]>
Commit | Line | Data |
---|---|---|
25ed3412 BY |
1 | #include "git-compat-util.h" |
2 | #include "line-range.h" | |
13b8f68c | 3 | #include "xdiff-interface.h" |
13b8f68c | 4 | #include "userdiff.h" |
25ed3412 BY |
5 | |
6 | /* | |
7 | * Parse one item in the -L option | |
815834e9 ES |
8 | * |
9 | * 'begin' is applicable only to relative range anchors. Absolute anchors | |
10 | * ignore this value. | |
11 | * | |
12 | * When parsing "-L A,B", parse_loc() is called once for A and once for B. | |
13 | * | |
14 | * When parsing A, 'begin' must be a negative number, the absolute value of | |
15 | * which is the line at which relative start-of-range anchors should be | |
16 | * based. Beginning of file is represented by -1. | |
17 | * | |
18 | * When parsing B, 'begin' must be the positive line number immediately | |
19 | * following the line computed for 'A'. | |
25ed3412 BY |
20 | */ |
21 | static const char *parse_loc(const char *spec, nth_line_fn_t nth_line, | |
22 | void *data, long lines, long begin, long *ret) | |
23 | { | |
24 | char *term; | |
25 | const char *line; | |
26 | long num; | |
27 | int reg_error; | |
28 | regex_t regexp; | |
29 | regmatch_t match[1]; | |
30 | ||
31 | /* Allow "-L <something>,+20" to mean starting at <something> | |
32 | * for 20 lines, or "-L <something>,-5" for 5 lines ending at | |
33 | * <something>. | |
34 | */ | |
5d57cac6 | 35 | if (1 <= begin && (spec[0] == '+' || spec[0] == '-')) { |
25ed3412 BY |
36 | num = strtol(spec + 1, &term, 10); |
37 | if (term != spec + 1) { | |
12da1d1f TR |
38 | if (!ret) |
39 | return term; | |
abba3539 ES |
40 | if (num == 0) |
41 | die("-L invalid empty range"); | |
25ed3412 BY |
42 | if (spec[0] == '-') |
43 | num = 0 - num; | |
44 | if (0 < num) | |
45 | *ret = begin + num - 2; | |
46 | else if (!num) | |
47 | *ret = begin; | |
48 | else | |
96cfa94e | 49 | *ret = begin + num > 0 ? begin + num : 1; |
25ed3412 BY |
50 | return term; |
51 | } | |
52 | return spec; | |
53 | } | |
54 | num = strtol(spec, &term, 10); | |
55 | if (term != spec) { | |
5ce922a0 ES |
56 | if (ret) { |
57 | if (num <= 0) | |
58 | die("-L invalid line number: %ld", num); | |
12da1d1f | 59 | *ret = num; |
5ce922a0 | 60 | } |
25ed3412 BY |
61 | return term; |
62 | } | |
815834e9 | 63 | |
a6ac5f98 ES |
64 | if (begin < 0) { |
65 | if (spec[0] != '^') | |
66 | begin = -begin; | |
67 | else { | |
68 | begin = 1; | |
69 | spec++; | |
70 | } | |
71 | } | |
815834e9 | 72 | |
25ed3412 BY |
73 | if (spec[0] != '/') |
74 | return spec; | |
75 | ||
76 | /* it could be a regexp of form /.../ */ | |
77 | for (term = (char *) spec + 1; *term && *term != '/'; term++) { | |
78 | if (*term == '\\') | |
79 | term++; | |
80 | } | |
81 | if (*term != '/') | |
82 | return spec; | |
83 | ||
12da1d1f TR |
84 | /* in the scan-only case we are not interested in the regex */ |
85 | if (!ret) | |
86 | return term+1; | |
87 | ||
25ed3412 BY |
88 | /* try [spec+1 .. term-1] as regexp */ |
89 | *term = 0; | |
90 | begin--; /* input is in human terms */ | |
91 | line = nth_line(data, begin); | |
92 | ||
93 | if (!(reg_error = regcomp(®exp, spec + 1, REG_NEWLINE)) && | |
94 | !(reg_error = regexec(®exp, line, 1, match, 0))) { | |
95 | const char *cp = line + match[0].rm_so; | |
96 | const char *nline; | |
