]>
Commit | Line | Data |
---|---|---|
2744b234 LT |
1 | /* |
2 | * Another stupid program, this one parsing the headers of an | |
3 | * email to figure out authorship and subject | |
4 | */ | |
f1f909e3 | 5 | #include "cache.h" |
34488e3c | 6 | #include "builtin.h" |
b45974a6 | 7 | #include "utf8.h" |
3b6121f6 | 8 | #include "strbuf.h" |
2744b234 | 9 | |
34488e3c | 10 | static FILE *cmitmsg, *patchfile, *fin, *fout; |
2744b234 | 11 | |
96f1e58f DR |
12 | static int keep_subject; |
13 | static const char *metainfo_charset; | |
3b6121f6 LS |
14 | static struct strbuf line = STRBUF_INIT; |
15 | static struct strbuf name = STRBUF_INIT; | |
16 | static struct strbuf email = STRBUF_INIT; | |
2744b234 | 17 | |
d4a9ce78 JH |
18 | static enum { |
19 | TE_DONTCARE, TE_QP, TE_BASE64, | |
20 | } transfer_encoding; | |
87ab7992 DZ |
21 | static enum { |
22 | TYPE_TEXT, TYPE_OTHER, | |
23 | } message_type; | |
d4a9ce78 | 24 | |
3b6121f6 | 25 | static struct strbuf charset = STRBUF_INIT; |
96f1e58f | 26 | static int patch_lines; |
3b6121f6 | 27 | static struct strbuf **p_hdr_data, **s_hdr_data; |
87ab7992 DZ |
28 | |
29 | #define MAX_HDR_PARSED 10 | |
30 | #define MAX_BOUNDARIES 5 | |
d4a9ce78 | 31 | |
3b6121f6 | 32 | static void get_sane_name(struct strbuf *out, struct strbuf *name, struct strbuf *email) |
2744b234 | 33 | { |
3b6121f6 LS |
34 | struct strbuf *src = name; |
35 | if (name->len < 3 || 60 < name->len || strchr(name->buf, '@') || | |
36 | strchr(name->buf, '<') || strchr(name->buf, '>')) | |
37 | src = email; | |
38 | else if (name == out) | |
39 | return; | |
40 | strbuf_reset(out); | |
41 | strbuf_addbuf(out, src); | |
2744b234 LT |
42 | } |
43 | ||
3b6121f6 | 44 | static void parse_bogus_from(const struct strbuf *line) |
e0e3ba20 JH |
45 | { |
46 | /* John Doe <johndoe> */ | |
e0e3ba20 | 47 | |
3b6121f6 | 48 | char *bra, *ket; |
e0e3ba20 JH |
49 | /* This is fallback, so do not bother if we already have an |
50 | * e-mail address. | |
34488e3c | 51 | */ |
3b6121f6 LS |
52 | if (email.len) |
53 | return; | |
e0e3ba20 | 54 | |
3b6121f6 | 55 | bra = strchr(line->buf, '<'); |
e0e3ba20 | 56 | if (!bra) |
3b6121f6 | 57 | return; |
e0e3ba20 JH |
58 | ket = strchr(bra, '>'); |
59 | if (!ket) | |
3b6121f6 | 60 | return; |
e0e3ba20 | 61 | |
3b6121f6 LS |
62 | strbuf_reset(&email); |
63 | strbuf_add(&email, bra + 1, ket - bra - 1); | |
64 | ||
65 | strbuf_reset(&name); | |
66 | strbuf_add(&name, line->buf, bra - line->buf); | |
67 | strbuf_trim(&name); | |
68 | get_sane_name(&name, &name, &email); | |
e0e3ba20 JH |
69 | } |
70 | ||
3b6121f6 | 71 | static void handle_from(const struct strbuf *from) |
2744b234 | 72 | { |
2dec02b1 | 73 | char *at; |
3b6121f6 LS |
74 | size_t el; |
75 | struct strbuf f; | |
2744b234 | 76 | |
3b6121f6 LS |
77 | strbuf_init(&f, from->len); |
78 | strbuf_addbuf(&f, from); | |
79 | ||
80 | at = strchr(f.buf, '@'); | |
81 | if (!at) { | |
82 | parse_bogus_from(from); | |
83 | return; | |
84 | } | |
2744b234 LT |
85 | |
86 | /* | |
87 | * If we already have one email, don't take any confusing lines | |
88 | */ | |
3b6121f6 LS |
89 | if (email.len && strchr(at + 1, '@')) { |
90 | strbuf_release(&f); | |
91 | return; | |
92 | } | |
2744b234 | 93 | |
d4a9ce78 | 94 | /* Pick up the string around '@', possibly delimited with <> |
3b6121f6 | 95 | * pair; that is the email part. |
d4a9ce78 | 96 | */ |
3b6121f6 | 97 | while (at > f.buf) { |
2744b234 | 98 | char c = at[-1]; |
d4a9ce78 JH |
99 | if (isspace(c)) |
100 | break; | |
101 | if (c == '<') { | |
102 | at[-1] = ' '; | |
2744b234 | 103 | break; |
d4a9ce78 | 104 | } |
2744b234 LT |
105 | at--; |
106 | } | |
3b6121f6 LS |
107 | el = strcspn(at, " \n\t\r\v\f>"); |
108 | strbuf_reset(&email); | |
109 | strbuf_add(&email, at, el); | |
110 | strbuf_remove(&f, at - f.buf, el + 1); | |
2744b234 | 111 | |
d4a9ce78 | 112 | /* The remainder is name. It could be "John Doe <john.doe@xz>" |
3b6121f6 | 113 | * or "john.doe@xz (John Doe)", but we have removed the |
d4a9ce78 JH |
114 | * email part, so trim from both ends, possibly removing |
115 | * the () pair at the end. | |
116 | */ | |
3b6121f6 | 117 | strbuf_trim(&f); |
0d4ede9f PB |
