3b6121f69b · matheusbsbs/matheus-bagliano

@@ -5,14 +5,15 @@

5 #include "cache.h"

6 #include "builtin.h"

7 #include "utf8.h"

8 +#include "strbuf.h"

9

10 static FILE *cmitmsg, *patchfile, *fin, *fout;

11

12 static int keep_subject;

13 static const char *metainfo_charset;

13 -static char line[1000];

14 -static char name[1000];

15 -static char email[1000];

14 +static struct strbuf line = STRBUF_INIT;

15 +static struct strbuf name = STRBUF_INIT;

16 +static struct strbuf email = STRBUF_INIT;

17

18 static enum {

19 TE_DONTCARE, TE_QP, TE_BASE64,

@@ -21,74 +22,79 @@ static enum {

22 TYPE_TEXT, TYPE_OTHER,

23 } message_type;

24

24 -static char charset[256];

25 +static struct strbuf charset = STRBUF_INIT;

26 static int patch_lines;

26 -static char **p_hdr_data, **s_hdr_data;

27 +static struct strbuf **p_hdr_data, **s_hdr_data;

28

29 #define MAX_HDR_PARSED 10

30 #define MAX_BOUNDARIES 5

31

31 -static char *sanity_check(char *name, char *email)

32 +static void get_sane_name(struct strbuf *out, struct strbuf *name, struct strbuf *email)

33 {

33 - int len = strlen(name);

34 - if (len < 3 || len > 60)

35 - return email;

36 - if (strchr(name, '@') || strchr(name, '<') || strchr(name, '>'))

37 - return email;

38 - return name;

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);

42 }

43

41 -static int bogus_from(char *line)

44 +static void parse_bogus_from(const struct strbuf *line)

45 {

46 /* John Doe <johndoe> */

44 - char *bra, *ket, *dst, *cp;

47

48 + char *bra, *ket;

49 /* This is fallback, so do not bother if we already have an

50 * e-mail address.

51 */

49 - if (*email)

50 - return 0;

52 + if (email.len)

53 + return;

54

52 - bra = strchr(line, '<');

55 + bra = strchr(line->buf, '<');

56 if (!bra)

54 - return 0;

57 + return;

58 ket = strchr(bra, '>');

59 if (!ket)

57 - return 0;

60 + return;

61

59 - for (dst = email, cp = bra+1; cp < ket; )

60 - *dst++ = *cp++;

61 - *dst = 0;

62 - for (cp = line; isspace(*cp); cp++)

63 - ;

64 - for (bra--; isspace(*bra); bra--)

65 - *bra = 0;

66 - cp = sanity_check(cp, email);

67 - strcpy(name, cp);

68 - return 1;

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);

69 }

70

71 -static int handle_from(char *in_line)

71 +static void handle_from(const struct strbuf *from)

72 {

73 - char line[1000];

73 char *at;

75 - char *dst;

74 + size_t el;

75 + struct strbuf f;

76

77 - strcpy(line, in_line);

78 - at = strchr(line, '@');

79 - if (!at)

80 - return bogus_from(line);

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 + }

85

86 /*

87 * If we already have one email, don't take any confusing lines

88 */

85 - if (*email && strchr(at+1, '@'))

86 - return 0;

89 + if (email.len && strchr(at + 1, '@')) {

90 + strbuf_release(&f);

91 + return;

92 + }

93

94 /* Pick up the string around '@', possibly delimited with <>

89 - * pair; that is the email part. White them out while copying.

95 + * pair; that is the email part.

96 */

91 - while (at > line) {

97 + while (at > f.buf) {

98 char c = at[-1];

99 if (isspace(c))

100 break;

@@ -98,56 +104,35 @@ static int handle_from(char *in_line)

104 }

105 at--;

106 }

101 - dst = email;

102 - for (;;) {

103 - unsigned char c = *at;

104 - if (!c || c == '>' || isspace(c)) {

105 - if (c == '>')

106 - *at = ' ';

107 - break;

108 - }

109 - *at++ = ' ';

110 - *dst++ = c;

111 - }

112 - *dst++ = 0;

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);

111

112 /* The remainder is name. It could be "John Doe <john.doe@xz>"

115 - * or "john.doe@xz (John Doe)", but we have whited out the

113 + * or "john.doe@xz (John Doe)", but we have removed the

114 * email part, so trim from both ends, possibly removing

115 * the () pair at the end.

116 */

119 - at = line + strlen(line);

120 - while (at > line) {

121 - unsigned char c = *--at;

122 - if (!isspace(c)) {

123 - at[(c == ')') ? 0 : 1] = 0;

124 - break;

125 - }

126 - }

117 + strbuf_trim(&f);

118 + if (f.buf[0] == '(')

119 + strbuf_remove(&name, 0, 1);

120 + if (f.len && f.buf[f.len - 1] == ')')

121 + strbuf_setlen(&f, f.len - 1);

122

128 - at = line;

129 - for (;;) {

130 - unsigned char c = *at;

131 - if (!c || !isspace(c)) {

132 - if (c == '(')

133 - at++;

134 - break;

135 - }

136 - at++;

137 - }

138 - at = sanity_check(at, email);

139 - strcpy(name, at);

140 - return 1;

123 + get_sane_name(&name, &f, &email);

124 + strbuf_release(&f);

125 }

126

143 -static int handle_header(char *line, char *data, int ofs)

127 +static void handle_header(struct strbuf **out, const struct strbuf *line)

128 {

145 - if (!line || !data)

146 - return 1;

129 + if (!*out) {

130 + *out = xmalloc(sizeof(struct strbuf));

131 + strbuf_init(*out, line->len);

132 + } else

133 + strbuf_reset(*out);

134

148 - strcpy(data, line+ofs);

149 -

150 - return 0;

135 + strbuf_addbuf(*out, line);

136 }

137

138 /* NOTE NOTE NOTE. We do not claim we do full MIME. We just attempt

@@ -156,13 +141,13 @@ static int handle_header(char *line, char *data, int ofs)

141 * case insensitively.

