |
|
string name; |
181 |
|
int n = -1; |
182 |
|
|
183 |
|
try { |
184 |
|
// The parser is written such that there are few explicit range checks |
185 |
|
// and depends on 'IndexOutOfRangeException' being thrown. |
186 |
|
|
187 |
|
if (Char.IsDigit (replacement [ptr])) { |
188 |
|
n = Parser.ParseDecimal (replacement, ref ptr); |
189 |
|
name = ""; |
190 |
|
} else { |
191 |
|
name = Parser.ParseName (replacement, ref ptr); |
192 |
|
} |
193 |
|
} catch (IndexOutOfRangeException) { |
194 |
|
ptr = replacement.Length; |
195 |
|
return 0; |
196 |
|
} |
197 |
|
|
198 |
|
if (ptr == replacement.Length || replacement[ptr] != '}' || name == null) |
199 |
|
return 0; |
200 |
|
++ptr; // Swallow the '}' |
201 |
|
|
202 |
|
if (name != "") |
203 |
|
n = regex.GroupNumberFromName (name); |
204 |
|
|
205 |
|
if (n < 0 || n > regex.GroupCount) |
206 |
|
return 0; |
207 |
|
|
208 |
|
return -n - 4; |
209 |
|
} |
210 |
|
|
211 |
|
case '&': // entire match. Value should be same as $0 |
212 |
|
return -4; |
213 |
|
|
214 |
|
case '`': // text before match |
215 |
|
return -2; |
216 |
|
|
217 |
|
case '\'': // text after match |
218 |
|
return -3; |
219 |
|
|
220 |
|
case '+': // last group |
221 |
|
return -regex.GroupCount - 4; |
222 |
|
|
223 |
|
case '_': // entire text |
224 |
|
return -1; |
225 |
|
|
226 |
|
default: |
227 |
|
return 0; |
228 |
|
} |
229 |
|
} |
230 |
|
|
231 |
|
private Regex regex; |
232 |
|
int n_pieces; |
233 |
|
private int [] pieces; |
234 |
|
string replacement; |
235 |
|
} |
236 |
|
} |