118 | if (f.buf[0] == '(' && f.len && f.buf[f.len - 1] == ')') { |
119 | strbuf_remove(&f, 0, 1); | |
3b6121f6 | 120 | strbuf_setlen(&f, f.len - 1); |
0d4ede9f | 121 | } |
2744b234 | 122 | |
3b6121f6 LS |
123 | get_sane_name(&name, &f, &email); |
124 | strbuf_release(&f); | |
2744b234 LT |
125 | } |
126 | ||
3b6121f6 | 127 | static void handle_header(struct strbuf **out, const struct strbuf *line) |
62c1f6b4 | 128 | { |
3b6121f6 LS |
129 | if (!*out) { |
130 | *out = xmalloc(sizeof(struct strbuf)); | |
131 | strbuf_init(*out, line->len); | |
132 | } else | |
133 | strbuf_reset(*out); | |
87ab7992 | 134 | |
3b6121f6 | 135 | strbuf_addbuf(*out, line); |
d4a9ce78 JH |
136 | } |
137 | ||
138 | /* NOTE NOTE NOTE. We do not claim we do full MIME. We just attempt | |
139 | * to have enough heuristics to grok MIME encoded patches often found | |
140 | * on our mailing lists. For example, we do not even treat header lines | |
141 | * case insensitively. | |
142 | */ | |
143 | ||
3b6121f6 | 144 | static int slurp_attr(const char *line, const char *name, struct strbuf *attr) |
d4a9ce78 | 145 | { |
554fe20d | 146 | const char *ends, *ap = strcasestr(line, name); |
d4a9ce78 JH |
147 | size_t sz; |
148 | ||
149 | if (!ap) { | |
3b6121f6 | 150 | strbuf_setlen(attr, 0); |
d4a9ce78 JH |
151 | return 0; |
152 | } | |
153 | ap += strlen(name); | |
154 | if (*ap == '"') { | |
155 | ap++; | |
156 | ends = "\""; | |
157 | } | |
158 | else | |
159 | ends = "; \t"; | |
160 | sz = strcspn(ap, ends); | |
3b6121f6 | 161 | strbuf_add(attr, ap, sz); |
d4a9ce78 JH |
162 | return 1; |
163 | } | |
164 | ||
3b6121f6 | 165 | static struct strbuf *content[MAX_BOUNDARIES]; |
87ab7992 | 166 | |
3b6121f6 | 167 | static struct strbuf **content_top = content; |
87ab7992 | 168 | |
3b6121f6 | 169 | static void handle_content_type(struct strbuf *line) |
d4a9ce78 | 170 | { |
3b6121f6 LS |
171 | struct strbuf *boundary = xmalloc(sizeof(struct strbuf)); |
172 | strbuf_init(boundary, line->len); | |
87ab7992 | 173 | |
3b6121f6 | 174 | if (!strcasestr(line->buf, "text/")) |
87ab7992 | 175 | message_type = TYPE_OTHER; |
3b6121f6 LS |
176 | if (slurp_attr(line->buf, "boundary=", boundary)) { |
177 | strbuf_insert(boundary, 0, "--", 2); | |
87ab7992 DZ |
178 | if (content_top++ >= &content[MAX_BOUNDARIES]) { |
179 | fprintf(stderr, "Too many boundaries to handle\n"); | |
180 | exit(1); | |
181 | } | |
3b6121f6 LS |
182 | *content_top = boundary; |
183 | boundary = NULL; | |
b893f091 | 184 | } |
3b6121f6 LS |
185 | if (slurp_attr(line->buf, "charset=", &charset)) |
186 | strbuf_tolower(&charset); | |
187 | ||
188 | if (boundary) { | |
189 | strbuf_release(boundary); | |
190 | free(boundary); | |
d4a9ce78 | 191 | } |
d4a9ce78 JH |
192 | } |
193 | ||
3b6121f6 | 194 | static void handle_content_transfer_encoding(const struct strbuf *line) |
d4a9ce78 | 195 | { |
3b6121f6 | 196 | if (strcasestr(line->buf, "base64")) |
d4a9ce78 | 197 | transfer_encoding = TE_BASE64; |
3b6121f6 | 198 | else if (strcasestr(line->buf, "quoted-printable")) |
d4a9ce78 JH |
199 | transfer_encoding = TE_QP; |
200 | else | |
201 | transfer_encoding = TE_DONTCARE; | |
2744b234 LT |
202 | } |
203 | ||
3b6121f6 | 204 | static int is_multipart_boundary(const struct strbuf *line) |
d4a9ce78 | 205 | { |
3b6121f6 | 206 | return !strbuf_cmp(line, *content_top); |
d4a9ce78 JH |
207 | } |
208 | ||
3b6121f6 | 209 | static void cleanup_subject(struct strbuf *subject) |
2744b234 | 210 | { |
3b6121f6 LS |
211 | char *pos; |
212 | size_t remove; | |
213 | while (subject->len) { | |
214 | switch (*subject->buf) { | |
2744b234 | 215 | case 'r': case 'R': |
3b6121f6 LS |
216 | if (subject->len <= 3) |
217 | break; | |
218 | if (!memcmp(subject->buf + 1, "e:", 2)) { | |
219 | strbuf_remove(subject, 0, 3); | |
2744b234 LT |
220 | continue; |
221 | } | |
222 | break; | |
223 | case ' ': case '\t': case ':': | |
3b6121f6 | 224 | strbuf_remove(subject, 0, 1); |
2744b234 | 225 | continue; |
2744b234 | 226 | case '[': |
3b6121f6 | 227 | if ((pos = strchr(subject->buf, ']'))) { |
1e102bf7 JH |
228 | remove = pos - subject->buf; |
229 | if (remove <= (subject->len - remove) * 2) { | |