142 */

143

159 -static int slurp_attr(const char *line, const char *name, char *attr)

144 +static int slurp_attr(const char *line, const char *name, struct strbuf *attr)

145 {

146 const char *ends, *ap = strcasestr(line, name);

147 size_t sz;

148

149 if (!ap) {

165 - *attr = 0;

150 + strbuf_setlen(attr, 0);

151 return 0;

152 }

153 ap += strlen(name);

@@ -173,180 +158,172 @@ static int slurp_attr(const char *line, const char *name, char *attr)

158 else

159 ends = "; \t";

160 sz = strcspn(ap, ends);

176 - memcpy(attr, ap, sz);

177 - attr[sz] = 0;

161 + strbuf_add(attr, ap, sz);

162 return 1;

163 }

164

181 -struct content_type {

182 - char *boundary;

183 - int boundary_len;

184 -};

185 -

186 -static struct content_type content[MAX_BOUNDARIES];

165 +static struct strbuf *content[MAX_BOUNDARIES];

166

188 -static struct content_type *content_top = content;

167 +static struct strbuf **content_top = content;

168

190 -static int handle_content_type(char *line)

169 +static void handle_content_type(struct strbuf *line)

170 {

192 - char boundary[256];

171 + struct strbuf *boundary = xmalloc(sizeof(struct strbuf));

172 + strbuf_init(boundary, line->len);

173

194 - if (strcasestr(line, "text/") == NULL)

174 + if (!strcasestr(line->buf, "text/"))

175 message_type = TYPE_OTHER;

196 - if (slurp_attr(line, "boundary=", boundary + 2)) {

197 - memcpy(boundary, "--", 2);

176 + if (slurp_attr(line->buf, "boundary=", boundary)) {

177 + strbuf_insert(boundary, 0, "--", 2);

178 if (content_top++ >= &content[MAX_BOUNDARIES]) {

179 fprintf(stderr, "Too many boundaries to handle\n");

180 exit(1);

181 }

202 - content_top->boundary_len = strlen(boundary);

203 - content_top->boundary = xmalloc(content_top->boundary_len+1);

204 - strcpy(content_top->boundary, boundary);

182 + *content_top = boundary;

183 + boundary = NULL;

184 }

206 - if (slurp_attr(line, "charset=", charset)) {

207 - int i, c;

208 - for (i = 0; (c = charset[i]) != 0; i++)

209 - charset[i] = tolower(c);

185 + if (slurp_attr(line->buf, "charset=", &charset))

186 + strbuf_tolower(&charset);

187 +

188 + if (boundary) {

189 + strbuf_release(boundary);

190 + free(boundary);

191 }

211 - return 0;

192 }

193

214 -static int handle_content_transfer_encoding(char *line)

194 +static void handle_content_transfer_encoding(const struct strbuf *line)

195 {

216 - if (strcasestr(line, "base64"))

196 + if (strcasestr(line->buf, "base64"))

197 transfer_encoding = TE_BASE64;

218 - else if (strcasestr(line, "quoted-printable"))

198 + else if (strcasestr(line->buf, "quoted-printable"))

199 transfer_encoding = TE_QP;

200 else

201 transfer_encoding = TE_DONTCARE;

222 - return 0;

202 }

203

225 -static int is_multipart_boundary(const char *line)

204 +static int is_multipart_boundary(const struct strbuf *line)

205 {

227 - return (!memcmp(line, content_top->boundary, content_top->boundary_len));

206 + return !strbuf_cmp(line, *content_top);

207 }

208

230 -static int eatspace(char *line)

209 +static void cleanup_subject(struct strbuf *subject)

210 {

232 - int len = strlen(line);

233 - while (len > 0 && isspace(line[len-1]))

234 - line[--len] = 0;

235 - return len;

236 -}

237 -

238 -static char *cleanup_subject(char *subject)

239 -{

240 - for (;;) {

241 - char *p;

242 - int len, remove;

243 - switch (*subject) {

211 + char *pos;

212 + size_t remove;

213 + while (subject->len) {

214 + switch (*subject->buf) {

215 case 'r': case 'R':

245 - if (!memcmp("e:", subject+1, 2)) {

246 - subject += 3;

216 + if (subject->len <= 3)

217 + break;

218 + if (!memcmp(subject->buf + 1, "e:", 2)) {

219 + strbuf_remove(subject, 0, 3);

220 continue;

221 }

222 break;

223 case ' ': case '\t': case ':':

251 - subject++;

224 + strbuf_remove(subject, 0, 1);

225 continue;

253 -

226 case '[':

255 - p = strchr(subject, ']');

256 - if (!p) {

257 - subject++;

258 - continue;

