Hello all.
A few days ago I needed to reverse a few files line-by-line (i.e.,
make first line become last and vice versa). There was nothing to do
that in base, and rev(1) utility only may reverse characters in lines.
So I decided to improve rev(1)... here it is.
While digging there I realized that rev(1) didn't respect the case
when input is not newline-terminated, and always append '\n'. I preserved
this behaviour for now, but maybe this should be fixed?
Thanks for your time!
--
WBR, Vadim Zhukov
Index: rev.1
===================================================================
RCS file: /cvs/src/usr.bin/rev/rev.1,v
retrieving revision 1.7
diff -u -p -r1.7 rev.1
--- rev.1 16 Aug 2009 09:41:08 -0000 1.7
+++ rev.1 29 Dec 2010 00:35:57 -0000
@@ -38,12 +38,42 @@
.Nd reverse lines of a file
.Sh SYNOPSIS
.Nm rev
+.Op Fl cms
.Op Ar
.Sh DESCRIPTION
The
.Nm rev
utility copies the specified files to the standard output, reversing the
-order of characters in every line.
+order of characters in every line or the order of lines in the all input files.
If no files are specified, the standard input is read.
+.Pp
+The options are as follows:
+.Bl -tag -width Ds
+.It Fl c
+Reverse order of characters in every line.
+This is default.
+.It Fl m
+Reverse order of lines.
+Lines are kept in memory.
+This method is fast but is obviously limited by system and user memory limits.
+.It Fl s
+Reverse order of lines.
+Lines are kept in temporary file.
+This method is slower than
+.Fl m
+but is suitable for large input files (if you have enough space on your drive,
+of course).
+.El
+.Pp
+Only last option specified is used.
+.Sh ENVIRONMENT
+.Bl -tag -width Fl
+.It Ev TMPDIR
+Path in which to store temporary file
+.Fl ( s
+mode only).
+.El
+.Sh EXIT STATUS
+.Ex -std rev
.Sh SEE ALSO
.Xr cat 1
Index: rev.c
===================================================================
RCS file: /cvs/src/usr.bin/rev/rev.c,v
retrieving revision 1.10
diff -u -p -r1.10 rev.c
--- rev.c 27 Oct 2009 23:59:42 -0000 1.10
+++ rev.c 29 Dec 2010 00:35:57 -0000
@@ -2,6 +2,7 @@
/* $NetBSD: rev.c,v 1.5 1995/09/28 08:49:40 tls Exp $ */
/*-
+ * Copyright (c) 2010 Vadim Zhukov <persg...@gmail.com>
* Copyright (c) 1987, 1992, 1993
* The Regents of the University of California. All rights reserved.
*
@@ -31,27 +32,66 @@
*/
#include <sys/types.h>
+#include <sys/queue.h>
#include <err.h>
#include <errno.h>
+#include <stddef.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
-void usage(void);
+static void usage(void);
+static int rev_line_chars(const char *p, size_t len);
+static int mem_file_line(const char *p, size_t len);
+static int swap_file_line(const char *p, size_t len);
+static void print_mem(void);
+static void print_swap(void);
+
+#define fatal(msg...) err(EXIT_FAILURE, msg)
+
+size_t max_line_length_seen;
+
+/* Non-swap containers */
+SLIST_HEAD(, line) lines = SLIST_HEAD_INITIALIZER(lines);
+struct line {
+ SLIST_ENTRY(line) next;
+ char text[];
+};
+
+/* Swap containers */
+FILE *sfp;
+
int
main(int argc, char *argv[])
{
- char *filename, *p, *t;
+ char *filename, *p;
FILE *fp;
size_t len;
- int ch, rval;
+ int ch, rval, frval, ffailed;
+
+ enum {
+ RM_CHARS,
+ RM_LINES_MEM,
+ RM_LINES_SWAP
+ } rev_mode = RM_CHARS;
- while ((ch = getopt(argc, argv, "")) != -1)
+ while ((ch = getopt(argc, argv, "cms")) != -1)
switch(ch) {
- case '?':
+ case 'c':
+ rev_mode = RM_CHARS;
+ break;
+
+ case 'm':
