| [9] | 1 | /* split - split a file Author: Michiel Huisjes */
|
|---|
| 2 |
|
|---|
| 3 | #include <sys/types.h>
|
|---|
| 4 | #include <fcntl.h>
|
|---|
| 5 | #include <stdlib.h>
|
|---|
| 6 | #include <string.h>
|
|---|
| 7 | #include <unistd.h>
|
|---|
| 8 | #include <minix/minlib.h>
|
|---|
| 9 |
|
|---|
| 10 | #define CHUNK_SIZE 1024
|
|---|
| 11 |
|
|---|
| 12 | int cut_line = 1000;
|
|---|
| 13 | int infile;
|
|---|
| 14 | char out_file[100];
|
|---|
| 15 | char *suffix;
|
|---|
| 16 |
|
|---|
| 17 | _PROTOTYPE(int main, (int argc, char **argv));
|
|---|
| 18 | _PROTOTYPE(void split, (void));
|
|---|
| 19 | _PROTOTYPE(int newfile, (void));
|
|---|
| 20 | _PROTOTYPE(void usage, (void));
|
|---|
| 21 | _PROTOTYPE(void quit, (void));
|
|---|
| 22 |
|
|---|
| 23 | int main(argc, argv)
|
|---|
| 24 | int argc;
|
|---|
| 25 | char **argv;
|
|---|
| 26 | {
|
|---|
| 27 | unsigned short i;
|
|---|
| 28 |
|
|---|
| 29 | out_file[0] = 'x';
|
|---|
| 30 | infile = -1;
|
|---|
| 31 |
|
|---|
| 32 | if (argc > 4) usage();
|
|---|
| 33 | for (i = 1; i < argc; i++) {
|
|---|
| 34 | if (argv[i][0] == '-') {
|
|---|
| 35 | if (argv[i][1] >= '0' && argv[i][1] <= '9'
|
|---|
| 36 | && cut_line == 1000)
|
|---|
| 37 | cut_line = atoi(argv[i]);
|
|---|
| 38 | else if (argv[i][1] == '\0' && infile == -1)
|
|---|
| 39 | infile = 0;
|
|---|
| 40 | else
|
|---|
| 41 | usage();
|
|---|
| 42 | } else if (infile == -1) {
|
|---|
| 43 | if ((infile = open(argv[i], O_RDONLY)) < 0) {
|
|---|
| 44 | std_err("Cannot open input file.\n");
|
|---|
| 45 | exit(1);
|
|---|
| 46 | }
|
|---|
| 47 | } else
|
|---|
| 48 | strcpy(out_file, argv[i]);
|
|---|
| 49 | }
|
|---|
| 50 | if (infile == -1) infile = 0;
|
|---|
| 51 | strcat(out_file, "aa");
|
|---|
| 52 | for (suffix = out_file; *suffix; suffix++);
|
|---|
| 53 | suffix--;
|
|---|
| 54 |
|
|---|
| 55 | /* Appendix now points to last `a' of "aa". We have to decrement it by one */
|
|---|
| 56 | *suffix = 'a' - 1;
|
|---|
| 57 | split();
|
|---|
| 58 | return(0);
|
|---|
| 59 | }
|
|---|
| 60 |
|
|---|
| 61 | void split()
|
|---|
| 62 | {
|
|---|
| 63 | char buf[CHUNK_SIZE];
|
|---|
| 64 | register char *index, *base;
|
|---|
| 65 | register int n;
|
|---|
| 66 | int fd;
|
|---|
| 67 | long lines = 0L;
|
|---|
| 68 |
|
|---|
| 69 | fd = -1;
|
|---|
| 70 | while ((n = read(infile, buf, CHUNK_SIZE)) > 0) {
|
|---|
| 71 | base = index = buf;
|
|---|
| 72 | while (--n >= 0) {
|
|---|
| 73 | if (*index++ == '\n')
|
|---|
| 74 | if (++lines % cut_line == 0) {
|
|---|
| 75 | if (fd == -1) fd = newfile();
|
|---|
| 76 | if (write(fd, base, (int) (index - base)) != (int) (index - base))
|
|---|
| 77 | quit();
|
|---|
| 78 | base = index;
|
|---|
| 79 | close(fd);
|
|---|
| 80 | fd = -1;
|
|---|
| 81 | }
|
|---|
| 82 | }
|
|---|
| 83 | if (index == base) continue;
|
|---|
| 84 | if (fd == -1) fd = newfile();
|
|---|
| 85 | if (write(fd, base, (int) (index - base)) != (int) (index - base))
|
|---|
| 86 | quit();
|
|---|
| 87 | }
|
|---|
| 88 | }
|
|---|
| 89 |
|
|---|
| 90 | int newfile()
|
|---|
| 91 | {
|
|---|
| 92 | int fd;
|
|---|
| 93 |
|
|---|
| 94 | if (++*suffix > 'z') { /* Increment letter */
|
|---|
| 95 | *suffix = 'a'; /* Reset last letter */
|
|---|
| 96 | ++*(suffix - 1); /* Previous letter must be incremented */
|
|---|
| 97 | /* E.g. was `filename.az' */
|
|---|
| 98 | /* Now `filename.ba' */
|
|---|
| 99 | }
|
|---|
| 100 | if ((fd = creat(out_file, 0644)) < 0) {
|
|---|
| 101 | std_err("Cannot create new file.\n");
|
|---|
| 102 | exit(2);
|
|---|
| 103 | }
|
|---|
| 104 | return fd;
|
|---|
| 105 | }
|
|---|
| 106 |
|
|---|
| 107 | void usage()
|
|---|
| 108 | {
|
|---|
| 109 | std_err("Usage: split [-n] [file [name]].\n");
|
|---|
| 110 | exit(1);
|
|---|
| 111 | }
|
|---|
| 112 |
|
|---|
| 113 | void quit()
|
|---|
| 114 | {
|
|---|
| 115 | std_err("split: write error\n");
|
|---|
| 116 | exit(1);
|
|---|
| 117 | }
|
|---|