230 | strbuf_remove(subject, 0, remove + 1); | |
3b6121f6 LS |
231 | continue; |
232 | } | |
233 | } else | |
234 | strbuf_remove(subject, 0, 1); | |
2744b234 LT |
235 | break; |
236 | } | |
3b6121f6 LS |
237 | strbuf_trim(subject); |
238 | return; | |
2744b234 | 239 | } |
34488e3c | 240 | } |
2744b234 | 241 | |
3b6121f6 | 242 | static void cleanup_space(struct strbuf *sb) |
2744b234 | 243 | { |
3b6121f6 LS |
244 | size_t pos, cnt; |
245 | for (pos = 0; pos < sb->len; pos++) { | |
246 | if (isspace(sb->buf[pos])) { | |
247 | sb->buf[pos] = ' '; | |
248 | for (cnt = 0; isspace(sb->buf[pos + cnt + 1]); cnt++); | |
249 | strbuf_remove(sb, pos + 1, cnt); | |
2744b234 LT |
250 | } |
251 | } | |
252 | } | |
253 | ||
3b6121f6 | 254 | static void decode_header(struct strbuf *line); |
538dfe73 | 255 | static const char *header[MAX_HDR_PARSED] = { |
87ab7992 | 256 | "From","Subject","Date", |
d4a9ce78 JH |
257 | }; |
258 | ||
3b6121f6 | 259 | static inline int cmp_header(const struct strbuf *line, const char *hdr) |
d4a9ce78 | 260 | { |
3b6121f6 LS |
261 | int len = strlen(hdr); |
262 | return !strncasecmp(line->buf, hdr, len) && line->len > len && | |
263 | line->buf[len] == ':' && isspace(line->buf[len + 1]); | |
264 | } | |
d4a9ce78 | 265 | |
3b6121f6 LS |
266 | static int check_header(const struct strbuf *line, |
267 | struct strbuf *hdr_data[], int overwrite) | |
268 | { | |
269 | int i, ret = 0, len; | |
270 | struct strbuf sb = STRBUF_INIT; | |
87ab7992 DZ |
271 | /* search for the interesting parts */ |
272 | for (i = 0; header[i]; i++) { | |
273 | int len = strlen(header[i]); | |
3b6121f6 | 274 | if ((!hdr_data[i] || overwrite) && cmp_header(line, header[i])) { |
33504530 EB |
275 | /* Unwrap inline B and Q encoding, and optionally |
276 | * normalize the meta information to utf8. | |
277 | */ | |
3b6121f6 LS |
278 | strbuf_add(&sb, line->buf + len + 2, line->len - len - 2); |
279 | decode_header(&sb); | |
280 | handle_header(&hdr_data[i], &sb); | |
281 | ret = 1; | |
282 | goto check_header_out; | |
d4a9ce78 JH |
283 | } |
284 | } | |
d4a9ce78 | 285 | |
87ab7992 | 286 | /* Content stuff */ |
3b6121f6 LS |
287 | if (cmp_header(line, "Content-Type")) { |
288 | len = strlen("Content-Type: "); | |
289 | strbuf_add(&sb, line->buf + len, line->len - len); | |
290 | decode_header(&sb); | |
291 | strbuf_insert(&sb, 0, "Content-Type: ", len); | |
292 | handle_content_type(&sb); | |
293 | ret = 1; | |
294 | goto check_header_out; | |
295 | } | |
296 | if (cmp_header(line, "Content-Transfer-Encoding")) { | |
297 | len = strlen("Content-Transfer-Encoding: "); | |
298 | strbuf_add(&sb, line->buf + len, line->len - len); | |
299 | decode_header(&sb); | |
300 | handle_content_transfer_encoding(&sb); | |
301 | ret = 1; | |
302 | goto check_header_out; | |
87ab7992 DZ |
303 | } |
304 | ||
305 | /* for inbody stuff */ | |
3b6121f6 LS |
306 | if (!prefixcmp(line->buf, ">From") && isspace(line->buf[5])) { |
307 | ret = 1; /* Should this return 0? */ | |
308 | goto check_header_out; | |
309 | } | |
310 | if (!prefixcmp(line->buf, "[PATCH]") && isspace(line->buf[7])) { | |
87ab7992 | 311 | for (i = 0; header[i]; i++) { |
e9fe804a | 312 | if (!memcmp("Subject", header[i], 7)) { |
3b6121f6 LS |
313 | handle_header(&hdr_data[i], line); |
314 | ret = 1; | |
315 | goto check_header_out; | |
87ab7992 DZ |
316 | } |
317 | } | |
318 | } | |
319 | ||
3b6121f6 LS |
320 | check_header_out: |
321 | strbuf_release(&sb); | |
322 | return ret; | |
d4a9ce78 JH |
323 | } |
324 | ||
3b6121f6 | 325 | static int is_rfc2822_header(const struct strbuf *line) |
ef29c117 JH |
326 | { |
327 | /* | |
328 | * The section that defines the loosest possible | |
329 | * field name is "3.6.8 Optional fields". | |
330 | * | |
331 | * optional-field = field-name ":" unstructured CRLF | |
332 | * field-name = 1*ftext | |
333 | * ftext = %d33-57 / %59-126 | |
334 | */ | |
335 | int ch; | |
3b6121f6 | 336 | char *cp = line->buf; |
34fc5cef LT |
337 | |
338 | /* Count mbox From headers as headers */ | |
3b6121f6 | 339 | if (!prefixcmp(cp, "From ") || !prefixcmp(cp, ">From ")) |
34fc5cef LT |
340 | return 1; |
341 | ||
ef29c117 JH |