97 | ||
98 | while (begin++ < lines) { | |
99 | nline = nth_line(data, begin); | |
100 | if (line <= cp && cp < nline) | |
101 | break; | |
102 | line = nline; | |
103 | } | |
104 | *ret = begin; | |
105 | regfree(®exp); | |
106 | *term++ = '/'; | |
107 | return term; | |
108 | } | |
109 | else { | |
110 | char errbuf[1024]; | |
111 | regerror(reg_error, ®exp, errbuf, 1024); | |
815834e9 ES |
112 | die("-L parameter '%s' starting at line %ld: %s", |
113 | spec + 1, begin + 1, errbuf); | |
25ed3412 BY |
114 | } |
115 | } | |
116 | ||
13b8f68c TR |
117 | static int match_funcname(xdemitconf_t *xecfg, const char *bol, const char *eol) |
118 | { | |
119 | if (xecfg) { | |
120 | char buf[1]; | |
121 | return xecfg->find_func(bol, eol - bol, buf, 1, | |
122 | xecfg->find_func_priv) >= 0; | |
123 | } | |
124 | ||
125 | if (bol == eol) | |
126 | return 0; | |
127 | if (isalpha(*bol) || *bol == '_' || *bol == '$') | |
128 | return 1; | |
129 | return 0; | |
130 | } | |
131 | ||
132 | static const char *find_funcname_matching_regexp(xdemitconf_t *xecfg, const char *start, | |
133 | regex_t *regexp) | |
134 | { | |
135 | int reg_error; | |
136 | regmatch_t match[1]; | |
4e57c88e | 137 | while (*start) { |
13b8f68c TR |
138 | const char *bol, *eol; |
139 | reg_error = regexec(regexp, start, 1, match, 0); | |
140 | if (reg_error == REG_NOMATCH) | |
141 | return NULL; | |
142 | else if (reg_error) { | |
143 | char errbuf[1024]; | |
144 | regerror(reg_error, regexp, errbuf, 1024); | |
145 | die("-L parameter: regexec() failed: %s", errbuf); | |
146 | } | |
147 | /* determine extent of line matched */ | |
148 | bol = start+match[0].rm_so; | |
149 | eol = start+match[0].rm_eo; | |
4e57c88e LKS |
150 | while (bol > start && *--bol != '\n') |
151 | ; /* nothing */ | |
13b8f68c TR |
152 | if (*bol == '\n') |
153 | bol++; | |
154 | while (*eol && *eol != '\n') | |
155 | eol++; | |
156 | if (*eol == '\n') | |
157 | eol++; | |
158 | /* is it a funcname line? */ | |
159 | if (match_funcname(xecfg, (char*) bol, (char*) eol)) | |
160 | return bol; | |
161 | start = eol; | |
162 | } | |
4e57c88e | 163 | return NULL; |
13b8f68c TR |
164 | } |
165 | ||
80e03855 NTND |
166 | static const char *parse_range_funcname( |
167 | const char *arg, nth_line_fn_t nth_line_cb, | |
168 | void *cb_data, long lines, long anchor, long *begin, long *end, | |
169 | const char *path, struct index_state *istate) | |
13b8f68c TR |
170 | { |
171 | char *pattern; | |
172 | const char *term; | |
173 | struct userdiff_driver *drv; | |
174 | xdemitconf_t *xecfg = NULL; | |
175 | const char *start; | |
176 | const char *p; | |
177 | int reg_error; | |
178 | regex_t regexp; | |
179 | ||
215e76c7 ES |
180 | if (*arg == '^') { |
181 | anchor = 1; | |
182 | arg++; | |
183 | } | |
184 | ||
13b8f68c TR |
185 | assert(*arg == ':'); |
186 | term = arg+1; | |
187 | while (*term && *term != ':') { | |
188 | if (*term == '\\' && *(term+1)) | |
189 | term++; | |
190 | term++; | |
191 | } | |
192 | if (term == arg+1) | |
193 | return NULL; | |
194 | if (!begin) /* skip_range_arg case */ | |
195 | return term; | |
196 | ||
197 | pattern = xstrndup(arg+1, term-(arg+1)); | |
198 | ||
1ce761a5 ES |
199 | anchor--; /* input is in human terms */ |
200 | start = nth_line_cb(cb_data, anchor); | |
