]>
Commit | Line | Data |
---|---|---|
1 | /* | |
2 | * Copyright (C) 2005 Junio C Hamano | |
3 | * Copyright (C) 2010 Google Inc. | |
4 | */ | |
5 | #include "cache.h" | |
6 | #include "diff.h" | |
7 | #include "diffcore.h" | |
8 | #include "xdiff-interface.h" | |
9 | #include "kwset.h" | |
10 | #include "commit.h" | |
11 | #include "quote.h" | |
12 | ||
13 | typedef int (*pickaxe_fn)(mmfile_t *one, mmfile_t *two, | |
14 | struct diff_options *o, | |
15 | regex_t *regexp, kwset_t kws); | |
16 | ||
17 | struct diffgrep_cb { | |
18 | regex_t *regexp; | |
19 | int hit; | |
20 | }; | |
21 | ||
22 | static void diffgrep_consume(void *priv, char *line, unsigned long len) | |
23 | { | |
24 | struct diffgrep_cb *data = priv; | |
25 | regmatch_t regmatch; | |
26 | ||
27 | if (line[0] != '+' && line[0] != '-') | |
28 | return; | |
29 | if (data->hit) | |
30 | /* | |
31 | * NEEDSWORK: we should have a way to terminate the | |
32 | * caller early. | |
33 | */ | |
34 | return; | |
35 | data->hit = !regexec_buf(data->regexp, line + 1, len - 1, 1, | |
36 | ®match, 0); | |
37 | } | |
38 | ||
39 | static int diff_grep(mmfile_t *one, mmfile_t *two, | |
40 | struct diff_options *o, | |
41 | regex_t *regexp, kwset_t kws) | |
42 | { | |
43 | regmatch_t regmatch; | |
44 | struct diffgrep_cb ecbdata; | |
45 | xpparam_t xpp; | |
46 | xdemitconf_t xecfg; | |
47 | ||
48 | if (!one) | |
49 | return !regexec_buf(regexp, two->ptr, two->size, | |
50 | 1, ®match, 0); | |
51 | if (!two) | |
52 | return !regexec_buf(regexp, one->ptr, one->size, | |
53 | 1, ®match, 0); | |
54 | ||
55 | /* | |
56 | * We have both sides; need to run textual diff and see if | |
57 | * the pattern appears on added/deleted lines. | |
58 | */ | |
59 | memset(&xpp, 0, sizeof(xpp)); | |
60 | memset(&xecfg, 0, sizeof(xecfg)); | |
61 | ecbdata.regexp = regexp; | |
62 | ecbdata.hit = 0; | |
63 | xecfg.ctxlen = o->context; | |
64 | xecfg.interhunkctxlen = o->interhunkcontext; | |
65 | if (xdi_diff_outf(one, two, discard_hunk_line, diffgrep_consume, | |
66 | &ecbdata, &xpp, &xecfg)) | |
67 | return 0; | |
68 | return ecbdata.hit; | |
69 | } | |
70 | ||
71 | static unsigned int contains(mmfile_t *mf, regex_t *regexp, kwset_t kws) | |
72 | { | |
73 | unsigned int cnt; | |
74 | unsigned long sz; | |
75 | const char *data; | |
76 | ||
77 | sz = mf->size; | |
78 | data = mf->ptr; | |
79 | cnt = 0; | |
80 | ||
81 | if (regexp) { | |
82 | regmatch_t regmatch; | |
83 | int flags = 0; | |
84 | ||
85 | while (sz && *data && | |
86 | !regexec_buf(regexp, data, sz, 1, ®match, flags)) { | |
87 | flags |= REG_NOTBOL; | |
88 | data += regmatch.rm_eo; | |
89 | sz -= regmatch.rm_eo; | |
90 | if (sz && *data && regmatch.rm_so == regmatch.rm_eo) { | |
91 | data++; | |
92 | sz--; | |
93 | } | |
94 | cnt++; | |
95 | } | |
96 | ||
97 | } else { /* Classic exact string match */ | |
98 | while (sz) { | |
99 | struct kwsmatch kwsm; | |
100 | size_t offset = kwsexec(kws, data, sz, &kwsm); | |