342 | while ((ch = *cp++)) { |
343 | if (ch == ':') | |
3b6121f6 | 344 | return 1; |
ef29c117 JH |
345 | if ((33 <= ch && ch <= 57) || |
346 | (59 <= ch && ch <= 126)) | |
347 | continue; | |
348 | break; | |
349 | } | |
350 | return 0; | |
351 | } | |
352 | ||
3b6121f6 | 353 | static int read_one_header_line(struct strbuf *line, FILE *in) |
d4a9ce78 | 354 | { |
34fc5cef | 355 | /* Get the first part of the line. */ |
3b6121f6 | 356 | if (strbuf_getline(line, in, '\n')) |
34fc5cef LT |
357 | return 0; |
358 | ||
359 | /* | |
360 | * Is it an empty line or not a valid rfc2822 header? | |
361 | * If so, stop here, and return false ("not a header") | |
362 | */ | |
3b6121f6 LS |
363 | strbuf_rtrim(line); |
364 | if (!line->len || !is_rfc2822_header(line)) { | |
34fc5cef | 365 | /* Re-add the newline */ |
3b6121f6 | 366 | strbuf_addch(line, '\n'); |
34fc5cef LT |
367 | return 0; |
368 | } | |
369 | ||
370 | /* | |
371 | * Now we need to eat all the continuation lines.. | |
372 | * Yuck, 2822 header "folding" | |
373 | */ | |
374 | for (;;) { | |
3b6121f6 LS |
375 | int peek; |
376 | struct strbuf continuation = STRBUF_INIT; | |
34fc5cef | 377 | |
f30b2028 EB |
378 | peek = fgetc(in); ungetc(peek, in); |
379 | if (peek != ' ' && peek != '\t') | |
380 | break; | |
3b6121f6 | 381 | if (strbuf_getline(&continuation, in, '\n')) |
34fc5cef | 382 | break; |
3b6121f6 LS |
383 | continuation.buf[0] = '\n'; |
384 | strbuf_rtrim(&continuation); | |
385 | strbuf_addbuf(line, &continuation); | |
d4a9ce78 | 386 | } |
34fc5cef LT |
387 | |
388 | return 1; | |
d4a9ce78 JH |
389 | } |
390 | ||
3b6121f6 | 391 | static struct strbuf *decode_q_segment(const struct strbuf *q_seg, int rfc2047) |
d4a9ce78 | 392 | { |
3b6121f6 | 393 | const char *in = q_seg->buf; |
d4a9ce78 | 394 | int c; |
3b6121f6 LS |
395 | struct strbuf *out = xmalloc(sizeof(struct strbuf)); |
396 | strbuf_init(out, q_seg->len); | |
397 | ||
398 | while ((c = *in++) != 0) { | |
d4a9ce78 JH |
399 | if (c == '=') { |
400 | int d = *in++; | |
401 | if (d == '\n' || !d) | |
402 | break; /* drop trailing newline */ | |
3b6121f6 | 403 | strbuf_addch(out, (hexval(d) << 4) | hexval(*in++)); |
75731930 | 404 | continue; |
d4a9ce78 | 405 | } |
75731930 JH |
406 | if (rfc2047 && c == '_') /* rfc2047 4.2 (2) */ |
407 | c = 0x20; | |
3b6121f6 | 408 | strbuf_addch(out, c); |
d4a9ce78 | 409 | } |
3b6121f6 | 410 | return out; |
d4a9ce78 JH |
411 | } |
412 | ||
3b6121f6 | 413 | static struct strbuf *decode_b_segment(const struct strbuf *b_seg) |
d4a9ce78 JH |
414 | { |
415 | /* Decode in..ep, possibly in-place to ot */ | |
416 | int c, pos = 0, acc = 0; | |
3b6121f6 LS |
417 | const char *in = b_seg->buf; |
418 | struct strbuf *out = xmalloc(sizeof(struct strbuf)); | |
419 | strbuf_init(out, b_seg->len); | |
d4a9ce78 | 420 | |
3b6121f6 | 421 | while ((c = *in++) != 0) { |
d4a9ce78 JH |
422 | if (c == '+') |
423 | c = 62; | |
424 | else if (c == '/') | |
425 | c = 63; | |
426 | else if ('A' <= c && c <= 'Z') | |
427 | c -= 'A'; | |
428 | else if ('a' <= c && c <= 'z') | |
429 | c -= 'a' - 26; | |
430 | else if ('0' <= c && c <= '9') | |
431 | c -= '0' - 52; | |
432 | else if (c == '=') { | |
433 | /* padding is almost like (c == 0), except we do | |
434 | * not output NUL resulting only from it; | |
435 | * for now we just trust the data. | |
436 | */ | |
437 | c = 0; | |
438 | } | |
439 | else | |
440 | continue; /* garbage */ | |
441 | switch (pos++) { | |
442 | case 0: | |
443 | acc = (c << 2); | |
444 | break; | |
445 | case 1: | |
3b6121f6 | 446 | strbuf_addch(out, (acc | (c >> 4))); |
d4a9ce78 JH |
447 | acc = (c & 15) << 4; |
448 | break; | |
449 | case 2: | |
3b6121f6 | 450 | strbuf_addch(out, (acc | (c >> 2))); |
d4a9ce78 JH |
451 | acc = (c & 3) << 6; |
452 | break; | |
453 | case 3: | |
3b6121f6 | 454 | strbuf_addch(out, (acc | c)); |
d4a9ce78 JH |
455 | acc = pos = 0; |
456 | break; | |
457 | } | |
458 | } | |
3b6121f6 | 459 | return out; |
d4a9ce78 JH |
460 | } |
461 | ||
b59d398b LT |
462 | /* |
463 | * When there is no known charset, guess. | |
464 | * | |
465 | * Right now we assume that if the target is UTF-8 (the default), | |
