* '_tent/cache-config-read' of https://github.com/sebres/fail2ban:
code review, change log entries added;
reset share/cache storage (if we use 'reload' in client with interactive mode)
normalize tabs/spaces in docstrings;
cache-config-read-v2 merged; logging normalized, set log level for loading (read or use shared) file(s) to INFO; prevent to read some files twice by read inside "_getIncludes" and by "read" self (occurred by only one file);
code review; more stable config sharing, configurator always shares it config readers now;
code review: use the same code (corresponding test cases - with sharing on and without it);
rewritten caching resp. sharing of ConfigReader and SafeConfigParserWithIncludes (v.2, first and second level cache, without fingerprinting etc.);
code review
ConfigReader/ConfigWrapper renamed as suggested from @yarikoptic; + code clarifying (suggested also);
Partially merge remote-tracking from 'sebres:cache-config-read-820': test cases extended, configurator.py adapted for test case.
ENH: keep spitting out logging to the screen in LogCaptureTestCases if HEAVYDEBUG
test case for check the read of config files will be cached;
more precise by test
ConfigWrapper class introduced: sharing of the same ConfigReader object between JailsReader and JailReader (don't read jail config each jail); sharing of the same DefinitionInitConfigReader (ActionReader, FilterReader) between all jails using that; cache of read a config files was optimized; test case extended for all types of config readers;
config cache optimized - prevent to read the same config file inside different resources multiple times; test case: read jail file only once;
test case for check the read of config files will be cached;
caching of read config files, to make start of fail2ban faster, see issue #820
* origin/master:
RF: moving logwatch setup/sample logs under files/logwatch
ENH: print rebans stats even if no "Failures" are logged, and reduce indentation in output
ENH: untabified and reindented entire script for sane formatting (no functional changes)
BF: logwatch -- fixing up regex for 'already banned'
Sample logfiles to test logwatch services script
Adjusting fail2ban logwatch script to match lines from 0.9 as well
logging normalized, set log level for loading (read or use shared) file(s) to INFO;
prevent to read some files twice by read inside "_getIncludes" and by "read" self (occurred by only one file);
sharing of the same DefinitionInitConfigReader (ActionReader, FilterReader) between all jails using that;
cache of read a config files was optimized;
test case extended for all types of config readers;
Merge remote-tracking branch 'sebres:cache-config-read-820' into ban-time-incr:
config cache optimized - prevent to read the same config file inside different resources multiple times;
test case: read jail file only once;
+ optimized merge: use OrderedDict.update instead of merge in cycle;
addJail (executed before addLog) early uses a "INSERT OR REPLACE" statement to update "enabled" to 1 (and add jail the first time used at once), but this syntax in sqlite always deletes an entry (cause of constraint) and inserts it again, so because of CASCADE all log entries with this jail will be also deleted from logs table.
test cases: extended with test to check action together with database functionality (ex.: to verify lambdas in checkBan);
database: getBansMerged should work within lock, using reentrant lock (cause call of getBans inside of getBansMerged);
sebres