101 | if (offset == -1) | |
102 | break; | |
103 | sz -= offset + kwsm.size[0]; | |
104 | data += offset + kwsm.size[0]; | |
105 | cnt++; | |
106 | } | |
107 | } | |
108 | return cnt; | |
109 | } | |
110 | ||
111 | static int has_changes(mmfile_t *one, mmfile_t *two, | |
112 | struct diff_options *o, | |
113 | regex_t *regexp, kwset_t kws) | |
114 | { | |
115 | unsigned int one_contains = one ? contains(one, regexp, kws) : 0; | |
116 | unsigned int two_contains = two ? contains(two, regexp, kws) : 0; | |
117 | return one_contains != two_contains; | |
118 | } | |
119 | ||
120 | static int pickaxe_match(struct diff_filepair *p, struct diff_options *o, | |
121 | regex_t *regexp, kwset_t kws, pickaxe_fn fn) | |
122 | { | |
123 | struct userdiff_driver *textconv_one = NULL; | |
124 | struct userdiff_driver *textconv_two = NULL; | |
125 | mmfile_t mf1, mf2; | |
126 | int ret; | |
127 | ||
128 | /* ignore unmerged */ | |
129 | if (!DIFF_FILE_VALID(p->one) && !DIFF_FILE_VALID(p->two)) | |
130 | return 0; | |
131 | ||
132 | if (o->objfind) { | |
133 | return (DIFF_FILE_VALID(p->one) && | |
134 | oidset_contains(o->objfind, &p->one->oid)) || | |
135 | (DIFF_FILE_VALID(p->two) && | |
136 | oidset_contains(o->objfind, &p->two->oid)); | |
137 | } | |
138 | ||
139 | if (!o->pickaxe[0]) | |
140 | return 0; | |
141 | ||
142 | if (o->flags.allow_textconv) { | |
143 | textconv_one = get_textconv(o->repo, p->one); | |
144 | textconv_two = get_textconv(o->repo, p->two); | |
145 | } | |
146 | ||
147 | /* | |
148 | * If we have an unmodified pair, we know that the count will be the | |
149 | * same and don't even have to load the blobs. Unless textconv is in | |
150 | * play, _and_ we are using two different textconv filters (e.g., | |
151 | * because a pair is an exact rename with different textconv attributes | |
152 | * for each side, which might generate different content). | |
153 | */ | |
154 | if (textconv_one == textconv_two && diff_unmodified_pair(p)) | |
155 | return 0; | |
156 | ||
157 | if ((o->pickaxe_opts & DIFF_PICKAXE_KIND_G) && | |
158 | !o->flags.text && | |
159 | ((!textconv_one && diff_filespec_is_binary(o->repo, p->one)) || | |
160 | (!textconv_two && diff_filespec_is_binary(o->repo, p->two)))) | |
161 | return 0; | |
162 | ||
163 | mf1.size = fill_textconv(o->repo, textconv_one, p->one, &mf1.ptr); | |
164 | mf2.size = fill_textconv(o->repo, textconv_two, p->two, &mf2.ptr); | |
165 | ||
166 | ret = fn(DIFF_FILE_VALID(p->one) ? &mf1 : NULL, | |
167 | DIFF_FILE_VALID(p->two) ? &mf2 : NULL, | |
168 | o, regexp, kws); | |
169 | ||
170 | if (textconv_one) | |
171 | free(mf1.ptr); | |
172 | if (textconv_two) | |
173 | free(mf2.ptr); | |
174 | diff_free_filespec_data(p->one); | |
175 | diff_free_filespec_data(p->two); | |
176 | ||
177 | return ret; | |
178 | } | |
179 | ||
180 | static void pickaxe(struct diff_queue_struct *q, struct diff_options *o, | |