466 | * and it already looks like UTF-8 (which includes US-ASCII as its | |
467 | * subset, of course) then that is what it is and there is nothing | |
468 | * to do. | |
469 | * | |
470 | * Otherwise, we default to assuming it is Latin1 for historical | |
471 | * reasons. | |
472 | */ | |
3b6121f6 | 473 | static const char *guess_charset(const struct strbuf *line, const char *target_charset) |
b59d398b LT |
474 | { |
475 | if (is_encoding_utf8(target_charset)) { | |
3b6121f6 | 476 | if (is_utf8(line->buf)) |
b59d398b LT |
477 | return NULL; |
478 | } | |
479 | return "latin1"; | |
480 | } | |
481 | ||
3b6121f6 | 482 | static void convert_to_utf8(struct strbuf *line, const char *charset) |
d4a9ce78 | 483 | { |
b59d398b LT |
484 | char *out; |
485 | ||
486 | if (!charset || !*charset) { | |
487 | charset = guess_charset(line, metainfo_charset); | |
488 | if (!charset) | |
489 | return; | |
490 | } | |
b45974a6 | 491 | |
7296096c JS |
492 | if (!strcmp(metainfo_charset, charset)) |
493 | return; | |
3b6121f6 | 494 | out = reencode_string(line->buf, metainfo_charset, charset); |
bb1091a4 JH |
495 | if (!out) |
496 | die("cannot convert from %s to %s\n", | |
b59d398b | 497 | charset, metainfo_charset); |
3b6121f6 | 498 | strbuf_attach(line, out, strlen(out), strlen(out)); |
d4a9ce78 JH |
499 | } |
500 | ||
3b6121f6 | 501 | static int decode_header_bq(struct strbuf *it) |
d4a9ce78 | 502 | { |
3b6121f6 LS |
503 | char *in, *ep, *cp; |
504 | struct strbuf outbuf = STRBUF_INIT, *dec; | |
505 | struct strbuf charset_q = STRBUF_INIT, piecebuf = STRBUF_INIT; | |
b75bf2c3 | 506 | int rfc2047 = 0; |
d4a9ce78 | 507 | |
3b6121f6 LS |
508 | in = it->buf; |
509 | while (in - it->buf <= it->len && (ep = strstr(in, "=?")) != NULL) { | |
510 | int encoding; | |
511 | strbuf_reset(&charset_q); | |
512 | strbuf_reset(&piecebuf); | |
b75bf2c3 JH |
513 | rfc2047 = 1; |
514 | ||
d4a9ce78 | 515 | if (in != ep) { |
3b6121f6 LS |
516 | strbuf_add(&outbuf, in, ep - in); |
517 | in = ep; | |
d4a9ce78 JH |
518 | } |
519 | /* E.g. | |
520 | * ep : "=?iso-2022-jp?B?GyR...?= foo" | |
521 | * ep : "=?ISO-8859-1?Q?Foo=FCbar?= baz" | |
522 | */ | |
523 | ep += 2; | |
3b6121f6 LS |
524 | |
525 | if (ep - it->buf >= it->len || !(cp = strchr(ep, '?'))) | |
526 | goto decode_header_bq_out; | |
527 | ||
528 | if (cp + 3 - it->buf > it->len) | |
529 | goto decode_header_bq_out; | |
530 | strbuf_add(&charset_q, ep, cp - ep); | |
531 | strbuf_tolower(&charset_q); | |
532 | ||
d4a9ce78 JH |
533 | encoding = cp[1]; |
534 | if (!encoding || cp[2] != '?') | |
3b6121f6 | 535 | goto decode_header_bq_out; |
d4a9ce78 JH |
536 | ep = strstr(cp + 3, "?="); |
537 | if (!ep) | |
3b6121f6 LS |
538 | goto decode_header_bq_out; |
539 | strbuf_add(&piecebuf, cp + 3, ep - cp - 3); | |
d4a9ce78 JH |
540 | switch (tolower(encoding)) { |
541 | default: | |
3b6121f6 | 542 | goto decode_header_bq_out; |
d4a9ce78 | 543 | case 'b': |
3b6121f6 | 544 | dec = decode_b_segment(&piecebuf); |
d4a9ce78 JH |
545 | break; |
546 | case 'q': | |
3b6121f6 | 547 | dec = decode_q_segment(&piecebuf, 1); |
d4a9ce78 JH |
548 | break; |
549 | } | |
650e4be5 | 550 | if (metainfo_charset) |
3b6121f6 | 551 | convert_to_utf8(dec, charset_q.buf); |
8dabdfcc | 552 | |
3b6121f6 LS |
553 | strbuf_addbuf(&outbuf, dec); |
554 | strbuf_release(dec); | |
555 | free(dec); | |
d4a9ce78 JH |
556 | in = ep + 2; |
557 | } | |
3b6121f6 LS |
558 | strbuf_addstr(&outbuf, in); |
559 | strbuf_reset(it); | |
560 | strbuf_addbuf(it, &outbuf); | |
561 | decode_header_bq_out: | |
562 | strbuf_release(&outbuf); | |
563 | strbuf_release(&charset_q); | |
564 | strbuf_release(&piecebuf); | |
b75bf2c3 JH |
565 | return rfc2047; |
566 | } | |
567 | ||
3b6121f6 | 568 | static void decode_header(struct strbuf *it) |
b75bf2c3 | 569 | { |
3b6121f6 | 570 | if (decode_header_bq(it)) |
b75bf2c3 JH |
571 | return; |
572 | /* otherwise "it" is a straight copy of the input. | |
573 | * This can be binary guck but there is no charset specified. | |
574 | */ | |
575 | if (metainfo_charset) | |
3b6121f6 | 576 | convert_to_utf8(it, ""); |
d4a9ce78 JH |
577 | } |
578 | ||
3b6121f6 | 579 | static void decode_transfer_encoding(struct strbuf *line) |
