* 'ainfo-copy' of https://github.com/kwirk/fail2ban:
TST: actions modifying aInfo test more robust
TST: Test for actions modifying (un)ban aInfo
BF: aInfo could be modified by actions, causing unexpected behaviour
* 'sebres-strptime-bug' of https://github.com/kwirk/fail2ban:
DOC: Tweak ChangeLog and THANKS
DOC: Update docs in reference to time zone related fix
TST: Fix tests due to @sebres fix and based from gh-349 reverts
strptime bug fix: if gmtoff is None we have 1 hour increment of time (through utctimetuple), compare: >>>> datetime.datetime.fromtimestamp(time.mktime(datetime.datetime.now().timetuple())).strftime("%Y-%m-%d %H:%M:%S") '2014-04-29 17:26:31' >>>> datetime.datetime.fromtimestamp(time.mktime(datetime.datetime.now().utctimetuple())).strftime("%Y-%m-%d %H:%M:%S") '2014-04-29 18:26:37'
Conflicts:
ChangeLog
This test file shows configuration information for the application, three log lines that DO match the pattern, and one log line that does NOT match the pattern (the first one).
* 'recursive-tag-fix' of https://github.com/kwirk/fail2ban:
ENH: explicitly define tags which should be escaped
DOC: ChangeLog update for recursive tag bug fix
BF: Tags not fully recursively substituted
Conflicts:
ChangeLog -- kept all as is
* 'master' of github.com:fail2ban/fail2ban:
ENH: Match non "Bye Bye" for sshd locked accounts failregex
Even stricter monit regex, now covers entire line
Tidy up filter.d/monit.conf, make regex more complete. Add ChangeLog / THANKS entry. Add test cases.
ENH: Move traceback formatter to from tests.utils to helpers
Block brute-force attempts against the Monit gui
* 'database-persistent-bans' of https://github.com/kwirk/fail2ban:
BF: bantime < 0 database should return all bans, as they are persistent
Conflicts:
ChangeLog - kept all ;)
Note: recursive check ignored for "matches", as tags would be escaped,
and hence shouldn't match "<%s>" as "<ip>" would become "\<ip\>". This
therefore maintains advantage of delayed call for {ip,jail,}matches.
Fixes gh-713
The reason is that internally it does round, so even 1.005 then would not
be equal to 1. Making it spaces==1 should be sufficient for up to 1.05
i.e. we would allow 50ms "drift"
As documented at
http://nginx.org/en/docs/http/server_names.html#miscellaneous_names "If
no server_name is defined in a server block then nginx uses the empty
name as the server name." This regex change allows us to match error
output for such a configuration.
The log line added to the tests was lifted from our logs verbatim; it
did not match without the patched regex.
Signed-off-by: Yung-Chin Oei <yungchin@yungchin.nl>
Yaroslav Halchenko