13b8f68c | 201 | |
80e03855 | 202 | drv = userdiff_find_by_path(istate, path); |
13b8f68c TR |
203 | if (drv && drv->funcname.pattern) { |
204 | const struct userdiff_funcname *pe = &drv->funcname; | |
ca56dadb | 205 | CALLOC_ARRAY(xecfg, 1); |
13b8f68c TR |
206 | xdiff_set_find_func(xecfg, pe->pattern, pe->cflags); |
207 | } | |
208 | ||
209 | reg_error = regcomp(®exp, pattern, REG_NEWLINE); | |
210 | if (reg_error) { | |
211 | char errbuf[1024]; | |
212 | regerror(reg_error, ®exp, errbuf, 1024); | |
213 | die("-L parameter '%s': %s", pattern, errbuf); | |
214 | } | |
215 | ||
216 | p = find_funcname_matching_regexp(xecfg, (char*) start, ®exp); | |
217 | if (!p) | |
1ce761a5 ES |
218 | die("-L parameter '%s' starting at line %ld: no match", |
219 | pattern, anchor + 1); | |
13b8f68c TR |
220 | *begin = 0; |
221 | while (p > nth_line_cb(cb_data, *begin)) | |
222 | (*begin)++; | |
223 | ||
224 | if (*begin >= lines) | |
225 | die("-L parameter '%s' matches at EOF", pattern); | |
226 | ||
227 | *end = *begin+1; | |
228 | while (*end < lines) { | |
229 | const char *bol = nth_line_cb(cb_data, *end); | |
230 | const char *eol = nth_line_cb(cb_data, *end+1); | |
231 | if (match_funcname(xecfg, bol, eol)) | |
232 | break; | |
233 | (*end)++; | |
234 | } | |
235 | ||
236 | regfree(®exp); | |
237 | free(xecfg); | |
238 | free(pattern); | |
239 | ||
240 | /* compensate for 1-based numbering */ | |
241 | (*begin)++; | |
242 | ||
243 | return term; | |
244 | } | |
245 | ||
25ed3412 | 246 | int parse_range_arg(const char *arg, nth_line_fn_t nth_line_cb, |
815834e9 | 247 | void *cb_data, long lines, long anchor, |
80e03855 NTND |
248 | long *begin, long *end, |
249 | const char *path, struct index_state *istate) | |
25ed3412 | 250 | { |
3bf65f9e ES |
251 | *begin = *end = 0; |
252 | ||
815834e9 ES |
253 | if (anchor < 1) |
254 | anchor = 1; | |
255 | if (anchor > lines) | |
256 | anchor = lines + 1; | |
257 | ||
215e76c7 | 258 | if (*arg == ':' || (*arg == '^' && *(arg + 1) == ':')) { |
80e03855 NTND |
259 | arg = parse_range_funcname(arg, nth_line_cb, cb_data, |
260 | lines, anchor, begin, end, | |
261 | path, istate); | |
13b8f68c TR |
262 | if (!arg || *arg) |
263 | return -1; | |
264 | return 0; | |
265 | } | |
266 | ||
815834e9 | 267 | arg = parse_loc(arg, nth_line_cb, cb_data, lines, -anchor, begin); |
25ed3412 BY |
268 | |
269 | if (*arg == ',') | |
270 | arg = parse_loc(arg + 1, nth_line_cb, cb_data, lines, *begin + 1, end); | |
271 | ||
272 | if (*arg) | |
273 | return -1; | |
274 | ||
3bf65f9e | 275 | if (*begin && *end && *end < *begin) { |
35d803bc | 276 | SWAP(*end, *begin); |
3bf65f9e ES |
277 | } |
278 | ||
25ed3412 BY |
279 | return 0; |
280 | } | |
12da1d1f | 281 | |
80e03855 | 282 | const char *skip_range_arg(const char *arg, struct index_state *istate) |
12da1d1f | 283 | { |
215e76c7 | 284 | if (*arg == ':' || (*arg == '^' && *(arg + 1) == ':')) |
80e03855 NTND |
285 | return parse_range_funcname(arg, NULL, NULL, |
286 | 0, 0, NULL, NULL, | |
287 | NULL, istate); | |
13b8f68c | 288 | |
12da1d1f TR |
289 | arg = parse_loc(arg, NULL, NULL, 0, -1, NULL); |
290 | ||
291 | if (*arg == ',') | |
292 | arg = parse_loc(arg+1, NULL, NULL, 0, 0, NULL); | |
293 | ||
294 | return arg; | |
295 | } |