+ rev_mode = RM_LINES_MEM;
+ break;
+
+ case 's':
+ rev_mode = RM_LINES_SWAP;
+ break;
+
default:
usage();
}
@@ -59,32 +99,66 @@ main(int argc, char *argv[])
argc -= optind;
argv += optind;
+ if (rev_mode == RM_LINES_SWAP)
+ if ((sfp = tmpfile()) == NULL)
+ fatal("creating swap file");
+
fp = stdin;
filename = "stdin";
rval = 0;
do {
if (*argv) {
if ((fp = fopen(*argv, "r")) == NULL) {
- warn("%s", *argv);
+ warn("fopen: %s", *argv);
rval = 1;
++argv;
continue;
}
filename = *argv++;
}
+
+ ffailed = 0;
while ((p = fgetln(fp, &len)) != NULL) {
if (p[len - 1] == '\n')
--len;
- for (t = p + len - 1; t >= p; --t)
- putchar(*t);
- putchar('\n');
- }
- if (ferror(fp)) {
- warn("%s", filename);
- rval = 1;
+
+ switch(rev_mode) {
+ case RM_CHARS:
+ frval = rev_line_chars(p, len);
+ break;
+
+ case RM_LINES_MEM:
+ frval = mem_file_line(p, len);
+ break;
+
+ case RM_LINES_SWAP:
+ frval = swap_file_line(p, len);
+ break;
+ }
+
+ if (frval == -1 && !ffailed) {
+ warn("reading: %s", filename);
+ rval = EXIT_FAILURE;
+ ffailed = 1;
+ }
}
+
(void)fclose(fp);
} while(*argv);
+
+ switch (rev_mode) {
+ case RM_CHARS: /* please compiler */
+ break;
+
+ case RM_LINES_MEM:
+ print_mem();
+ break;
+
+ case RM_LINES_SWAP:
+ print_swap();
+ break;
+ }
+
exit(rval);
}
@@ -93,6 +167,106 @@ usage(void)
{
extern char *__progname;
- (void)fprintf(stderr, "usage: %s [file ...]\n", __progname);
+ (void)fprintf(stderr, "usage: %s [-cms] [file ...]\n",
+ __progname);
exit(1);
+}
+
+int
+rev_line_chars(const char *p, size_t len) {
+ const char *t;
+
+ for (t = p + len - 1; t >= p; --t)
+ if (putchar(*t) == EOF)
+ return -1;
+ return (putchar('\n') == EOF ? -1 : 0);
+}
+
+int
+mem_file_line(const char *p, size_t len) {
+ struct line *l;
+
+ l = calloc(1, sizeof(struct line) + len + 1);
+ if (l == NULL)
+ return -1;
+ strlcpy(l->text, p, len + 1);
+ SLIST_INSERT_HEAD(&lines, l, next);
+ return 0;
+}
+
+void
+print_mem(void) {
+ struct line *l;
+
+ while (!SLIST_EMPTY(&lines)) {
+ l = SLIST_FIRST(&lines);
+ if (puts(l->text) == -1)
+ fatal("print_mem: puts");
+ SLIST_REMOVE_HEAD(&lines, next);
+ free(l);
+ }
+}
+
+int
+swap_file_line(const char *p, size_t len) {
+ if (len > max_line_length_seen)
+ max_line_length_seen = len;
+ if (len > 0 && fwrite(p, len, 1, sfp) < 1)
+ return -1;
+ if (fwrite(&len, sizeof(size_t), 1, sfp) < 1)
+ return -1;
+ return 0;
+}
+
+void
+print_swap(void) {
+ off_t off, pos;
+ size_t len;
+ char *buf;
+ const char *errstr = NULL;
+
+ buf = malloc(max_line_length_seen + 1);
+ if (buf == NULL)
+ fatal("print_swap: malloc");
+
+ len = 0;
+ while ((pos = ftello(sfp)) > (off_t)len) {
+ /* Read length of saved line */
+ off = -(off_t)(sizeof(size_t) + len);
+ if (fseeko(sfp, off, SEEK_CUR) == -1) {
+ errstr = "fseeko#1";
+ goto err;
+ }
+ if (fread(&len, sizeof(size_t), 1, sfp) < 1) {
+ errstr = "fread#1";
+ goto err;
+ }
+
+ if (len > max_line_length_seen) { /* paranoid check */
+ errstr = "len > max_line_length_seen";
+ goto err;
+ }
+
+ /* Read line itself */
+ off = -(off_t)(sizeof(size_t) + len);
+ if (fseeko(sfp, off, SEEK_CUR) == -1) {
+ errstr = "fseeko#2";
+ goto err;
+ }
+ if (len > 0 && fread(buf, len, 1, sfp) < 1) {
+ errstr = "fread#2";
+ goto err;
+ }
+ buf[len] = '\0';
+
+ if (puts(buf) == EOF)
+ fatal("print_swap: puts");
+ }
+ free(buf);
+ fclose(sfp);
+ if (pos == (off_t)len)
+ return;
+
+err:
+ fatal("print_swap: swap file damaged before %s", errstr);
}