d4a9ce78 | 580 | { |
3b6121f6 | 581 | struct strbuf *ret; |
d4a9ce78 JH |
582 | |
583 | switch (transfer_encoding) { | |
584 | case TE_QP: | |
3b6121f6 LS |
585 | ret = decode_q_segment(line, 0); |
586 | break; | |
d4a9ce78 | 587 | case TE_BASE64: |
3b6121f6 LS |
588 | ret = decode_b_segment(line); |
589 | break; | |
d4a9ce78 | 590 | case TE_DONTCARE: |
9aa23094 | 591 | default: |
3b6121f6 | 592 | return; |
d4a9ce78 | 593 | } |
3b6121f6 LS |
594 | strbuf_reset(line); |
595 | strbuf_addbuf(line, ret); | |
596 | strbuf_release(ret); | |
597 | free(ret); | |
d4a9ce78 JH |
598 | } |
599 | ||
3b6121f6 | 600 | static void handle_filter(struct strbuf *line); |
87ab7992 DZ |
601 | |
602 | static int find_boundary(void) | |
2744b234 | 603 | { |
3b6121f6 LS |
604 | while (!strbuf_getline(&line, fin, '\n')) { |
605 | if (is_multipart_boundary(&line)) | |
87ab7992 DZ |
606 | return 1; |
607 | } | |
608 | return 0; | |
609 | } | |
610 | ||
611 | static int handle_boundary(void) | |
612 | { | |
3b6121f6 LS |
613 | struct strbuf newline = STRBUF_INIT; |
614 | ||
615 | strbuf_addch(&newline, '\n'); | |
87ab7992 | 616 | again: |
3b6121f6 LS |
617 | if (line.len >= (*content_top)->len + 2 && |
618 | !memcmp(line.buf + (*content_top)->len, "--", 2)) { | |
87ab7992 DZ |
619 | /* we hit an end boundary */ |
620 | /* pop the current boundary off the stack */ | |
3b6121f6 LS |
621 | strbuf_release(*content_top); |
622 | free(*content_top); | |
623 | *content_top = NULL; | |
87ab7992 DZ |
624 | |
625 | /* technically won't happen as is_multipart_boundary() | |
626 | will fail first. But just in case.. | |
627 | */ | |
628 | if (content_top-- < content) { | |
629 | fprintf(stderr, "Detected mismatched boundaries, " | |
630 | "can't recover\n"); | |
631 | exit(1); | |
632 | } | |
3b6121f6 LS |
633 | handle_filter(&newline); |
634 | strbuf_release(&newline); | |
87ab7992 DZ |
635 | |
636 | /* skip to the next boundary */ | |
637 | if (!find_boundary()) | |
638 | return 0; | |
639 | goto again; | |
640 | } | |
641 | ||
642 | /* set some defaults */ | |
643 | transfer_encoding = TE_DONTCARE; | |
3b6121f6 | 644 | strbuf_reset(&charset); |
87ab7992 | 645 | message_type = TYPE_TEXT; |
d4a9ce78 | 646 | |
87ab7992 | 647 | /* slurp in this section's info */ |
3b6121f6 LS |
648 | while (read_one_header_line(&line, fin)) |
649 | check_header(&line, p_hdr_data, 0); | |
2744b234 | 650 | |
3b6121f6 | 651 | strbuf_release(&newline); |
87ab7992 | 652 | /* eat the blank line after section info */ |
3b6121f6 | 653 | return (strbuf_getline(&line, fin, '\n') == 0); |
d4a9ce78 JH |
654 | } |
655 | ||
3b6121f6 | 656 | static inline int patchbreak(const struct strbuf *line) |
f0658cf2 | 657 | { |
3b6121f6 LS |
658 | size_t i; |
659 | ||
f0658cf2 | 660 | /* Beginning of a "diff -" header? */ |
3b6121f6 | 661 | if (!prefixcmp(line->buf, "diff -")) |
f0658cf2 DZ |
662 | return 1; |
663 | ||
664 | /* CVS "Index: " line? */ | |
3b6121f6 | 665 | if (!prefixcmp(line->buf, "Index: ")) |
f0658cf2 DZ |
666 | return 1; |
667 | ||
668 | /* | |
669 | * "--- <filename>" starts patches without headers | |
670 | * "---<sp>*" is a manual separator | |
671 | */ | |
3b6121f6 LS |
672 | if (line->len < 4) |
673 | return 0; | |
674 | ||
675 | if (!prefixcmp(line->buf, "---")) { | |
f0658cf2 | 676 | /* space followed by a filename? */ |
3b6121f6 | 677 | if (line->buf[3] == ' ' && !isspace(line->buf[4])) |
f0658cf2 DZ |
678 | return 1; |
679 | /* Just whitespace? */ | |
3b6121f6 LS |
680 | for (i = 3; i < line->len; i++) { |
681 | unsigned char c = line->buf[i]; | |
f0658cf2 DZ |
682 | if (c == '\n') |
683 | return 1; | |
684 | if (!isspace(c)) | |
685 | break; | |
686 | } | |
687 | return 0; | |
688 | } | |
689 | return 0; | |
690 | } | |
691 | ||
3b6121f6 | 692 | static int handle_commit_msg(struct strbuf *line) |
d4a9ce78 | 693 | { |
87ab7992 DZ |
694 | static int still_looking = 1; |
695 | ||
d4a9ce78 JH |
696 | if (!cmitmsg) |
697 | return 0; | |
2744b234 | 698 | |
87ab7992 | 699 | if (still_looking) { |
3b6121f6 LS |
700 | strbuf_ltrim(line); |
701 | if (!line->len) | |
702 | return 0; | |
703 | if ((still_looking = check_header(line, s_hdr_data, 0)) != 0) | |
