summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2011-02-07Dynamically allocate inotify read bufferHEADmasterTobias Klauser1-1/+2
This fixes a sparse warning.
2011-02-07Remove duplicate const (found using sparse)Tobias Klauser1-1/+1
2011-02-07Add __noreturn attribute, update copyrightTobias Klauser2-8/+13
2011-02-07Only print error if there actually was oneTobias Klauser1-2/+4
write() returning 0 there is fine and not considered an error.
2010-02-09Add .gitattributes, thanks François MarierTobias Klauser1-0/+2
http://feeding.cloud.geek.nz/2010/02/excluding-files-from-git-archive.html Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
2009-12-17Makefile: Change to new git command formatTobias Klauser1-2/+2
2009-10-30inotail.c: Add TODO comment to handle_inotify_eventTobias Klauser1-0/+4
Reopening the watch should only be done in FOLLOW_NAME mode.
2009-10-30inotail: Support reopening rotated filesDominguez Bonini, David2-15/+42
We changed the behaviour of the "follow" option so that it detects that file has been moved or deleted it keeps monitoring for the file instead of ignoring it. This is useful when tailing syslog circular logs, where syslog periodically moves the log to another filename and recreates the log file. Without the patch inotail would issue a "file has been moved" or a "file has been deleted message" and stop tracking the file. Now it will close the file and refresh the inotify +descriptor. When a notification arrives with a CREATE or MODIFY event, it will reopen the file and set the size to zero. We are using the patched program in an embedded ARM system with satisfactory results.
2009-07-14inotail.h: Correct typoTobias Klauser1-1/+1
2009-04-23inotail: open with O_LARGEFILESage Weil1-1/+1
Opening with O_LARGEFILE avoids failing with -EFBIG on large (>2gb?) files on 32bit machines (tested on 32bit x86).
2009-04-06Large file support for inotailTobias Klauser1-0/+1
Patch by hondza <sedaj2@gmail.com> Inotail fails to fstat() large files with "Error: Could not stat file '/large/file' (Value too large for defined data type)". Following patch makes it work.
2009-02-08Implement argument parsing for --followTobias Klauser1-1/+17
2009-01-26Prepare for -F/--follow=name options (no effect yet)Tobias Klauser2-12/+25
2009-01-26Change a remaining mode_t to charTobias Klauser1-1/+1
2009-01-26inotail.c: More graceful error handling in watch_file()Tobias Klauser1-3/+8
Extracted from the patch by Joe Pelkey
2009-01-25Makefile: release target updateTobias Klauser1-11/+24
2009-01-25README: Mention POSIX/GNU compatibilityTobias Klauser1-1/+4
2009-01-25Reference head(1) in the manpage according to the original POSIX manpageTobias Klauser1-0/+1
2009-01-25Document the options in the READMETobias Klauser2-5/+25
2009-01-25Make sure that st_blksize is greater than 0Tobias Klauser2-1/+4
2009-01-25Implement remaining commandline optionsTobias Klauser2-8/+50
inotail is now 100% option compatible to GNU tail (at least to the documented options) Some options make no sense for inotail (-s, --max-unchanged-stats), others are impossible (--pid) and are thus not implemented. They just emit a warning on stderr.
2009-01-25Update copyrightTobias Klauser4-4/+5
2009-01-25Add -q, --quiet and --silent optionsTobias Klauser2-12/+22
inotail should become option compatible to GNU tail.
2009-01-25inotail.c: Macros do not set errnoTobias Klauser1-1/+1
2009-01-24inotail.c: Switch to postfix increment, close ifd on errorTobias Klauser1-10/+14
2009-01-24Move linked list structs for pipe_{bytes,lines} into these functionsTobias Klauser2-17/+11
2009-01-18Makefile: Add -Wextra to CFLAGSTobias Klauser1-1/+1
2009-01-18inotail.h: Use sys/inotify.hTobias Klauser1-1/+1
2009-01-18Merge branch 'master' of git+ssh://git.distanz.ch/home/tklauser/scm/inotailTobias Klauser5-178/+4
2008-11-14More comments and clearify a return statement.Tobias Klauser1-1/+3
2008-09-23Update changelogTobias Klauser1-1/+3
2008-09-23Merge branch 'master' of /home/tklauser/scm/inotailTobias Klauser4-44/+306
Conflicts: inotail.c
2008-09-23use sys/inotify.hAdrian Bunk4-177/+1
Some time ago glibc did not yet have inotify support, which forced several applications (including inotail) to ship their own copy of the required headers. Especially considering that this also requires updating the syscall numbers each time a new architecture wants to get supported it's the best to instead switch to #include <sys/inotify.h> now.
2008-08-31Prepare inotail 0.6Tobias Klauser2-1/+10
Presumably this version will not yet include the follow by name feature and the fixes for the file rotation issue (Debian bug #460527). Will have to find some time first to have a closer look at these two.
2008-08-31Check for file being a pipe before seekingTobias Klauser1-1/+1
Otherwise inotail -f on a named pipe (fifo) would not work. Thanks Alexander Sulfrian.
2008-05-10Resolve some merge conflicts.Tobias Klauser2-4/+3
2008-04-14inotail.c: Comment some oarts of tail_pipe_lines/tail_pipe_bytesTobias Klauser1-5/+20
2008-03-24Prevent double calls to ignore_file() in/after tail_file()Tobias Klauser1-6/+1
2008-03-01Merge branch 'pipe'Tobias Klauser2-19/+260
Conflicts: inotail.c
2008-03-01Move all #define's except PROGRAM_NAME to inotail.hTobias Klauser2-12/+11
2008-02-11inotail.c: Handle IN_IGNORED eventTobias Klauser1-0/+2
This event is e.g. emitted when we remove a watch on a file.
2008-02-11inotail.c: Check ignore flag before ignoreing the fileTobias Klauser1-2/+4
This makes ignore_file reentrant. Calling the function on a file which is already ignored shouldn't lead to n_ignored getting incremented.
2008-01-14Update copyright and bump version to pre0 (ti fix Debian bug #460527)Tobias Klauser2-3/+3
2008-01-14inotail.c: Correct format specifier for size_tTobias Klauser1-1/+1
2008-01-14inotail.c: Handle file truncation in follow modeTobias Klauser1-10/+15
2008-01-13inotail.c: Correct format for size_t (fixes warning on amd64)Tobias Klauser1-1/+1
2007-11-11Revert "inotail.c: Decrement n_units before calling tail_pipe_from_begin"pipeTobias Klauser1-2/+5
This reverts commit 21cddfba5f53ed0ad9d8f14bc1555c57cad9790e. Makes inotail -n +0 work again. Sometimes reducing code leads to unrecognised errors ;-)
2007-09-25inotail.c: typdef mode_tTobias Klauser2-3/+4
2007-09-25inotail.c: Decrement n_units before calling tail_pipe_from_beginTobias Klauser1-5/+2
Saves 2 lines of code
2007-09-19inotail.c: Check for n_units == 0 only if tailing from endTobias Klauser1-4/+9
inotail -n+0/-c+0 should be allowed
generated by cgit v1.2.3-54-g00ecf (git 2.39.0) at 2024-11-22 01:01:06 +0000