259 - }

260 - len = strlen(p);

261 - remove = p - subject;

262 - if (remove <= len *2) {

263 - subject = p+1;

264 - continue;

265 - }

227 + if ((pos = strchr(subject->buf, ']'))) {

228 + remove = pos - subject->buf + 1;

229 + /* Don't remove too much. */

230 + if (remove <= (subject->len - remove + 1) * 2) {

231 + strbuf_remove(subject, 0, remove);

232 + continue;

233 + }

234 + } else

235 + strbuf_remove(subject, 0, 1);

236 break;

237 }

268 - eatspace(subject);

269 - return subject;

238 + strbuf_trim(subject);

239 + return;

240 }

241 }

242

273 -static void cleanup_space(char *buf)

243 +static void cleanup_space(struct strbuf *sb)

244 {

275 - unsigned char c;

276 - while ((c = *buf) != 0) {

277 - buf++;

278 - if (isspace(c)) {

279 - buf[-1] = ' ';

280 - c = *buf;

281 - while (isspace(c)) {

282 - int len = strlen(buf);

283 - memmove(buf, buf+1, len);

284 - c = *buf;

285 - }

245 + size_t pos, cnt;

246 + for (pos = 0; pos < sb->len; pos++) {

247 + if (isspace(sb->buf[pos])) {

248 + sb->buf[pos] = ' ';

249 + for (cnt = 0; isspace(sb->buf[pos + cnt + 1]); cnt++);

250 + strbuf_remove(sb, pos + 1, cnt);

251 }

252 }

253 }

254

290 -static void decode_header(char *it, unsigned itsize);

255 +static void decode_header(struct strbuf *line);

256 static const char *header[MAX_HDR_PARSED] = {

257 "From","Subject","Date",

258 };

259

295 -static int check_header(char *line, unsigned linesize, char **hdr_data, int overwrite)

260 +static inline int cmp_header(const struct strbuf *line, const char *hdr)

261 {

297 - int i;

262 + int len = strlen(hdr);

263 + return !strncasecmp(line->buf, hdr, len) && line->len > len &&

264 + line->buf[len] == ':' && isspace(line->buf[len + 1]);

265 +}

266

267 +static int check_header(const struct strbuf *line,

268 + struct strbuf *hdr_data[], int overwrite)

269 +{

270 + int i, ret = 0, len;

271 + struct strbuf sb = STRBUF_INIT;

272 /* search for the interesting parts */

273 for (i = 0; header[i]; i++) {

274 int len = strlen(header[i]);

302 - if ((!hdr_data[i] || overwrite) &&

303 - !strncasecmp(line, header[i], len) &&

304 - line[len] == ':' && isspace(line[len + 1])) {

275 + if ((!hdr_data[i] || overwrite) && cmp_header(line, header[i])) {

276 /* Unwrap inline B and Q encoding, and optionally

277 * normalize the meta information to utf8.

278 */

308 - decode_header(line + len + 2, linesize - len - 2);

309 - hdr_data[i] = xmalloc(1000 * sizeof(char));

310 - if (! handle_header(line, hdr_data[i], len + 2)) {

311 - return 1;

312 - }

279 + strbuf_add(&sb, line->buf + len + 2, line->len - len - 2);

280 + decode_header(&sb);

281 + handle_header(&hdr_data[i], &sb);

282 + ret = 1;

283 + goto check_header_out;

284 }

285 }

286

287 /* Content stuff */

317 - if (!strncasecmp(line, "Content-Type", 12) &&

318 - line[12] == ':' && isspace(line[12 + 1])) {

319 - decode_header(line + 12 + 2, linesize - 12 - 2);

320 - if (! handle_content_type(line)) {

321 - return 1;

322 - }

323 - }

324 - if (!strncasecmp(line, "Content-Transfer-Encoding", 25) &&

325 - line[25] == ':' && isspace(line[25 + 1])) {

326 - decode_header(line + 25 + 2, linesize - 25 - 2);

327 - if (! handle_content_transfer_encoding(line)) {

328 - return 1;

329 - }

288 + if (cmp_header(line, "Content-Type")) {

289 + len = strlen("Content-Type: ");

290 + strbuf_add(&sb, line->buf + len, line->len - len);

291 + decode_header(&sb);

292 + strbuf_insert(&sb, 0, "Content-Type: ", len);

293 + handle_content_type(&sb);

294 + ret = 1;

295 + goto check_header_out;

296 + }

297 + if (cmp_header(line, "Content-Transfer-Encoding")) {

298 + len = strlen("Content-Transfer-Encoding: ");

299 + strbuf_add(&sb, line->buf + len, line->len - len);

300 + decode_header(&sb);

301 + handle_content_transfer_encoding(&sb);

302 + ret = 1;

303 + goto check_header_out;

304 }

305

306 /* for inbody stuff */

333 - if (!memcmp(">From", line, 5) && isspace(line[5]))

334 - return 1;

335 - if (!memcmp("[PATCH]", line, 7) && isspace(line[7])) {

307 + if (!prefixcmp(line->buf, ">From") && isspace(line->buf[5])) {

308 + ret = 1; /* Should this return 0? */

309 + goto check_header_out;

310 + }

311 + if (!prefixcmp(line->buf, "[PATCH]") && isspace(line->buf[7])) {

312 for (i = 0; header[i]; i++) {

337 - if (!memcmp("Subject: ", header[i], 9)) {

338 - if (! handle_header(line, hdr_data[i], 0)) {

339 - return 1;

