aboutsummaryrefslogtreecommitdiff
path: root/toys/posix/tail.c
diff options
context:
space:
mode:
authorRob Landley <rob@landley.net>2012-08-25 14:25:22 -0500
committerRob Landley <rob@landley.net>2012-08-25 14:25:22 -0500
commit3a9241add947cb6d24b5de7a8927517426a78795 (patch)
treed122ab6570439cd6b17c7d73ed8d4e085e0b8a95 /toys/posix/tail.c
parent689f095bc976417bf50810fe59a3b3ac32b21105 (diff)
downloadtoybox-3a9241add947cb6d24b5de7a8927517426a78795.tar.gz
Move commands into "posix", "lsb", and "other" menus/directories.
Diffstat (limited to 'toys/posix/tail.c')
-rw-r--r--toys/posix/tail.c232
1 files changed, 232 insertions, 0 deletions
diff --git a/toys/posix/tail.c b/toys/posix/tail.c
new file mode 100644
index 00000000..8783d6ba
--- /dev/null
+++ b/toys/posix/tail.c
@@ -0,0 +1,232 @@
+/* vi: set sw=4 ts=4:
+ *
+ * tail.c - copy last lines from input to stdout.
+ *
+ * Copyright 2012 Timothy Elliott <tle@holymonkey.com>
+ *
+ * See http://www.opengroup.org/onlinepubs/009695399/utilities/tail.html
+
+USE_TAIL(NEWTOY(tail, "fc-n-", TOYFLAG_BIN))
+
+config TAIL
+ bool "tail"
+ default y
+ help
+ usage: tail [-n|c number] [-f] [file...]
+
+ Copy last lines from files to stdout. If no files listed, copy from
+ stdin. Filename "-" is a synonym for stdin.
+
+ -n output the last X lines (default 10), +X counts from start.
+ -c output the last X bytes, +X counts from start
+ -f follow file, waiting for more data to be appended
+
+config TAIL_SEEK
+ bool "tail seek support"
+ default y
+ depends on TAIL
+ help
+ This version uses lseek, which is faster on large files.
+*/
+
+#include "toys.h"
+
+DEFINE_GLOBALS(
+ long lines;
+ long bytes;
+
+ int file_no;
+)
+
+#define TT this.tail
+
+#define FLAG_n 1
+#define FLAG_c 2
+#define FLAG_f 4
+
+struct line_list {
+ struct line_list *next, *prev;
+ char *data;
+ int len;
+};
+
+static struct line_list *get_chunk(int fd, int len)
+{
+ struct line_list *line = xmalloc(sizeof(struct line_list)+len);
+
+ line->data = ((char *)line) + sizeof(struct line_list);
+ line->len = readall(fd, line->data, len);
+
+ if (line->len < 1) {
+ free(line);
+ return 0;
+ }
+
+ return line;
+}
+
+static void dump_chunk(void *ptr)
+{
+ struct line_list *list = ptr;
+ xwrite(1, list->data, list->len);
+ free(list);
+}
+
+// Reading through very large files is slow. Using lseek can speed things
+// up a lot, but isn't applicable to all input (cat | tail).
+// Note: bytes and lines are negative here.
+static int try_lseek(int fd, long bytes, long lines)
+{
+ struct line_list *list = 0, *temp;
+ int flag = 0, chunk = sizeof(toybuf);
+ ssize_t pos = lseek(fd, 0, SEEK_END);
+
+ // If lseek() doesn't work on this stream, return now.
+ if (pos<0) return 0;
+
+ // Seek to the right spot, output data from there.
+ if (bytes) {
+ if (lseek(fd, bytes, SEEK_END)<0) lseek(fd, 0, SEEK_SET);
+ xsendfile(fd, 1);
+ return 1;
+ }
+
+ // Read from end to find enough lines, then output them.
+
+ bytes = pos;
+ while (lines && pos) {
+ int offset;
+
+ // Read in next chunk from end of file
+ if (chunk>pos) chunk = pos;
+ pos -= chunk;
+ if (pos != lseek(fd, pos, SEEK_SET)) {
+ perror_msg("seek failed");
+ break;
+ }
+ if (!(temp = get_chunk(fd, chunk))) break;
+ if (list) list->next = temp;
+ list = temp;
+
+ // Count newlines in this chunk.
+ offset = list->len;
+ while (offset--) {
+ // If the last line ends with a newline, that one doesn't count.
+ if (!flag) {
+ flag++;
+
+ continue;
+ }
+
+ // Start outputting data right after newline
+ if (list->data[offset] == '\n' && !++lines) {
+ offset++;
+ list->data += offset;
+ list->len -= offset;
+
+ break;
+ }
+ }
+ }
+
+ // Output stored data
+ llist_traverse(list, dump_chunk);
+
+ // In case of -f
+ lseek(fd, bytes, SEEK_SET);
+ return 1;
+}
+
+// Called for each file listed on command line, and/or stdin
+static void do_tail(int fd, char *name)
+{
+ long bytes = TT.bytes, lines = TT.lines;
+
+ if (toys.optc > 1) {
+ if (TT.file_no++) xputc('\n');
+ xprintf("==> %s <==\n", name);
+ }
+
+ // Are we measuring from the end of the file?
+
+ if (bytes<0 || lines<0) {
+ struct line_list *list = 0, *new;
+
+ // The slow codepath is always needed, and can handle all input,
+ // so make lseek support optional.
+ if (CFG_TAIL_SEEK && try_lseek(fd, bytes, lines));
+
+ // Read data until we run out, keep a trailing buffer
+ else for (;;) {
+ int len, count;
+ char *try;
+
+ if (!(new = get_chunk(fd, sizeof(toybuf)))) break;
+ // append in order
+ dlist_add_nomalloc((struct double_list **)&list,
+ (struct double_list *)new);
+
+ // Measure new chunk, discarding extra data from buffer
+ len = new->len;
+ try = new->data;
+ for (count=0; count<len; count++) {
+ if ((toys.optflags & FLAG_c) && bytes) {
+ bytes++;
+ continue;
+ }
+
+ if (lines) {
+ if(try[count] != '\n' && count != len-1) continue;
+ if (lines<0) {
+ if (!++lines) ++lines;
+ continue;
+ }
+ }
+
+ // Time to discard data; given that bytes and lines were
+ // nonzero coming in, we can't discard too much if we're
+ // measuring right.
+ do {
+ char c = *(list->data++);
+ if (!(--list->len)) {
+ struct line_list *next = list->next;
+ list->prev->next = next;
+ list->next->prev = list->prev;
+ free(list);
+ list = next;
+ }
+ if (c == '\n') break;
+ } while (lines);
+ }
+ }
+
+ // Output/free the buffer.
+ llist_traverse(list, dump_chunk);
+
+ // Measuring from the beginning of the file.
+ } else for (;;) {
+ int len, offset = 0;
+
+ // Error while reading does not exit. Error writing does.
+ len = read(fd, toybuf, sizeof(toybuf));
+ if (len<1) break;
+ while (bytes > 1 || lines > 1) {
+ bytes--;
+ if (toybuf[offset++] == '\n') lines--;
+ if (offset >= len) break;
+ }
+ if (offset<len) xwrite(1, toybuf+offset, len-offset);
+ }
+
+ // -f support: cache name/descriptor
+}
+
+void tail_main(void)
+{
+ // if nothing specified, default -n to -10
+ if (!(toys.optflags&(FLAG_n|FLAG_c))) TT.lines = -10;
+
+ loopfiles(toys.optargs, do_tail);
+
+ // do -f stuff
+}