Scans a mailbox (mbox format) and extracts all attachments with the help of munpack

mbox_attachments.c

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <errno.h>
#include <ctype.h>
#include <alloca.h>
#include <unistd.h>
#include <fcntl.h>
#include <dirent.h>
#include <sys/wait.h>
#include <sys/mman.h>
#include <sys/stat.h>

void print_usage(const char *argv0);
int proc_mbox(const char *fname);
int proc_mail(const char *mboxname, int idx, char *mail, long size);

int listonly;
int match_msgid = -1;
int dumptext;

int main(int argc, char **argv)
{
	int i;

	signal(SIGPIPE, SIG_IGN);

	for(i=1; i<argc; i++) {
		if(argv[i][0] == '-') {
			if(argv[i][2] == 0) {
				switch(argv[i][1]) {
				case 'l':
					listonly = 1;
					break;

				case 'a':
					dumptext = 1;
					break;

				case 'm':
					if(!isdigit(argv[++i][0])) {
						fprintf(stderr, "-m must be followed by the message number\n");
						return 1;
					}
					match_msgid = atoi(argv[i]);
					break;

				case 'h':
					print_usage(argv[0]);
					return 0;

				default:
					fprintf(stderr, "invalid option: %s\n", argv[i]);
					print_usage(argv[0]);
					return 1;
				}
			} else {
				fprintf(stderr, "invalid option: %s\n", argv[i]);
				print_usage(argv[0]);
				return 1;
			}

		} else {
			if(proc_mbox(argv[i]) == -1) {
				return 1;
			}
		}
	}
	return 0;
}

void print_usage(const char *argv0)
{
	printf("Usage: %s [options] <mbox> [mbox2 ...]\n", argv0);
	printf("Options:\n");
	printf(" -l: list messages only, do not extract attachments\n");
	printf(" -m <msg number>: process a single message out of the mailbox\n");
	printf(" -a: dump all parts of each mail, not only the attachments\n");
	printf(" -h: print usage information and exit\n");
}

int proc_mbox(const char *fname)
{
	int idx, fd;
	struct stat st;
	char *start, *ptr, *end, *msgend;

	const char *match = "\nFrom ";
	int match_len = strlen(match);

	if((fd = open(fname, O_RDONLY)) == -1) {
		fprintf(stderr, "failed to open mailbox: %s: %s\n", fname, strerror(errno));
		return -1;
	}
	fstat(fd, &st);

	if((start = mmap(0, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0)) == (void*)-1) {
		fprintf(stderr, "failed to map mailbox: %s: %s\n", fname, strerror(errno));
		close(fd);
		return -1;
	}
	ptr = start;
	end = start + st.st_size;

	idx = 0;
	while(end - ptr > match_len) {
		if(ptr == start || memcmp(ptr, match, match_len) == 0) {
			msgend = ptr + 1;
			while(msgend < end && memcmp(msgend, match, match_len) != 0) {
				msgend++;
			}

			if(match_msgid == -1 || match_msgid == idx) {
				if(proc_mail(fname, idx, ptr, msgend - ptr) == -1) {
					munmap(start, st.st_size);
					close(fd);
					return -1;
				}
			}
			idx++;
			ptr = msgend;
		} else {
			ptr++;
		}
	}

	munmap(start, st.st_size);
	close(fd);
	return 0;
}

void print_field(const char *s, const char *end)
{
	while(s < end && *s != '\n') {
		putchar(*s++);
	}
}

int proc_mail(const char *mboxname, int idx, char *mail, long size)
{
	int pid, pfd[2];
	long i;
	char *from = "(unknown sender)";
	char *subj = "Subject: (no subject)";
	char *dirname;
	DIR *dir;
	struct dirent *dent;

	for(i=0; i<size; i++) {
		if(memcmp(mail + i, "\nFrom: ", 7) == 0) {
			from = mail + i + 1;
		}
		if(memcmp(mail + i, "\nSubject: ", 10) == 0) {
			subj = mail + i + 1;
		}
	}

	printf("%d: ", idx);
	print_field(subj + 9, mail + size);
	printf(" - ");
	print_field(from, mail + size);
	printf("\n");

	if(listonly) return 0;

	if(pipe(pfd) == -1) {
		fprintf(stderr, "failed to create message pipe\n");
		return -1;
	}

	dirname = alloca(strlen(mboxname) + 16);
	sprintf(dirname, "%s%05d", mboxname, idx);
	if(mkdir(dirname, 0777) == -1 && errno != EEXIST) {
		fprintf(stderr, "failed to create directory: %s: %s\n", dirname, strerror(errno));
		return -1;
	}

	if(!(pid = fork())) {
		close(0);
		dup(pfd[0]);
		close(pfd[1]);
		execlp("munpack", "munpack", "-f", "-C", dirname, dumptext ? "-t" : (void*)0, (void*)0);
		fprintf(stderr, "failed to execute munpack: %s\n", strerror(errno));
		_exit(1);
	}
	close(pfd[0]);

	while(size > 0) {
		ssize_t wr = write(pfd[1], mail, size);
		if(wr < 0) {
			if(errno == EINTR) continue;
			fprintf(stderr, "failed to write to message pipe: %s\n", strerror(errno));
			kill(pid, SIGINT);
			break;
		}
		size -= wr;
		mail += wr;
	}

	close(pfd[1]);
	wait(0);

	if((dir = opendir(dirname))) {
		while((dent = readdir(dir))) {
			if(strcmp(dent->d_name, ".") != 0 && strcmp(dent->d_name, "..") != 0 &&
					strcmp(dent->d_name, "tempdesc.txt") != 0) {
				break;
			}
		}
		closedir(dir);

		if(!dent) {
			/* directory is empty, remove it */
			char *path = alloca(strlen(dirname) + 16);
			sprintf(path, "%s/tempdesc.txt", dirname);
			remove(path);
			rmdir(dirname);
		}
	}
	return 0;
}