340 - }

313 + if (!memcmp("Subject", header[i], 7)) {

314 + handle_header(&hdr_data[i], line);

315 + ret = 1;

316 + goto check_header_out;

317 }

318 }

319 }

320

345 - /* no match */

346 - return 0;

321 +check_header_out:

322 + strbuf_release(&sb);

323 + return ret;

324 }

325

349 -static int is_rfc2822_header(char *line)

326 +static int is_rfc2822_header(const struct strbuf *line)

327 {

328 /*

329 * The section that defines the loosest possible

@@ -357,15 +334,15 @@ static int is_rfc2822_header(char *line)

334 * ftext = %d33-57 / %59-126

335 */

336 int ch;

360 - char *cp = line;

337 + char *cp = line->buf;

338

339 /* Count mbox From headers as headers */

363 - if (!memcmp(line, "From ", 5) || !memcmp(line, ">From ", 6))

340 + if (!prefixcmp(cp, "From ") || !prefixcmp(cp, ">From "))

341 return 1;

342

343 while ((ch = *cp++)) {

344 if (ch == ':')

368 - return cp != line;

345 + return 1;

346 if ((33 <= ch && ch <= 57) ||

347 (59 <= ch && ch <= 126))

348 continue;

@@ -374,34 +351,20 @@ static int is_rfc2822_header(char *line)

351 return 0;

352 }

353

377 -/*

378 - * sz is size of 'line' buffer in bytes. Must be reasonably

379 - * long enough to hold one physical real-world e-mail line.

380 - */

381 -static int read_one_header_line(char *line, int sz, FILE *in)

354 +static int read_one_header_line(struct strbuf *line, FILE *in)

355 {

383 - int len;

384 -

385 - /*

386 - * We will read at most (sz-1) bytes and then potentially

387 - * re-add NUL after it. Accessing line[sz] after this is safe

388 - * and we can allow len to grow up to and including sz.

389 - */

390 - sz--;

391 -

356 /* Get the first part of the line. */

393 - if (!fgets(line, sz, in))

357 + if (strbuf_getline(line, in, '\n'))

358 return 0;

359

360 /*

361 * Is it an empty line or not a valid rfc2822 header?

362 * If so, stop here, and return false ("not a header")

363 */

400 - len = eatspace(line);

401 - if (!len || !is_rfc2822_header(line)) {

364 + strbuf_rtrim(line);

365 + if (!line->len || !is_rfc2822_header(line)) {

366 /* Re-add the newline */

403 - line[len] = '\n';

404 - line[len + 1] = '\0';

367 + strbuf_addch(line, '\n');

368 return 0;

369 }

370

@@ -410,65 +373,53 @@ static int read_one_header_line(char *line, int sz, FILE *in)

373 * Yuck, 2822 header "folding"

374 */

375 for (;;) {

413 - int peek, addlen;

414 - static char continuation[1000];

376 + int peek;

377 + struct strbuf continuation = STRBUF_INIT;

378

379 peek = fgetc(in); ungetc(peek, in);

380 if (peek != ' ' && peek != '\t')

381 break;

419 - if (!fgets(continuation, sizeof(continuation), in))

382 + if (strbuf_getline(&continuation, in, '\n'))

383 break;

421 - addlen = eatspace(continuation);

422 - if (len < sz - 1) {

423 - if (addlen >= sz - len)

424 - addlen = sz - len - 1;

425 - memcpy(line + len, continuation, addlen);

426 - line[len] = '\n';

427 - len += addlen;

428 - }

384 + continuation.buf[0] = '\n';

385 + strbuf_rtrim(&continuation);

386 + strbuf_addbuf(line, &continuation);

387 }

430 - line[len] = 0;

388

389 return 1;

390 }

391

435 -static int decode_q_segment(char *in, char *ot, unsigned otsize, char *ep, int rfc2047)

392 +static struct strbuf *decode_q_segment(const struct strbuf *q_seg, int rfc2047)

393 {

437 - char *otbegin = ot;

438 - char *otend = ot + otsize;

394 + const char *in = q_seg->buf;

395 int c;

440 - while ((c = *in++) != 0 && (in <= ep)) {

441 - if (ot == otend) {

442 - *--ot = '\0';

443 - return -1;

444 - }

396 + struct strbuf *out = xmalloc(sizeof(struct strbuf));

397 + strbuf_init(out, q_seg->len);

398 +

399 + while ((c = *in++) != 0) {

400 if (c == '=') {

401 int d = *in++;

402 if (d == '\n' || !d)

403 break; /* drop trailing newline */

449 - *ot++ = ((hexval(d) << 4) | hexval(*in++));

404 + strbuf_addch(out, (hexval(d) << 4) | hexval(*in++));

405 continue;

406 }

407 if (rfc2047 && c == '_') /* rfc2047 4.2 (2) */

408 c = 0x20;

454 - *ot++ = c;

409 + strbuf_addch(out, c);

410 }

456 - *ot = 0;

457 - return (ot - otbegin);

411 + return out;

412 }

413

460 -static int decode_b_segment(char *in, char *ot, unsigned otsize, char *ep)

414 +static struct strbuf *decode_b_segment(const struct strbuf *b_seg)

