2 * Totally braindamaged mbox splitter program.
4 * It just splits a mbox into a list of files: "0001" "0002" ..
5 * so you can process them further from there.
9 #include "string-list.h"
12 static const char git_mailsplit_usage
[] =
13 "git mailsplit [-d<prec>] [-f<n>] [-b] [--keep-cr] -o<directory> [(<mbox>|<Maildir>)...]";
15 static int is_from_line(const char *line
, int len
)
19 if (len
< 20 || memcmp("From ", line
, 5))
22 colon
= line
+ len
- 2;
31 if (!isdigit(colon
[-4]) ||
32 !isdigit(colon
[-2]) ||
33 !isdigit(colon
[-1]) ||
34 !isdigit(colon
[ 1]) ||
39 if (strtol(colon
+3, NULL
, 10) <= 90)
42 /* Ok, close enough */
46 static struct strbuf buf
= STRBUF_INIT
;
49 /* Called with the first line (potentially partial)
50 * already in buf[] -- normally that should begin with
51 * the Unix "From " line. Write it into the specified
54 static int split_one(FILE *mbox
, const char *name
, int allow_bare
)
59 int is_bare
= !is_from_line(buf
.buf
, buf
.len
);
61 if (is_bare
&& !allow_bare
)
64 fd
= open(name
, O_WRONLY
| O_CREAT
| O_EXCL
, 0666);
66 die_errno("cannot open output file '%s'", name
);
67 output
= xfdopen(fd
, "w");
69 /* Copy it out, while searching for a line that begins with
70 * "From " and having something that looks like a date format.
73 if (!keep_cr
&& buf
.len
> 1 && buf
.buf
[buf
.len
-1] == '\n' &&
74 buf
.buf
[buf
.len
-2] == '\r') {
75 strbuf_setlen(&buf
, buf
.len
-2);
76 strbuf_addch(&buf
, '\n');
79 if (fwrite(buf
.buf
, 1, buf
.len
, output
) != buf
.len
)
80 die_errno("cannot write output");
82 if (strbuf_getwholeline(&buf
, mbox
, '\n')) {
87 die_errno("cannot read mbox");
89 if (!is_bare
&& is_from_line(buf
.buf
, buf
.len
))
90 break; /* done with one message */
99 fprintf(stderr
, "corrupt mailbox\n");
103 static int populate_maildir_list(struct string_list
*list
, const char *path
)
108 char *subs
[] = { "cur", "new", NULL
};
111 for (sub
= subs
; *sub
; ++sub
) {
112 snprintf(name
, sizeof(name
), "%s/%s", path
, *sub
);
113 if ((dir
= opendir(name
)) == NULL
) {
116 error("cannot opendir %s (%s)", name
, strerror(errno
));
120 while ((dent
= readdir(dir
)) != NULL
) {
121 if (dent
->d_name
[0] == '.')
123 snprintf(name
, sizeof(name
), "%s/%s", *sub
, dent
->d_name
);
124 string_list_insert(list
, name
);
133 static int maildir_filename_cmp(const char *a
, const char *b
)
136 if (isdigit(*a
) && isdigit(*b
)) {
138 na
= strtol(a
, (char **)&a
, 10);
139 nb
= strtol(b
, (char **)&b
, 10);
142 /* strtol advanced our pointers */
146 return (unsigned char)*a
- (unsigned char)*b
;
151 return (unsigned char)*a
- (unsigned char)*b
;
154 static int split_maildir(const char *maildir
, const char *dir
,
155 int nr_prec
, int skip
)
161 struct string_list list
= STRING_LIST_INIT_DUP
;
163 list
.cmp
= maildir_filename_cmp
;
165 if (populate_maildir_list(&list
, maildir
) < 0)
168 for (i
= 0; i
< list
.nr
; i
++) {
170 snprintf(file
, sizeof(file
), "%s/%s", maildir
, list
.items
[i
].string
);
171 f
= fopen(file
, "r");
173 error("cannot open mail %s (%s)", file
, strerror(errno
));
177 if (strbuf_getwholeline(&buf
, f
, '\n')) {
178 error("cannot read mail %s (%s)", file
, strerror(errno
));
182 sprintf(name
, "%s/%0*d", dir
, nr_prec
, ++skip
);
183 split_one(f
, name
, 1);
190 string_list_clear(&list
, 1);
194 static int split_mbox(const char *file
, const char *dir
, int allow_bare
,
195 int nr_prec
, int skip
)
201 FILE *f
= !strcmp(file
, "-") ? stdin
: fopen(file
, "r");
205 error("cannot open mbox %s", file
);
211 } while (isspace(peek
));
214 if (strbuf_getwholeline(&buf
, f
, '\n')) {
215 /* empty stdin is OK */
217 error("cannot read mbox %s", file
);
224 sprintf(name
, "%s/%0*d", dir
, nr_prec
, ++skip
);
225 file_done
= split_one(f
, name
, allow_bare
);
236 int cmd_mailsplit(int argc
, const char **argv
, const char *prefix
)
238 int nr
= 0, nr_prec
= 4, num
= 0;
240 const char *dir
= NULL
;
242 static const char *stdin_only
[] = { "-", NULL
};
244 for (argp
= argv
+1; *argp
; argp
++) {
245 const char *arg
= *argp
;
250 if ( arg
[1] == 'd' ) {
251 nr_prec
= strtol(arg
+2, NULL
, 10);
252 if (nr_prec
< 3 || 10 <= nr_prec
)
253 usage(git_mailsplit_usage
);
255 } else if ( arg
[1] == 'f' ) {
256 nr
= strtol(arg
+2, NULL
, 10);
257 } else if ( arg
[1] == 'h' ) {
258 usage(git_mailsplit_usage
);
259 } else if ( arg
[1] == 'b' && !arg
[2] ) {
261 } else if (!strcmp(arg
, "--keep-cr")) {
263 } else if ( arg
[1] == 'o' && arg
[2] ) {
265 } else if ( arg
[1] == '-' && !arg
[2] ) {
266 argp
++; /* -- marks end of options */
269 die("unknown option: %s", arg
);
274 /* Backwards compatibility: if no -o specified, accept
275 <mbox> <dir> or just <dir> */
276 switch (argc
- (argp
-argv
)) {
282 stdin_only
[0] = argp
[0];
287 usage(git_mailsplit_usage
);
290 /* New usage: if no more argument, parse stdin */
296 const char *arg
= *argp
++;
300 if (arg
[0] == '-' && arg
[1] == 0) {
301 ret
= split_mbox(arg
, dir
, allow_bare
, nr_prec
, nr
);
303 error("cannot split patches from stdin");
311 if (stat(arg
, &argstat
) == -1) {
312 error("cannot stat %s (%s)", arg
, strerror(errno
));
316 if (S_ISDIR(argstat
.st_mode
))
317 ret
= split_maildir(arg
, dir
, nr_prec
, nr
);
319 ret
= split_mbox(arg
, dir
, allow_bare
, nr_prec
, nr
);
322 error("cannot split patches from %s", arg
);