busybox-w32/sysklogd, branch FRP

logread: implement dumpfollow mode of operation

2015-03-22T16:37:20+00:00

This is basically a combination of the default (dump mode) and -f (follow mode). Specifying -F makes logread first dump the log buffer and then immediately start following it. function old new delta packed_usage 30412 30443 +31 logread_main 491 497 +6 Signed-off-by: Phil Sutter Signed-off-by: Denys Vlasenko

syslogd: make "reopen log file every second" logic work for multiple logs

2014-07-03T12:51:47+00:00

Move last_log_time from a single global to *each logFile_t* so that we can actually apply the logic to every log-file in multi-file configurations, rather than working only for the first file written in each 1-second interval and then leaving the others connected to possibly wrong files. Signed-off-by: Joshua Judson Rosen Signed-off-by: Denys Vlasenko

syslogd: syslogd: don't decrement log_file->size on write failures

2014-07-02T17:41:41+00:00

Even if we fail to write to a log-file, and it's not growing, it's not *shrinking* either.... Signed-off-by: Joshua Judson Rosen Signed-off-by: Denys Vlasenko

syslogd: make "-O -" log to stdout

2014-07-02T13:21:30+00:00

function old new delta packed_usage 29871 29908 +37 log_locally 404 440 +36 syslogd_main 1966 1956 -10 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 2/1 up/down: 73/-10) Total: 63 bytes Signed-off-by: Denys Vlasenko

syslogd: Unify unlink/truncate + unlock log-rotation logic

2014-06-02T01:11:40+00:00

Always unlink + reopen, rather than sometimes using ftruncate(); using a single code-path reduces the opportunity for either mistakes or duplicate code. Signed-off-by: Joshua Judson Rosen Signed-off-by: Denys Vlasenko

syslogd: remember to un-writelock log-files even when called with "-b 0"

2014-06-02T01:11:39+00:00

Signed-off-by: Joshua Judson Rosen Signed-off-by: Denys Vlasenko

syslogd: avoid spurious ftrunctate() calls for "-b 0"

2014-06-02T01:11:38+00:00

Forgetting to re-set log_file->size after truncating to zero discards log-data for the next 1 second following an oversize-induced purge, when we shouldn't necessarily throw that data away. Signed-off-by: Joshua Judson Rosen Signed-off-by: Denys Vlasenko

Use unsigned printf/scanf conversion where more appropriate

2013-11-29T15:45:45+00:00

Signed-off-by: Denys Vlasenko

logread: don't call shmdt() before exit, kernel does it for us.

2013-11-26T11:02:18+00:00

function old new delta logread_main 472 466 -6 interrupted 20 9 -11 error_exit 20 - -20 ------------------------------------------------------------------------------ (add/remove: 0/1 grow/shrink: 0/2 up/down: 0/-37) Total: -37 bytes Signed-off-by: Denys Vlasenko

logread: flush output. closes 6710

2013-11-26T10:44:27+00:00