415 {

416 /* Decode in..ep, possibly in-place to ot */

417 int c, pos = 0, acc = 0;

464 - char *otbegin = ot;

465 - char *otend = ot + otsize;

418 + const char *in = b_seg->buf;

419 + struct strbuf *out = xmalloc(sizeof(struct strbuf));

420 + strbuf_init(out, b_seg->len);

421

467 - while ((c = *in++) != 0 && (in <= ep)) {

468 - if (ot == otend) {

469 - *--ot = '\0';

470 - return -1;

471 - }

422 + while ((c = *in++) != 0) {

423 if (c == '+')

424 c = 62;

425 else if (c == '/')

@@ -493,21 +444,20 @@ static int decode_b_segment(char *in, char *ot, unsigned otsize, char *ep)

444 acc = (c << 2);

445 break;

446 case 1:

496 - *ot++ = (acc | (c >> 4));

447 + strbuf_addch(out, (acc | (c >> 4)));

448 acc = (c & 15) << 4;

449 break;

450 case 2:

500 - *ot++ = (acc | (c >> 2));

451 + strbuf_addch(out, (acc | (c >> 2)));

452 acc = (c & 3) << 6;

453 break;

454 case 3:

504 - *ot++ = (acc | c);

455 + strbuf_addch(out, (acc | c));

456 acc = pos = 0;

457 break;

458 }

459 }

509 - *ot = 0;

510 - return (ot - otbegin);

460 + return out;

461 }

462

463 /*

@@ -521,16 +471,16 @@ static int decode_b_segment(char *in, char *ot, unsigned otsize, char *ep)

471 * Otherwise, we default to assuming it is Latin1 for historical

472 * reasons.

473 */

524 -static const char *guess_charset(const char *line, const char *target_charset)

474 +static const char *guess_charset(const struct strbuf *line, const char *target_charset)

475 {

476 if (is_encoding_utf8(target_charset)) {

527 - if (is_utf8(line))

477 + if (is_utf8(line->buf))

478 return NULL;

479 }

480 return "latin1";

481 }

482

533 -static void convert_to_utf8(char *line, unsigned linesize, const char *charset)

483 +static void convert_to_utf8(struct strbuf *line, const char *charset)

484 {

485 char *out;

486

@@ -542,112 +492,118 @@ static void convert_to_utf8(char *line, unsigned linesize, const char *charset)

492

493 if (!strcmp(metainfo_charset, charset))

494 return;

545 - out = reencode_string(line, metainfo_charset, charset);

495 + out = reencode_string(line->buf, metainfo_charset, charset);

496 if (!out)

497 die("cannot convert from %s to %s\n",

498 charset, metainfo_charset);

549 - strlcpy(line, out, linesize);

550 - free(out);

499 + strbuf_attach(line, out, strlen(out), strlen(out));

500 }

501

553 -static int decode_header_bq(char *it, unsigned itsize)

502 +static int decode_header_bq(struct strbuf *it)

503 {

555 - char *in, *out, *ep, *cp, *sp;

556 - char outbuf[1000];

504 + char *in, *ep, *cp;

505 + struct strbuf outbuf = STRBUF_INIT, *dec;

506 + struct strbuf charset_q = STRBUF_INIT, piecebuf = STRBUF_INIT;

507 int rfc2047 = 0;

508

559 - in = it;

560 - out = outbuf;

561 - while ((ep = strstr(in, "=?")) != NULL) {

562 - int sz, encoding;

563 - char charset_q[256], piecebuf[256];

509 + in = it->buf;

510 + while (in - it->buf <= it->len && (ep = strstr(in, "=?")) != NULL) {

511 + int encoding;

512 + strbuf_reset(&charset_q);

513 + strbuf_reset(&piecebuf);

514 rfc2047 = 1;

515

516 if (in != ep) {

567 - sz = ep - in;

568 - memcpy(out, in, sz);

569 - out += sz;

570 - in += sz;

517 + strbuf_add(&outbuf, in, ep - in);

518 + in = ep;

519 }

520 /* E.g.

521 * ep : "=?iso-2022-jp?B?GyR...?= foo"

522 * ep : "=?ISO-8859-1?Q?Foo=FCbar?= baz"

523 */

524 ep += 2;

577 - cp = strchr(ep, '?');

578 - if (!cp)

579 - return rfc2047; /* no munging */

580 - for (sp = ep; sp < cp; sp++)

581 - charset_q[sp - ep] = tolower(*sp);

582 - charset_q[cp - ep] = 0;

525 +

526 + if (ep - it->buf >= it->len || !(cp = strchr(ep, '?')))

527 + goto decode_header_bq_out;

528 +

529 + if (cp + 3 - it->buf > it->len)

530 + goto decode_header_bq_out;

531 + strbuf_add(&charset_q, ep, cp - ep);

532 + strbuf_tolower(&charset_q);

533 +

534 encoding = cp[1];

535 if (!encoding || cp[2] != '?')

