Commit Graph

42 Commits

Author SHA1 Message Date
Brian Somers
22da50cfc8 Change the behaviour of -F slightly; it now persists (forever) in
trying to open files rather than giving up when it encounters an
error.  ENOENT errors are not reported.

As a result, files that are moved away then recreated are not at
risk of being 'lost' to tail.  Files that are recreated and
temporarily have unreadable permissions will be shown when they
are fixed.

This behaviour is consistent with the GNU version of tail but
without the verbiage that goes with the GNU version.

This change also fixes error messages accompanying -f and -F.
They no longer report problems with (null)!

MFC after:	3 weeks
2009-06-05 09:08:53 +00:00
Marcel Moolenaar
d5d2cea180 Better handle the -F case:
o  When stat(2) fails (i.e. the file has been moved) there's no new
   file with the same name yet, so keep showing the file that's open.
   This yields the same behaviour as -f, for which we don't stat(2).
o  When a new file with the same name has been created (i.e stat(2)
   succeeds but the inode or device numbers differ from the opened
   file), show any new lines in the opened file (i.e. the old or
   rotated file) before reopening the new file.

These changes fix the observed behaviour that tail(1) doesn't show
the very last lines of the rotated (log) files.

PR: bin/101979
Tested by: Jos Backus <jos@catnook.com>
MFC after: 2 months
2006-10-21 18:05:06 +00:00
Florent Thoumie
aa1d7ce72f Add a -q option to suppress header lines when multiple files are specified.
Approved by:	cperciva (mentor)
MFC after:	1 week
2006-06-29 22:07:49 +00:00
Paul Saab
13829828e7 Unbreak tail -f on non-local filesystems. 2005-08-26 08:15:57 +00:00
Xin LI
76628ce773 Save last displayed file by having the "last" pointer global and save it
during we show the first file's tail.  Instead of:

	tarsier% tail -f 1 2
	==> 1 <==
	foo
	bar

	==> 2 <==
	bar
	foo

	==> 2 <==
	bar2
	foo2

Now with this change, we have:

	tarsier% tail -f 1 2
	==> 1 <==
	foo
	bar

	==> 2 <==
	bar
	foo
	bar2
	foo2

While I'm there, move a comment to where it should belong to.  Also,
const'ify the "last" static because we will never need to change the
contents it points to.

MFC After:	1 week
2005-02-04 14:01:06 +00:00
Brian Somers
9e777bc6e2 Don't reprint file names unnecessarily.
PR:		75028
Submitted by:	mteterin at 250-217 dot customer dot cloud9 dot net
MFC after:	7 days
2005-01-12 04:06:31 +00:00
David Malone
a2fb00d93f Some variables became unused or global in the last change to this
file. Remove the now redundant declarations.

Add declarations for the new show and set_events functions and make
them static.
2005-01-10 20:16:08 +00:00
Paul Richards
15a55f7926 Add support for following more than one file i.e.
tail -f file1 file2
2004-11-04 19:18:19 +00:00
Paul Richards
4bba8e595e Convert to ANSI style function definitions. 2004-11-03 15:23:11 +00:00
David Greenman
7b51f66476 Removed check of st_rdev changing in the -F support. st_rdev for regular
files is usually the first direct block pointer. Since FreeBSD does
automatic block reallocation to reduce filesystem fragmentation, the
file being tailed can be relocated to different blocks 'on-the-fly',
making the check for st_rdev unreliable. The result of this bug is
tail -F pseudo-randomnly thinking the file was rotated when it wasn't,
and as a result, spews out the entire file trying to catch up.

MFC after:	3 days
2003-08-14 11:02:03 +00:00
Tim J. Robbins
f33ae51c77 When doing tail -F, return if the freopen() of the file fails instead of
trying to use a null file pointer.
2002-12-15 11:26:57 +00:00
Murray Stokely
b77b9b9a90 GCC 3.1 cleanup - add a break after default: at the end of a switch
statement.
2002-04-08 08:34:00 +00:00
Warner Losh
3f330d7d1a remove __P 2002-03-22 01:42:45 +00:00
Mark Murray
814e3a92a9 WARNS=2 fixes, use __FBSDID(). 2001-12-12 00:01:16 +00:00
Ian Dowse
aee1bac713 Since kqueue support was added to tail, the "-F" option (check for
renames/rotations) only detected cases where the file itself was
moved or deleted. If part of the path to the file (or a symlink
in the path) was changed instead, tail would not notice.

Fix this by ensuring that we stat the path at least once every
second in the -F case to check for changes. We still use kqueue
when possible to inform us quickly when the file has changed.

PR:		bin/24955
Submitted by:	Maxim Konovalov <maxim@macomnet.ru>
MFC after:	1 week
2001-11-25 18:03:28 +00:00
Andrey A. Chernov
bd9dc97512 File positions are off_t nowdays, not long, so:
long -> off_t
strtol -> strtoll
fseek -> fseeko

NOTE: that fseek not works for >long offsets files per POSIX:

[EOVERFLOW] For fseek( ), the resulting file offset would be a value which
cannot be represented correctly in an object of type long.
2001-09-01 22:22:45 +00:00
Jonathan Lemon
a8fb4e2810 o If the file being tailed is moved back to the original location,
resume using kqueue to monitor it, if possible.
  o auto-clear the read filter on EOF; this is needed when tailing a FIFO.
2001-06-06 17:47:13 +00:00
Andrey A. Chernov
a74da62e70 Back out my fseeko -> fseek(END) change - we need to position on what we
displayed last, not to the end of file
2001-03-28 13:10:17 +00:00
Andrey A. Chernov
462da59fb1 Restore part of my fix spammed in v1.23:
fseeko(file_size, SEEK_SET) -> fseek(0L, SEEK_END)
1) File may grows between operations, so fseeko to file_size may miss
2) 0L, SEEK_END is the same code using in tail in all other places
2001-03-27 23:24:25 +00:00
David Malone
726098d35e Fix tail to work on files bigger than 2GB.
PR:		14786
Reviewed by:	iedowse
2001-03-27 20:37:34 +00:00
Andrey A. Chernov
6bea9ab43b Add comment explaining why size not passed directly to mmap, i.e. not relay
on its check.
KNF multi-line comments
(inspired by bde)
2001-03-27 15:53:48 +00:00
Andrey A. Chernov
1410c7cc9a Use SSIZE_MAX instead of INT_MAX, as kernel does ssize_t check
Better explanation comment of FIXME section
2001-03-27 01:16:44 +00:00
Andrey A. Chernov
ad859a471d Treat mmap() error as fatal too, i.e. do exit(1) instead of return 2001-03-26 19:36:27 +00:00
Andrey A. Chernov
4f2554b7b7 rlines() checks:
1) really check for size overflow by checking negative value.
2) since mmap() not support files over INT_MAX size, add check for it
until either mmap() will be fixed or tail will be rewritted to handle
large files alternatively.
3) replace fseek(... file_size, SEEK_SET) with fseek(... 0L, SEEK_END)
to avoid off_t -> long cast
4) Use exit() if file is too big instead of warning and wrong logic
afterwards.
2001-03-26 19:29:49 +00:00
Jonathan Lemon
6029e6933c Use EV_SET() macro to initialize kevent structure. 2001-02-28 16:16:31 +00:00
Jeroen Ruigrok van der Werven
48a1ef2284 Remove register keyword usage with prejudice.
Modern compilers are smarter when it comes to allocating register
usage.
2000-12-03 17:05:45 +00:00
Jeroen Ruigrok van der Werven
247e7cb1d3 Properly constify some static arrays.
Also #if out some sccsid's and add rcsid's.
2000-12-02 19:10:12 +00:00
Jonathan Lemon
727a4ed619 Convert calls to match new kqueue API. 2000-07-18 19:38:38 +00:00
Jonathan Lemon
ea9bd2df1d If the kevent() registration fails (probably due to the underlying
filesystem not being kq-aware), then fall back to using sleep.  This
allows tail to work with NFS filesystems again without chewing up CPU time.

When given the -F flag, resort to sleep/stat after the file was moved
or deleted.  This allows a window where the file being tailed does not
exist at all, which is typically the case during log rotation.  Switch
back to using kq (if possible) after the file is reopened.
2000-06-22 18:46:03 +00:00
Jonathan Lemon
32462e82fb Change <event.h> --> <sys/event.h> 2000-04-18 16:56:55 +00:00
Jonathan Lemon
b446630f2f Change tail to use kqueue/kevent to obtain a notification when
the file changes (when doing tail -{f|F}).
2000-04-16 19:04:49 +00:00
Peter Wemm
eb1c943900 Add a new -F flag which is a superset of -f. It will cause tail to
stat() the file being followed and do a close/reopen if the file has been
renamed and/or rotated.  This is damn useful for leaving running on files
in /var/log when newsyslog(8) rotates them.
1998-04-06 16:13:49 +00:00
Andrey A. Chernov
156b574b56 Change select(0, NULL, ...) to usleep() 1997-12-12 11:47:28 +00:00
Poul-Henning Kamp
ee7b3b0639 Fix a misleading comment. Rename a variable to make more sense.
Pointed out by by:	 bde
1997-10-11 07:32:43 +00:00
Poul-Henning Kamp
97bc6e857c Fix a comment.
Submitted by:		Mikael Karpberg <karpen@ocean.campus.luth.se>
1997-09-15 08:20:27 +00:00
Poul-Henning Kamp
a3dd501081 In these days, waiting one full second for more to appear is far too long.
Let's try 250ms.
1997-09-14 19:02:13 +00:00
Alexander Langer
8abdc2eb40 Sweep through the tree fixing mmap() usage:
- Use MAP_FAILED instead of the constant -1 to indicate
    failure (required by POSIX).
  - Removed flag arguments of '0' (required by POSIX).
  - Fixed code which expected an error return of 0.
  - Fixed code which thought any address with the high bit set
    was an error.
  - Check for failure where no checks were present.

Discussed with:	bde
1997-01-16 21:58:40 +00:00
Adam David
44cf272f78 complete err() changes, actually use warn() in most instances 1996-08-26 20:37:48 +00:00
Peter Wemm
c3fd5728c2 Dont forget to #include <err.h> 1996-08-25 21:14:19 +00:00
Peter Wemm
0a6f6077af Fix some bogons:
- timeval in select loop was depending on not having the remaining time
  returned from select(), causing a busy spin on an implementation that
  does implement it.
- the err() usage was pretty bogus, some of the error messages had
  strerror attached manually and then reattached by err().
1996-08-25 21:03:50 +00:00
Adam David
6439f56e80 one file somehow slipped by me in the previous commit
tidy up
1996-07-30 15:44:30 +00:00
Rodney W. Grimes
9b50d90275 BSD 4.4 Lite Usr.bin Sources 1994-05-27 12:33:43 +00:00