87ab7992 DZ |
704 | return 0; |
705 | } | |
8b4525fb | 706 | |
86747c13 DZ |
707 | /* normalize the log message to UTF-8. */ |
708 | if (metainfo_charset) | |
3b6121f6 | 709 | convert_to_utf8(line, charset.buf); |
86747c13 | 710 | |
f0658cf2 | 711 | if (patchbreak(line)) { |
87ab7992 DZ |
712 | fclose(cmitmsg); |
713 | cmitmsg = NULL; | |
714 | return 1; | |
715 | } | |
8b4525fb | 716 | |
3b6121f6 | 717 | fputs(line->buf, cmitmsg); |
d4a9ce78 | 718 | return 0; |
2744b234 LT |
719 | } |
720 | ||
3b6121f6 | 721 | static void handle_patch(const struct strbuf *line) |
2744b234 | 722 | { |
3b6121f6 | 723 | fwrite(line->buf, 1, line->len, patchfile); |
87ab7992 | 724 | patch_lines++; |
2744b234 LT |
725 | } |
726 | ||
3b6121f6 | 727 | static void handle_filter(struct strbuf *line) |
2744b234 | 728 | { |
87ab7992 | 729 | static int filter = 0; |
2744b234 | 730 | |
3b6121f6 | 731 | /* filter tells us which part we left off on */ |
87ab7992 DZ |
732 | switch (filter) { |
733 | case 0: | |
3b6121f6 | 734 | if (!handle_commit_msg(line)) |
d4a9ce78 | 735 | break; |
87ab7992 DZ |
736 | filter++; |
737 | case 1: | |
3b6121f6 LS |
738 | handle_patch(line); |
739 | break; | |
2744b234 LT |
740 | } |
741 | } | |
742 | ||
87ab7992 | 743 | static void handle_body(void) |
1d8fa411 | 744 | { |
3b6121f6 LS |
745 | int len = 0; |
746 | struct strbuf prev = STRBUF_INIT; | |
d4a9ce78 JH |
747 | |
748 | /* Skip up to the first boundary */ | |
3b6121f6 | 749 | if (*content_top) { |
87ab7992 | 750 | if (!find_boundary()) |
3b6121f6 | 751 | goto handle_body_out; |
87ab7992 DZ |
752 | } |
753 | ||
754 | do { | |
3b6121f6 LS |
755 | strbuf_setlen(&line, line.len + len); |
756 | ||
87ab7992 | 757 | /* process any boundary lines */ |
3b6121f6 | 758 | if (*content_top && is_multipart_boundary(&line)) { |
87ab7992 | 759 | /* flush any leftover */ |
3b6121f6 LS |
760 | if (line.len) |
761 | handle_filter(&line); | |
762 | ||
87ab7992 | 763 | if (!handle_boundary()) |
3b6121f6 | 764 | goto handle_body_out; |
87ab7992 DZ |
765 | } |
766 | ||
86747c13 | 767 | /* Unwrap transfer encoding */ |
3b6121f6 | 768 | decode_transfer_encoding(&line); |
87ab7992 DZ |
769 | |
770 | switch (transfer_encoding) { | |
771 | case TE_BASE64: | |
87f1b884 | 772 | case TE_QP: |
87ab7992 | 773 | { |
3b6121f6 LS |
774 | struct strbuf **lines, **it, *sb; |
775 | ||
776 | /* Prepend any previous partial lines */ | |
777 | strbuf_insert(&line, 0, prev.buf, prev.len); | |
778 | strbuf_reset(&prev); | |
87ab7992 DZ |
779 | |
780 | /* binary data most likely doesn't have newlines */ | |
781 | if (message_type != TYPE_TEXT) { | |
3b6121f6 | 782 | handle_filter(&line); |
87ab7992 DZ |
783 | break; |
784 | } | |
9aa23094 JH |
785 | /* |
786 | * This is a decoded line that may contain | |
87ab7992 DZ |
787 | * multiple new lines. Pass only one chunk |
788 | * at a time to handle_filter() | |
789 | */ | |
3b6121f6 LS |
790 | lines = strbuf_split(&line, '\n'); |
791 | for (it = lines; (sb = *it); it++) { | |
792 | if (*(it + 1) == NULL) /* The last line */ | |
793 | if (sb->buf[sb->len - 1] != '\n') { | |
794 | /* Partial line, save it for later. */ | |
795 | strbuf_addbuf(&prev, sb); | |
796 | break; | |
797 | } | |
798 | handle_filter(sb); | |
799 | } | |
9aa23094 | 800 | /* |
3b6121f6 | 801 | * The partial chunk is saved in "prev" and will be |
9aa23094 | 802 | * appended by the next iteration of read_line_with_nul(). |
87ab7992 | 803 | */ |
3b6121f6 | 804 | strbuf_list_free(lines); |
d4a9ce78 | 805 | break; |
1d8fa411 | 806 | } |
87ab7992 | 807 | default: |
3b6121f6 | 808 | handle_filter(&line); |
d4a9ce78 | 809 | } |
87ab7992 | 810 | |
3b6121f6 LS |
811 | strbuf_reset(&line); |
812 | if (strbuf_avail(&line) < 100) | |
813 | strbuf_grow(&line, 100); | |
814 | } while ((len = read_line_with_nul(line.buf, strbuf_avail(&line), fin))); | |
815 | ||
816 | handle_body_out: | |
817 | strbuf_release(&prev); | |
1d8fa411 JH |
818 | } |
819 | ||
3b6121f6 | 820 | static void output_header_lines(FILE *fout, const char *hdr, const struct strbuf *data) |
d7f6bae2 | 821 | { |
3b6121f6 | 822 | const char *sp = data->buf; |
d7f6bae2 | 823 | while (1) { |