585 - return rfc2047; /* no munging */

536 + goto decode_header_bq_out;

537 ep = strstr(cp + 3, "?=");

538 if (!ep)

588 - return rfc2047; /* no munging */

539 + goto decode_header_bq_out;

540 + strbuf_add(&piecebuf, cp + 3, ep - cp - 3);

541 switch (tolower(encoding)) {

542 default:

591 - return rfc2047; /* no munging */

543 + goto decode_header_bq_out;

544 case 'b':

593 - sz = decode_b_segment(cp + 3, piecebuf, sizeof(piecebuf), ep);

545 + dec = decode_b_segment(&piecebuf);

546 break;

547 case 'q':

596 - sz = decode_q_segment(cp + 3, piecebuf, sizeof(piecebuf), ep, 1);

548 + dec = decode_q_segment(&piecebuf, 1);

549 break;

550 }

599 - if (sz < 0)

600 - return rfc2047;

551 if (metainfo_charset)

602 - convert_to_utf8(piecebuf, sizeof(piecebuf), charset_q);

552 + convert_to_utf8(dec, charset_q.buf);

553

604 - sz = strlen(piecebuf);

605 - if (outbuf + sizeof(outbuf) <= out + sz)

606 - return rfc2047; /* no munging */

607 - strcpy(out, piecebuf);

608 - out += sz;

554 + strbuf_addbuf(&outbuf, dec);

555 + strbuf_release(dec);

556 + free(dec);

557 in = ep + 2;

558 }

611 - strcpy(out, in);

612 - strlcpy(it, outbuf, itsize);

559 + strbuf_addstr(&outbuf, in);

560 + strbuf_reset(it);

561 + strbuf_addbuf(it, &outbuf);

562 +decode_header_bq_out:

563 + strbuf_release(&outbuf);

564 + strbuf_release(&charset_q);

565 + strbuf_release(&piecebuf);

566 return rfc2047;

567 }

568

616 -static void decode_header(char *it, unsigned itsize)

569 +static void decode_header(struct strbuf *it)

570 {

618 -

619 - if (decode_header_bq(it, itsize))

571 + if (decode_header_bq(it))

572 return;

573 /* otherwise "it" is a straight copy of the input.

574 * This can be binary guck but there is no charset specified.

575 */

576 if (metainfo_charset)

625 - convert_to_utf8(it, itsize, "");

577 + convert_to_utf8(it, "");

578 }

579

628 -static int decode_transfer_encoding(char *line, unsigned linesize, int inputlen)

580 +static void decode_transfer_encoding(struct strbuf *line)

581 {

630 - char *ep;

582 + struct strbuf *ret;

583

584 switch (transfer_encoding) {

585 case TE_QP:

634 - ep = line + inputlen;

635 - return decode_q_segment(line, line, linesize, ep, 0);

586 + ret = decode_q_segment(line, 0);

587 + break;

588 case TE_BASE64:

637 - ep = line + inputlen;

638 - return decode_b_segment(line, line, linesize, ep);

589 + ret = decode_b_segment(line);

590 + break;

591 case TE_DONTCARE:

592 default:

641 - return inputlen;

593 + return;

594 }

595 + strbuf_reset(line);

596 + strbuf_addbuf(line, ret);

597 + strbuf_release(ret);

598 + free(ret);

599 }

600

645 -static int handle_filter(char *line, unsigned linesize, int linelen);

601 +static void handle_filter(struct strbuf *line);

602

603 static int find_boundary(void)

604 {

649 - while(fgets(line, sizeof(line), fin) != NULL) {

650 - if (is_multipart_boundary(line))

605 + while (!strbuf_getline(&line, fin, '\n')) {

606 + if (is_multipart_boundary(&line))

607 return 1;

608 }

609 return 0;

@@ -655,12 +611,17 @@ static int find_boundary(void)

611

612 static int handle_boundary(void)

613 {

658 - char newline[]="\n";

614 + struct strbuf newline = STRBUF_INIT;

615 +

616 + strbuf_addch(&newline, '\n');

617 again:

660 - if (!memcmp(line+content_top->boundary_len, "--", 2)) {

618 + if (line.len >= (*content_top)->len + 2 &&

619 + !memcmp(line.buf + (*content_top)->len, "--", 2)) {

620 /* we hit an end boundary */

621 /* pop the current boundary off the stack */

663 - free(content_top->boundary);

622 + strbuf_release(*content_top);

623 + free(*content_top);

624 + *content_top = NULL;

625

626 /* technically won't happen as is_multipart_boundary()

627 will fail first. But just in case..

@@ -670,7 +631,8 @@ again:

631 "can't recover\n");

632 exit(1);

633 }

673 - handle_filter(newline, sizeof(newline), strlen(newline));

634 + handle_filter(&newline);

635 + strbuf_release(&newline);

636

637 /* skip to the next boundary */

638 if (!find_boundary())

@@ -680,39 +642,44 @@ again:

642

643 /* set some defaults */

644 transfer_encoding = TE_DONTCARE;

683 - charset[0] = 0;

645 + strbuf_reset(&charset);

646 message_type = TYPE_TEXT;

647

648 /* slurp in this section's info */

687 - while (read_one_header_line(line, sizeof(line), fin))

688 - check_header(line, sizeof(line), p_hdr_data, 0);

649 + while (read_one_header_line(&line, fin))

650 + check_header(&line, p_hdr_data, 0);

651

652 + strbuf_release(&newline);

653 /* eat the blank line after section info */

691 - return (fgets(line, sizeof(line), fin) != NULL);

654 + return (strbuf_getline(&line, fin, '\n') == 0);

655 }

656

694 -static inline int patchbreak(const char *line)

657 +static inline int patchbreak(const struct strbuf *line)

658 {

659 + size_t i;

