summaryrefslogtreecommitdiff
path: root/usr.sbin/cron
AgeCommit message (Collapse)Author
2005-11-30timezone -> time zoneJason McIntyre
2005-11-29crontab.c uses the rcsid in the crontab head comment so don't protectTodd C. Miller
it from lint. Found by cloder@
2005-11-15More typos (especially indicies -> indices), checked with jmc@Miod Vallat
2005-11-15If two or more changes are made withing a single second it is possibleTodd C. Miller
that only the first change will be loaded due to the granularity of st_mtime. Work around this by zeroing the saved mtime when we get a poke on the socket from crontab or at since that means there has been a change. Fix from Daniel Lucq.
2005-06-08Force pid file fd to be > 2 so we don't clobber it when dup'ingTodd C. Miller
fds 0-2 to /dev/null. Tested and OK aaron@
2005-03-10put umask restores in right place; millert okTheo de Raadt
2005-01-30Detect editors that move the edited file out from underneath us and warn.Todd C. Miller
Based on an idea from Han Boetes.
2005-01-30bzero s_un before use; Andrey MatveevTodd C. Miller
2005-01-30Proper cleanup on malloc failure; Andrey MatveevTodd C. Miller
2005-01-30fix some whitespaceTodd C. Miller
2004-11-04Treat EOF at prompt the same as 'n' instead of looping forever.Todd C. Miller
Noticed by mepch@. OK pedro@ and deraadt@
2004-09-16nice off_t castTheo de Raadt
2004-09-16ftruncate() with ftello() instead of ftell(); ok millertTheo de Raadt
2004-07-22Fix incorrect snprintf return value check noticed by Jarno Huuskonen.Todd C. Miller
Also make a utime() failure non-fatal since the important thing is the poke on the socket now.
2004-07-09ansiTheo de Raadt
2004-06-22Recent glue_strings change reversed a few checks; fix up conditionalsAnil Madhavapeddy
to check for overflow as others do by convention. Noticed by Andreas Kahari <ak+openbsd at freeshell.org>, millert@ ok
2004-06-22use snprintf(3) instead of home-grown glue_strings() functionAnil Madhavapeddy
millert@ ok
2004-06-17UUpdate ISC copyright year to 2004Todd C. Miller
Remove unused macros Skip_Line and MkLower Remove trailing whitespace
2004-06-06Fix typo in last commit for non-compiled code. Dmitry V. LevinTodd C. Miller
2004-06-03Document behavior if MAILTO is set to the empty string.Todd C. Miller
2004-06-03More changes from Dmitry V. Levin:Todd C. Miller
Check return values for setgid, initgroups and setuid in code we don't compile. Print the correct filename for the at job in mail sent. Add some #if DEBUGGING in cron.c's usage(). Set sunlen each time before using it in accept(). Don't send mail at all if MAILTO is set but empty.
2004-06-03Though crontab files are created mode 0600 we should accept ones thatTodd C. Miller
are mode 0400 too. From Walt Howard.
2004-05-13Pasto, change a cast from uid_t to gid_t (no real change...)Todd C. Miller
2004-05-13Set close on exec flag for cron's socket. Jarno HuuskonenTodd C. Miller
2004-05-13If we get EAGAIN reading from the cron socket check both cron andTodd C. Miller
at databases for jobs to run. Jarno Huuskonen.
2004-05-10Make the accept socket non-blocking. Should not matter since weTodd C. Miller
use select anyway but seems like a good idea since we really don't want cron to block...
2004-05-03Make socket that reads reload requests from crontab be non-blocking.Todd C. Miller
Problem found by Jarno Huuskonen.
2004-04-26isalnum() does not match '_' so check it explicitly. This allows cronTodd C. Miller
to send mail to usernames with a '_' in them. From David Gwynne.
2004-02-24remove reference to -x bit;Jason McIntyre
ok millert@
2003-07-30Use (char *)NULL instead of (char *)0 in execl and execle; Andrey MatveevTodd C. Miller
2003-06-26two missing protosTheo de Raadt
2003-06-17Sync with share/misc/license.template and add missing DARPA creditTodd C. Miller
where applicable.
2003-06-12- section reorderJason McIntyre
- macro fixes - kill whitespace at EOL - new sentence, new line
2003-06-03Don't need 'all right reserved'; brings this into line with my otherTodd C. Miller
ISC-style copyrights.
2003-06-02I'm a dork, I removed the 4th clause of the UCB license instead ofTodd C. Miller
the 3rd one.
2003-05-12feof() can only be used after you actually hit EOF so the checkTodd C. Miller
for no output from the at command was busted. Instead of using feof(), just fread() a buffer's worth and check to see if we got anything back. Closes PR 3252.
2003-04-17Change a few return values from char to int since these functionsTodd C. Miller
can return EOF. Fixes gcc warnings (and potential problems) on ppc.
2003-04-15nasty lies from hobbits, clean them out of commentsTheo de Raadt
2003-04-14Use setproctitle() instead of the CAPITALIZE_FOR_PS hack. InspiredTodd C. Miller
by similar changes in FreeBSD and NetBSD. For at jobs, include the job number in the proctitle.
2003-03-15Use strlcpy() instead of pre-checking the src len and using strcpy()Todd C. Miller
2003-03-12Move range sanity check out of get_number() and into get_range() sinceTodd C. Miller
it was causing problems with symbolic names (and get_range() is really where it belongs anyway). Also allow a range to be followed by a ',' as pointed out by mpech@
2003-03-11Catch more syntax errors that were previously ignored. I've rewrittenTodd C. Miller
get_number() to be more careful about what it accepts and to take a string of terminating characters so it knows what is valid. Also added a range check so that ranges like "60-50" where the first number is greater than the second are now flagged. Thanks to mpech@ for testing finding ways to abuse the parser :-)
2003-03-10these pages all had bad section numbers in the .Xr's;Jason McIntyre
lots of help and ok millert@
2003-03-10Log that cron started up after we detach from the controlling terminalTodd C. Miller
and dup fds to /dev/null, not before. This may fix a possible race when cron is started as the last thing in /etc/rc.
2003-03-10Change a e_username to e_cmd.Todd C. Miller
2003-03-10Back out changes to get_number() that I didn't intend to commit.Todd C. Miller
2003-03-10Do the check for missing command before the un_getchar(); otherwiseTodd C. Miller
the error messages will say line 0. This is because un_getchar('\n', file) will decrement the line count. Noticed by mpech@
2003-03-09Fix error line number reporting for syntax errors (noticed by mpech@).Todd C. Miller
This is relatively tricky due to the way EOF is overloaded. Seems to give a correct line number for all errors I threw at it.
2003-03-09Catch lines w/o EOF. Previously they were silently ignored.Todd C. Miller
2003-03-04Deref the correct passwd pointer in a Debug statement.Todd C. Miller