181 | regex_t *regexp, kwset_t kws, pickaxe_fn fn) | |
182 | { | |
183 | int i; | |
184 | struct diff_queue_struct outq; | |
185 | ||
186 | DIFF_QUEUE_CLEAR(&outq); | |
187 | ||
188 | if (o->pickaxe_opts & DIFF_PICKAXE_ALL) { | |
189 | /* Showing the whole changeset if needle exists */ | |
190 | for (i = 0; i < q->nr; i++) { | |
191 | struct diff_filepair *p = q->queue[i]; | |
192 | if (pickaxe_match(p, o, regexp, kws, fn)) | |
193 | return; /* do not munge the queue */ | |
194 | } | |
195 | ||
196 | /* | |
197 | * Otherwise we will clear the whole queue by copying | |
198 | * the empty outq at the end of this function, but | |
199 | * first clear the current entries in the queue. | |
200 | */ | |
201 | for (i = 0; i < q->nr; i++) | |
202 | diff_free_filepair(q->queue[i]); | |
203 | } else { | |
204 | /* Showing only the filepairs that has the needle */ | |
205 | for (i = 0; i < q->nr; i++) { | |
206 | struct diff_filepair *p = q->queue[i]; | |
207 | if (pickaxe_match(p, o, regexp, kws, fn)) | |
208 | diff_q(&outq, p); | |
209 | else | |
210 | diff_free_filepair(p); | |
211 | } | |
212 | } | |
213 | ||
214 | free(q->queue); | |
215 | *q = outq; | |
216 | } | |
217 | ||
218 | static void regcomp_or_die(regex_t *regex, const char *needle, int cflags) | |
219 | { | |
220 | int err = regcomp(regex, needle, cflags); | |
221 | if (err) { | |
222 | /* The POSIX.2 people are surely sick */ | |
223 | char errbuf[1024]; | |
224 | regerror(err, regex, errbuf, 1024); | |
225 | die("invalid regex: %s", errbuf); | |
226 | } | |
227 | } | |
228 | ||
229 | void diffcore_pickaxe(struct diff_options *o) | |
230 | { | |
231 | const char *needle = o->pickaxe; | |
232 | int opts = o->pickaxe_opts; | |
233 | regex_t regex, *regexp = NULL; | |
234 | kwset_t kws = NULL; | |
235 | ||
236 | if (opts & (DIFF_PICKAXE_REGEX | DIFF_PICKAXE_KIND_G)) { | |
237 | int cflags = REG_EXTENDED | REG_NEWLINE; | |
238 | if (o->pickaxe_opts & DIFF_PICKAXE_IGNORE_CASE) | |
239 | cflags |= REG_ICASE; | |
240 | regcomp_or_die(®ex, needle, cflags); | |
241 | regexp = ®ex; | |
242 | } else if (opts & DIFF_PICKAXE_KIND_S) { | |
243 | if (o->pickaxe_opts & DIFF_PICKAXE_IGNORE_CASE && | |
244 | has_non_ascii(needle)) { | |
245 | struct strbuf sb = STRBUF_INIT; | |
246 | int cflags = REG_NEWLINE | REG_ICASE; | |
247 | ||
248 | basic_regex_quote_buf(&sb, needle); | |
249 | regcomp_or_die(®ex, sb.buf, cflags); | |
250 | strbuf_release(&sb); | |
251 | regexp = ®ex; | |
252 | } else { | |
253 | kws = kwsalloc(o->pickaxe_opts & DIFF_PICKAXE_IGNORE_CASE | |
254 | ? tolower_trans_tbl : NULL); | |
255 | kwsincr(kws, needle, strlen(needle)); | |
256 | kwsprep(kws); | |
257 | } | |
258 | } | |
259 | ||
260 | pickaxe(&diff_queued_diff, o, regexp, kws, | |
261 | (opts & DIFF_PICKAXE_KIND_G) ? diff_grep : has_changes); | |
262 | ||
263 | if (regexp) | |
264 | regfree(regexp); | |
265 | if (kws) | |
266 | kwsfree(kws); | |
267 | return; | |
268 | } |