660 +

661 /* Beginning of a "diff -" header? */

697 - if (!memcmp("diff -", line, 6))

662 + if (!prefixcmp(line->buf, "diff -"))

663 return 1;

664

665 /* CVS "Index: " line? */

701 - if (!memcmp("Index: ", line, 7))

666 + if (!prefixcmp(line->buf, "Index: "))

667 return 1;

668

669 /*

670 * "--- <filename>" starts patches without headers

671 * "---<sp>*" is a manual separator

672 */

708 - if (!memcmp("---", line, 3)) {

709 - line += 3;

673 + if (line->len < 4)

674 + return 0;

675 +

676 + if (!prefixcmp(line->buf, "---")) {

677 /* space followed by a filename? */

711 - if (line[0] == ' ' && !isspace(line[1]))

678 + if (line->buf[3] == ' ' && !isspace(line->buf[4]))

679 return 1;

680 /* Just whitespace? */

714 - for (;;) {

715 - unsigned char c = *line++;

681 + for (i = 3; i < line->len; i++) {

682 + unsigned char c = line->buf[i];

683 if (c == '\n')

684 return 1;

685 if (!isspace(c))

@@ -723,32 +690,24 @@ static inline int patchbreak(const char *line)

690 return 0;

691 }

692

726 -

727 -static int handle_commit_msg(char *line, unsigned linesize)

693 +static int handle_commit_msg(struct strbuf *line)

694 {

695 static int still_looking = 1;

730 - char *endline = line + linesize;

696

697 if (!cmitmsg)

698 return 0;

699

700 if (still_looking) {

736 - char *cp = line;

737 - if (isspace(*line)) {

738 - for (cp = line + 1; *cp; cp++) {

739 - if (!isspace(*cp))

740 - break;

741 - }

742 - if (!*cp)

743 - return 0;

744 - }

745 - if ((still_looking = check_header(cp, endline - cp, s_hdr_data, 0)) != 0)

701 + strbuf_ltrim(line);

702 + if (!line->len)

703 + return 0;

704 + if ((still_looking = check_header(line, s_hdr_data, 0)) != 0)

705 return 0;

706 }

707

708 /* normalize the log message to UTF-8. */

709 if (metainfo_charset)

751 - convert_to_utf8(line, endline - line, charset);

710 + convert_to_utf8(line, charset.buf);

711

712 if (patchbreak(line)) {

713 fclose(cmitmsg);

@@ -756,142 +715,132 @@ static int handle_commit_msg(char *line, unsigned linesize)

715 return 1;

716 }

717

759 - fputs(line, cmitmsg);

718 + fputs(line->buf, cmitmsg);

719 return 0;

720 }

721

763 -static int handle_patch(char *line, int len)

722 +static void handle_patch(const struct strbuf *line)

723 {

765 - fwrite(line, 1, len, patchfile);

724 + fwrite(line->buf, 1, line->len, patchfile);

725 patch_lines++;

767 - return 0;

726 }

727

770 -static int handle_filter(char *line, unsigned linesize, int linelen)

728 +static void handle_filter(struct strbuf *line)

729 {

730 static int filter = 0;

731

774 - /* filter tells us which part we left off on

775 - * a non-zero return indicates we hit a filter point

776 - */

732 + /* filter tells us which part we left off on */

733 switch (filter) {

734 case 0:

779 - if (!handle_commit_msg(line, linesize))

735 + if (!handle_commit_msg(line))

736 break;

737 filter++;

738 case 1:

783 - if (!handle_patch(line, linelen))

784 - break;

785 - filter++;

786 - default:

787 - return 1;

739 + handle_patch(line);

740 + break;

741 }

789 -

790 - return 0;

742 }

743

744 static void handle_body(void)

745 {

795 - int rc = 0;

796 - static char newline[2000];

797 - static char *np = newline;

798 - int len = strlen(line);

746 + int len = 0;

747 + struct strbuf prev = STRBUF_INIT;

748

749 /* Skip up to the first boundary */

801 - if (content_top->boundary) {

750 + if (*content_top) {

751 if (!find_boundary())

803 - return;

752 + goto handle_body_out;

753 }

754

755 do {

756 + strbuf_setlen(&line, line.len + len);

757 +

758 /* process any boundary lines */

808 - if (content_top->boundary && is_multipart_boundary(line)) {

759 + if (*content_top && is_multipart_boundary(&line)) {

760 /* flush any leftover */

810 - if (np != newline)

811 - handle_filter(newline, sizeof(newline),

812 - np - newline);

761 + if (line.len)

762 + handle_filter(&line);

763 +

764 if (!handle_boundary())

814 - return;

815 - len = strlen(line);

765 + goto handle_body_out;

766 }

767

768 /* Unwrap transfer encoding */

819 - len = decode_transfer_encoding(line, sizeof(line), len);

820 - if (len < 0) {

821 - error("Malformed input line");

822 - return;

823 - }

769 + decode_transfer_encoding(&line);

770

771 switch (transfer_encoding) {

772 case TE_BASE64:

773 case TE_QP:

774 {

829 - char *op = line;

775 + struct strbuf **lines, **it, *sb;

776 +

777 + /* Prepend any previous partial lines */

778 + strbuf_insert(&line, 0, prev.buf, prev.len);

779 + strbuf_reset(&prev);

780

781 /* binary data most likely doesn't have newlines */

782 if (message_type != TYPE_TEXT) {

833 - rc = handle_filter(line, sizeof(line), len);