3b6121f6 | 824 | char *ep = strchr(sp, '\n'); |
d7f6bae2 JH |
825 | int len; |
826 | if (!ep) | |
3b6121f6 | 827 | len = strlen(sp); |
d7f6bae2 | 828 | else |
3b6121f6 LS |
829 | len = ep - sp; |
830 | fprintf(fout, "%s: %.*s\n", hdr, len, sp); | |
d7f6bae2 JH |
831 | if (!ep) |
832 | break; | |
3b6121f6 | 833 | sp = ep + 1; |
d7f6bae2 JH |
834 | } |
835 | } | |
836 | ||
87ab7992 | 837 | static void handle_info(void) |
2744b234 | 838 | { |
3b6121f6 | 839 | struct strbuf *hdr; |
87ab7992 DZ |
840 | int i; |
841 | ||
842 | for (i = 0; header[i]; i++) { | |
87ab7992 DZ |
843 | /* only print inbody headers if we output a patch file */ |
844 | if (patch_lines && s_hdr_data[i]) | |
845 | hdr = s_hdr_data[i]; | |
846 | else if (p_hdr_data[i]) | |
847 | hdr = p_hdr_data[i]; | |
848 | else | |
849 | continue; | |
850 | ||
851 | if (!memcmp(header[i], "Subject", 7)) { | |
3b6121f6 LS |
852 | if (!keep_subject) { |
853 | cleanup_subject(hdr); | |
854 | cleanup_space(hdr); | |
d7f6bae2 | 855 | } |
3b6121f6 | 856 | output_header_lines(fout, "Subject", hdr); |
87ab7992 DZ |
857 | } else if (!memcmp(header[i], "From", 4)) { |
858 | handle_from(hdr); | |
3b6121f6 LS |
859 | fprintf(fout, "Author: %s\n", name.buf); |
860 | fprintf(fout, "Email: %s\n", email.buf); | |
87ab7992 DZ |
861 | } else { |
862 | cleanup_space(hdr); | |
3b6121f6 | 863 | fprintf(fout, "%s: %s\n", header[i], hdr->buf); |
87ab7992 | 864 | } |
d4a9ce78 | 865 | } |
87ab7992 | 866 | fprintf(fout, "\n"); |
2744b234 LT |
867 | } |
868 | ||
fcd056a6 JH |
869 | static int mailinfo(FILE *in, FILE *out, int ks, const char *encoding, |
870 | const char *msg, const char *patch) | |
34488e3c | 871 | { |
f88a545a | 872 | int peek; |
34488e3c LS |
873 | keep_subject = ks; |
874 | metainfo_charset = encoding; | |
875 | fin = in; | |
876 | fout = out; | |
877 | ||
878 | cmitmsg = fopen(msg, "w"); | |
879 | if (!cmitmsg) { | |
880 | perror(msg); | |
881 | return -1; | |
882 | } | |
883 | patchfile = fopen(patch, "w"); | |
884 | if (!patchfile) { | |
885 | perror(patch); | |
886 | fclose(cmitmsg); | |
887 | return -1; | |
888 | } | |
87ab7992 | 889 | |
3b6121f6 LS |
890 | p_hdr_data = xcalloc(MAX_HDR_PARSED, sizeof(*p_hdr_data)); |
891 | s_hdr_data = xcalloc(MAX_HDR_PARSED, sizeof(*s_hdr_data)); | |
87ab7992 | 892 | |
f88a545a SS |
893 | do { |
894 | peek = fgetc(in); | |
895 | } while (isspace(peek)); | |
896 | ungetc(peek, in); | |
897 | ||
87ab7992 | 898 | /* process the email header */ |
3b6121f6 LS |
899 | while (read_one_header_line(&line, fin)) |
900 | check_header(&line, p_hdr_data, 1); | |
87ab7992 DZ |
901 | |
902 | handle_body(); | |
903 | handle_info(); | |
34488e3c LS |
904 | |
905 | return 0; | |
906 | } | |
907 | ||
6bff6a60 | 908 | static const char mailinfo_usage[] = |
588c038a | 909 | "git mailinfo [-k] [-u | --encoding=<encoding> | -n] msg patch <mail >info"; |
d4a9ce78 | 910 | |
a633fca0 | 911 | int cmd_mailinfo(int argc, const char **argv, const char *prefix) |
2744b234 | 912 | { |
bb1091a4 JH |
913 | const char *def_charset; |
914 | ||
f1f909e3 JH |
915 | /* NEEDSWORK: might want to do the optional .git/ directory |
916 | * discovery | |
917 | */ | |
ef90d6d4 | 918 | git_config(git_default_config, NULL); |
f1f909e3 | 919 | |
bb1091a4 JH |
920 | def_charset = (git_commit_encoding ? git_commit_encoding : "utf-8"); |
921 | metainfo_charset = def_charset; | |
922 | ||
6bff6a60 JH |
923 | while (1 < argc && argv[1][0] == '-') { |
924 | if (!strcmp(argv[1], "-k")) | |
925 | keep_subject = 1; | |
d4a9ce78 | 926 | else if (!strcmp(argv[1], "-u")) |
bb1091a4 JH |
927 | metainfo_charset = def_charset; |
928 | else if (!strcmp(argv[1], "-n")) | |
929 | metainfo_charset = NULL; | |
cc44c765 | 930 | else if (!prefixcmp(argv[1], "--encoding=")) |
9f63892b | 931 | metainfo_charset = argv[1] + 11; |
d4a9ce78 | 932 | else |
f1f909e3 | 933 | usage(mailinfo_usage); |
6bff6a60 JH |
934 | argc--; argv++; |
935 | } | |
936 | ||
a196d8d4 | 937 | if (argc != 3) |
f1f909e3 | 938 | usage(mailinfo_usage); |
34488e3c LS |
939 | |
940 | return !!mailinfo(stdin, stdout, keep_subject, metainfo_charset, argv[1], argv[2]); | |
2744b234 | 941 | } |