Commit | Line | Data |
---|---|---|
460b9539 | 1 | /* |
2 | * This file is part of DisOrder | |
5aff007d | 3 | * Copyright (C) 2004, 2005, 2007, 20008 Richard Kettlewell |
460b9539 | 4 | * |
e7eb3a27 | 5 | * This program is free software: you can redistribute it and/or modify |
460b9539 | 6 | * it under the terms of the GNU General Public License as published by |
e7eb3a27 | 7 | * the Free Software Foundation, either version 3 of the License, or |
460b9539 | 8 | * (at your option) any later version. |
e7eb3a27 RK |
9 | * |
10 | * This program is distributed in the hope that it will be useful, | |
11 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
13 | * GNU General Public License for more details. | |
14 | * | |
460b9539 | 15 | * You should have received a copy of the GNU General Public License |
e7eb3a27 | 16 | * along with this program. If not, see <http://www.gnu.org/licenses/>. |
460b9539 | 17 | */ |
132a5a4a RK |
18 | /** @file lib/regsub.c |
19 | * @brief Regexp substitution | |
20 | */ | |
05b75f8d | 21 | #include "common.h" |
460b9539 | 22 | |
a2e9d147 | 23 | #include "regexp.h" |
460b9539 | 24 | #include "regsub.h" |
25 | #include "mem.h" | |
26 | #include "vector.h" | |
27 | #include "log.h" | |
28 | ||
29 | #define PREMATCH (-1) /* fictitious pre-match substring */ | |
30 | #define POSTMATCH (-2) /* fictitious post-match substring */ | |
31 | ||
a2e9d147 MW |
32 | static inline size_t substring_start(const char attribute((unused)) *subject, |
33 | const size_t *ovector, | |
34 | size_t n) { | |
460b9539 | 35 | switch(n) { |
36 | case PREMATCH: return 0; | |
37 | case POSTMATCH: return ovector[1]; | |
38 | default: return ovector[2 * n]; | |
39 | } | |
40 | } | |
41 | ||
a2e9d147 MW |
42 | static inline size_t substring_end(const char *subject, |
43 | const size_t *ovector, | |
44 | size_t n) { | |
460b9539 | 45 | switch(n) { |
46 | case PREMATCH: return ovector[0]; | |
47 | case POSTMATCH: return strlen(subject); | |
48 | default: return ovector[2 * n + 1]; | |
49 | } | |
50 | } | |
51 | ||
52 | static void transform_append(struct dynstr *d, | |
53 | const char *subject, | |
a2e9d147 MW |
54 | const size_t *ovector, |
55 | size_t n) { | |
460b9539 | 56 | int start = substring_start(subject, ovector, n); |
57 | int end = substring_end(subject, ovector, n); | |
58 | ||
59 | if(start != -1) | |
60 | dynstr_append_bytes(d, subject + start, end - start); | |
61 | } | |
62 | ||
63 | static void replace_core(struct dynstr *d, | |
64 | const char *subject, | |
65 | const char *replace, | |
a2e9d147 MW |
66 | size_t rc, |
67 | const size_t *ovector) { | |
68 | size_t substr; | |
460b9539 | 69 | |
70 | while(*replace) { | |
71 | if(*replace == '$') | |
72 | switch(replace[1]) { | |
73 | case '&': | |
74 | transform_append(d, subject, ovector, 0); | |
75 | replace += 2; | |
76 | break; | |
77 | case '1': case '2': case '3': | |
78 | case '4': case '5': case '6': | |
79 | case '7': case '8': case '9': | |
80 | substr = replace[1] - '0'; | |
81 | if(substr < rc) | |
82 | transform_append(d, subject, ovector, substr); | |
83 | replace += 2; | |
84 | break; | |
85 | case '$': | |
86 | dynstr_append(d, '$'); | |
87 | replace += 2; | |
88 | break; | |
89 | default: | |
90 | dynstr_append(d, *replace++); | |
91 | break; | |
92 | } | |
93 | else | |
94 | dynstr_append(d, *replace++); | |
95 | } | |
96 | } | |
97 | ||
98 | unsigned regsub_flags(const char *flags) { | |
99 | unsigned f = 0; | |
100 | ||
101 | while(*flags) { | |
102 | switch(*flags++) { | |
103 | case 'g': f |= REGSUB_GLOBAL; break; | |
104 | case 'i': f |= REGSUB_CASE_INDEPENDENT; break; | |
105 | default: break; | |
106 | } | |
107 | } | |
108 | return f; | |
109 | } | |
110 | ||
111 | int regsub_compile_options(unsigned flags) { | |
112 | int options = 0; | |
113 | ||
114 | if(flags & REGSUB_CASE_INDEPENDENT) | |
a2e9d147 | 115 | options |= RXF_CASELESS; |
460b9539 | 116 | return options; |
117 | } | |
118 | ||
a2e9d147 MW |
119 | const char *regsub(const regexp *re, const char *subject, |
120 | const char *replace, unsigned flags) { | |
121 | int rc, matches; | |
122 | size_t ovector[99]; | |
460b9539 | 123 | struct dynstr d; |
124 | ||
125 | dynstr_init(&d); | |
126 | matches = 0; | |
127 | /* find the next match */ | |
a2e9d147 MW |
128 | while((rc = regexp_match(re, subject, strlen(subject), 0, |
129 | ovector, sizeof ovector / sizeof (ovector[0]))) > 0) { | |
460b9539 | 130 | /* text just before the match */ |
131 | if(!(flags & REGSUB_REPLACE)) | |
132 | transform_append(&d, subject, ovector, PREMATCH); | |
133 | /* the replacement text */ | |
134 | replace_core(&d, subject, replace, rc, ovector); | |
135 | ++matches; | |
136 | if(!*subject) /* end of subject */ | |
137 | break; | |
138 | if(flags & REGSUB_REPLACE) /* replace subject entirely */ | |
139 | break; | |
140 | /* step over the matched substring */ | |
141 | subject += substring_start(subject, ovector, POSTMATCH); | |
142 | if(!(flags & REGSUB_GLOBAL)) | |
143 | break; | |
144 | } | |
a2e9d147 MW |
145 | if(rc <= 0 && rc != RXERR_NOMATCH) { |
146 | disorder_error(0, "regexp_match returned %d, subject '%s'", rc, subject); | |
460b9539 | 147 | return 0; |
148 | } | |
149 | if((flags & REGSUB_MUST_MATCH) && matches == 0) | |
150 | return 0; | |
151 | /* append the remainder of the subject */ | |
152 | if(!(flags & REGSUB_REPLACE)) | |
153 | dynstr_append_string(&d, subject); | |
154 | dynstr_terminate(&d); | |
155 | return d.vec; | |
156 | } | |
157 | ||
158 | /* | |
159 | Local Variables: | |
160 | c-basic-offset:2 | |
161 | comment-column:40 | |
162 | End: | |
163 | */ |