summaryrefslogtreecommitdiffstats
path: root/system/libraries
AgeCommit message (Collapse)AuthorFilesLines
2015-05-07Issue/PR #3836Andrey Andreev1-2/+6
2015-04-30Fix #3823Andrey Andreev2-2/+2
2015-04-29Fix #3816Andrey Andreev1-1/+1
2015-04-28Removed extra adapter validation. Adapter validation already happens in the ↵Tyler Brownell1-18/+2
system Driver library.
2015-04-27Adds error-level log messages when invalid adapters are set.Tyler Brownell1-0/+8
2015-04-24Reverts previous variable renaming and removes the foreach loop in favor of ↵Tyler Brownell1-24/+23
a simple if condition.
2015-04-23Renames $_adapter to $_driver. It's what we call it everywhere else.Tyler Brownell1-14/+14
2015-04-23Cache Library Defaults FixedTyler Brownell1-15/+6
- Comments! - Updates the cache library to validate *both* adapters. - No longer attempts to set an undefined "memcached" class variable. - $key variable renamed to $driver_type (more descriptive).
2015-04-23Updates the fallback driver variable name to match the config option name.Tyler Brownell1-6/+6
2015-04-20Fix #3787Andrey Andreev1-1/+1
2015-04-14Fix #3771Andrey Andreev1-10/+4
2015-04-09Merge branch 'develop' of https://github.com/bcit-ci/CodeIgniter into developftwbzhao5-12/+10
2015-04-09Merge pull request #3753 from ftwbzhao/feature/ci/3.0Andrey Andreev1-2/+2
[ci skip] Fix Cache/Redis get_metadata() for non-existing properties
2015-04-09[fix] redis get_metadataftwbzhao1-1/+1
2015-04-09use = instead of +=ftwbzhao1-1/+1
2015-04-07[feature] check redis auth failedftwbzhao1-1/+5
2015-04-06Disallow empty FV rules ... for consistencyAndrey Andreev1-7/+5
Related: #3736
2015-04-04Matched root_path's slashes with the name being replacedKyle Gadd1-1/+1
2015-04-01Fixed typoAchraf Almouloudi1-1/+1
2015-03-30Fix #3717Andrey Andreev1-1/+1
2015-03-30Fixes pagination with relative URLMiguel González1-1/+1
When base_url is a URL based on protocol, like "//www.google.com" the double slash regex kills the first "//".
2015-03-26Fix #3703Andrey Andreev1-1/+1
2015-03-26Remove an unnecessary checkAndrey Andreev1-5/+1
#3700 #3706
2015-03-26Apply #2737 fix to CI_XmlrpcsAndrey Andreev1-13/+14
3aecedbbb017567925f76ae68b726bd22b4cc80c Also related: #3703
2015-03-26Improve Session GC for files driverAndrey Andreev1-2/+4
Close #3701
2015-03-25use isset instead of array_key_existGabriel Potkány1-1/+1
2015-03-25Fix language overriding for calendar libraryGabriel Potkány1-1/+1
2015-03-22Fixed redis cache save logicMathew White1-1/+1
It was trying to use sAdd as a check if the key was created, but that will return false if it is already present in the set.
2015-03-21[ci skip] Forbid DB session usage with cache_on enabledAndrey Andreev1-0/+4
2015-03-13[ci skip] Remove NS usage in CI_SessionAndrey Andreev1-1/+1
It was accidental
2015-03-13Close #3663Andrey Andreev1-1/+1
2015-03-12Throw an exception on invalid session driver configAndrey Andreev1-6/+2
2015-03-10Add array typehint to CI_Email::__construct()Andrey Andreev1-1/+1
2015-03-09Reduce once MB_ENABLED checkingbjjay1-9/+3
The checking is done in the compat file system/core/compat/mbstring.php
2015-03-07Housekeeping.Master Yoda1-1/+1
Corrected typo in user guide for sessions, corrected misepelled key in calendar language file, added two links & updated wording on the repo readme. Signed-off-by:Master Yoda <jim_parry@bcit.ca>
2015-03-04Fix #3627: Keep timed locks for more than 5 secondsAndrey Andreev3-49/+23
Emulated locks for Redis and Memcached now have a TTL of 300 seconds (the default HTTP request timeout value on many environments) and 30 attemps, each separated by sleep(1), are made by the blocked request to try and obtain a lock if it has been freed. Additionaly, the blocking time for MySQL's locks, which are also timed, is also set to 300 seconds.
2015-02-23[ci skip] Fix #3618Andrey Andreev1-1/+1
2015-02-22Cache: is_supportedFieah2-6/+4
1. Cache_redis: Standardize the style as other driver. 2. Cache_wincache: Also check wincache.ucenabled
2015-02-21Fix #3610Andrey Andreev1-1/+3
2015-02-19Remove 'autoinit' DB settingAndrey Andreev1-1/+1
It doesn't make sense to do a load->database() call but not connect to the database. IIRC there was more stuff in CI_DB_driver::initialize() at some point, so that was probably the reason why the setting existed in the first place. However, now it only results in users making invalid bug reports because they don't understand the feature ... Examples during just the past 2 weeks: #3571 #3601 #3607
2015-02-17Finishing PR #3596 <!DOCTYPE html>Ivan Tcholakov1-1/+2
2015-02-15add <!DOCTYPE html>Tjoosten5-5/+10
2015-02-13Fix a typo in CI_SessionAndrey Andreev1-1/+1
2015-02-11Fix undefined variable notice in Session redis, memcached driversAndrey Andreev2-2/+2
2015-02-07Add 'sess_regenerate_destroy' settingAndrey Andreev1-1/+1
2015-02-06Polish changes following PR #3573Andrey Andreev1-5/+12
2015-02-06Fix HELO localhost.localdomain violates RFC standardsJames Wade1-2/+6
2015-02-05Use session cookie name in gc checkTom Atkinson1-2/+8
2015-02-05Merge pull request #3551 from gadelat/voidsAndrey Andreev8-8/+9
Fixed inconsistent return types
2015-02-05[ci skip] Fix a bug where session IDs are not regeneratedAndrey Andreev1-2/+1
Reported via the forums: http://forum.codeigniter.com/thread-996.html