783 + handle_filter(&line);

784 break;

785 }

836 -

786 /*

787 * This is a decoded line that may contain

788 * multiple new lines. Pass only one chunk

789 * at a time to handle_filter()

790 */

842 - do {

843 - while (op < line + len && *op != '\n')

844 - *np++ = *op++;

845 - *np = *op;

846 - if (*np != 0) {

847 - /* should be sitting on a new line */

848 - *(++np) = 0;

849 - op++;

850 - rc = handle_filter(newline, sizeof(newline), np - newline);

851 - np = newline;

852 - }

853 - } while (op < line + len);

791 + lines = strbuf_split(&line, '\n');

792 + for (it = lines; (sb = *it); it++) {

793 + if (*(it + 1) == NULL) /* The last line */

794 + if (sb->buf[sb->len - 1] != '\n') {

795 + /* Partial line, save it for later. */

796 + strbuf_addbuf(&prev, sb);

797 + break;

798 + }

799 + handle_filter(sb);

800 + }

801 /*

855 - * The partial chunk is saved in newline and will be

802 + * The partial chunk is saved in "prev" and will be

803 * appended by the next iteration of read_line_with_nul().

804 */

805 + strbuf_list_free(lines);

806 break;

807 }

808 default:

861 - rc = handle_filter(line, sizeof(line), len);

809 + handle_filter(&line);

810 }

863 - if (rc)

864 - /* nothing left to filter */

865 - break;

866 - } while ((len = read_line_with_nul(line, sizeof(line), fin)));

811

868 - return;

812 + strbuf_reset(&line);

813 + if (strbuf_avail(&line) < 100)

814 + strbuf_grow(&line, 100);

815 + } while ((len = read_line_with_nul(line.buf, strbuf_avail(&line), fin)));

816 +

817 +handle_body_out:

818 + strbuf_release(&prev);

819 }

820

871 -static void output_header_lines(FILE *fout, const char *hdr, char *data)

821 +static void output_header_lines(FILE *fout, const char *hdr, const struct strbuf *data)

822 {

823 + const char *sp = data->buf;

824 while (1) {

874 - char *ep = strchr(data, '\n');

825 + char *ep = strchr(sp, '\n');

826 int len;

827 if (!ep)

877 - len = strlen(data);

828 + len = strlen(sp);

829 else

879 - len = ep - data;

880 - fprintf(fout, "%s: %.*s\n", hdr, len, data);

830 + len = ep - sp;

831 + fprintf(fout, "%s: %.*s\n", hdr, len, sp);

832 if (!ep)

833 break;

883 - data = ep + 1;

834 + sp = ep + 1;

835 }

836 }

837

838 static void handle_info(void)

839 {

889 - char *sub;

890 - char *hdr;

840 + struct strbuf *hdr;

841 int i;

842

843 for (i = 0; header[i]; i++) {

894 -

844 /* only print inbody headers if we output a patch file */

845 if (patch_lines && s_hdr_data[i])

846 hdr = s_hdr_data[i];

@@ -901,20 +850,18 @@ static void handle_info(void)

850 continue;

851

852 if (!memcmp(header[i], "Subject", 7)) {

904 - if (keep_subject)

905 - sub = hdr;

906 - else {

907 - sub = cleanup_subject(hdr);

908 - cleanup_space(sub);

853 + if (!keep_subject) {

854 + cleanup_subject(hdr);

855 + cleanup_space(hdr);

856 }

910 - output_header_lines(fout, "Subject", sub);

857 + output_header_lines(fout, "Subject", hdr);

858 } else if (!memcmp(header[i], "From", 4)) {

859 handle_from(hdr);

913 - fprintf(fout, "Author: %s\n", name);

914 - fprintf(fout, "Email: %s\n", email);

860 + fprintf(fout, "Author: %s\n", name.buf);

861 + fprintf(fout, "Email: %s\n", email.buf);

862 } else {

863 cleanup_space(hdr);

917 - fprintf(fout, "%s: %s\n", header[i], hdr);

864 + fprintf(fout, "%s: %s\n", header[i], hdr->buf);

865 }

866 }

867 fprintf(fout, "\n");

@@ -941,8 +888,8 @@ static int mailinfo(FILE *in, FILE *out, int ks, const char *encoding,

888 return -1;

889 }

890

944 - p_hdr_data = xcalloc(MAX_HDR_PARSED, sizeof(char *));

945 - s_hdr_data = xcalloc(MAX_HDR_PARSED, sizeof(char *));

891 + p_hdr_data = xcalloc(MAX_HDR_PARSED, sizeof(*p_hdr_data));

892 + s_hdr_data = xcalloc(MAX_HDR_PARSED, sizeof(*s_hdr_data));

893

894 do {

895 peek = fgetc(in);

@@ -950,8 +897,8 @@ static int mailinfo(FILE *in, FILE *out, int ks, const char *encoding,

897 ungetc(peek, in);

898

899 /* process the email header */

953 - while (read_one_header_line(line, sizeof(line), fin))

954 - check_header(line, sizeof(line), p_hdr_data, 1);

900 + while (read_one_header_line(&line, fin))

901 + check_header(&line, p_hdr_data, 1);

902

903 handle_body();

904 handle_info();

git-mailinfo: use strbuf's instead of fixed buffers