summaryrefslogtreecommitdiffstats
path: root/system/libraries
AgeCommit message (Collapse)AuthorFilesLines
2016-03-12Renamed $this->_memcache_conf to $this->_configMasterklavi1-5/+5
2016-03-12Added variables to Docblocks (memcached)Masterklavi1-2/+2
2016-03-12Added the destructor (disconnect from memcache(d))Masterklavi1-0/+21
2016-03-12Added variables to Docblocks where they are skippedMasterklavi1-4/+4
2016-03-11Merge branch '3.0-stable' into developAndrey Andreev9-86/+137
Resolved conflicts: system/core/CodeIgniter.php user_guide_src/source/changelog.rst user_guide_src/source/conf.py user_guide_src/source/installation/downloads.rst user_guide_src/source/installation/upgrading.rst
2016-03-11Fix a number of CI_Cache bugsAndrey Andreev4-16/+45
Fixes #4277 Supersedes #4474 Really fixes #4066
2016-03-11Revert an unintended change from a027a7fd0d770cec0d71e888d8b6f4aa1568ce9fAndrey Andreev1-1/+1
2016-03-10Fix a logical error from last commitAndrey Andreev1-1/+1
2016-03-10Improve ext/session error messagesAndrey Andreev4-32/+51
2016-03-07[ci skip] Fix Profiler not applying htmlspecialchars() to all inputsAndrey Andreev1-26/+18
2016-02-29[ci skip] Apply #4491 to Memcached driverAndrey Andreev1-1/+1
2016-02-29Merge pull request #4491 from roastduck/developAndrey Andreev1-1/+1
[ci skip] Clean current lock key on close() in redis session driver
2016-02-29fix typo__RD1-1/+1
2016-02-29delete lock directly -> call _release_lock()__RD1-2/+1
2016-02-28clean current lock key in redis session driverroastduck1-0/+1
set $this->_lock_key to NULL after close Signed-off-by: roastduck <beantang.tang@gmail.com>
2016-02-24Merge pull request #4480 from versalle88/developAndrey Andreev1-1/+1
Changed class_exists() calls to ignore __autoload()
2016-02-23Fixed typoversalle881-1/+1
2016-02-23Removedversalle881-1/+1
2016-02-23Changed calls to class_exists to ignore __autoload() to match other callsversalle881-1/+1
2016-02-23Changed calls to class_exists to ignore __autoload() to match other callsversalle881-1/+1
2016-02-13[ci skip] Fix Memcached replace() result code checks in CI_SessionAndrey Andreev1-2/+2
Related #3919
2016-02-10Fix a bug where CI_Session_memcached_driver doesn't write empty sessionsAndrey Andreev1-4/+7
Related: #3919
2016-02-10[ci skip] Fix Memcached session lock handling and error checking around ↵Andrey Andreev1-9/+17
replace() usage
2016-02-09Rename back a variable changed by the last PR mergeAndrey Andreev1-4/+4
Because.
2016-02-09Rename back a variable changed by the last PR mergeAndrey Andreev1-4/+4
Because.
2016-02-09Merge pull request #4342 from jspreddy/sai/form_validation_refactorAndrey Andreev1-30/+36
Abstract error message fetching in Form_validation
2016-02-09Merge pull request #4342 from jspreddy/sai/form_validation_refactorAndrey Andreev1-30/+36
Abstract error message fetching in Form_validation
2016-02-09Merge branch '3.0-stable' into developAndrey Andreev2-0/+14
Fixed conflicts: user_guide_src/source/overview/at_a_glance.rst
2016-02-09Code formatting changes, again. I am bad at this.Sai Phaninder Reddy Jonnala1-6/+5
2016-02-08code styling changes as suggested by @narfbgSai Phaninder Reddy Jonnala1-21/+15
2016-02-05Merge pull request #4424 from jonty-comp/developAndrey Andreev1-3/+16
[ci skip] Fix PHP session_write_close() warning when writing a new session to Redis
2016-02-05Fix regression on PHP7 when regenerating the session (#4362)Jonty Sewell1-1/+3
2016-02-05Setting the flag to FALSE is unnecessary since it defaults to FALSE, ↵Jonty Sewell1-10/+1
therefore this block of code can be reduced to a single statement Signed-off-by: Jonty Sewell <jontysewell@gmail.com>
2016-02-05Set the _key_exists flag to TRUE when the key does in fact exist. Set it to ↵Jonty Sewell1-2/+7
FALSE if the ID is being regenerated, and set it to TRUE once it's been written. Signed-off-by: Jonty Sewell <jontysewell@gmail.com>
2016-02-05Add a flag to determine whether the redis key currently exists, and if not, ↵Jonty Sewell1-2/+17
force creation of it at write-time Signed-off-by: Jonty Sewell <jontysewell@gmail.com>
2016-02-05Revert previous changes - fixing the source of the problem rather than ↵Jonty Sewell1-9/+3
working around it Signed-off-by: Jonty Sewell <jontysewell@gmail.com>
2016-02-04Fix #4430Andrey Andreev1-0/+6
2016-02-03If attempting to write an empty session to Redis, a key will not actually be ↵Jonty Sewell1-3/+9
created, so when the driver tries to set the expiration timeout on the key, 0 is returned, triggering a warning from session_write_close Signed-off-by: Jonty Sewell <jontysewell@gmail.com>
2016-01-30Fix #4415 and add unit tests for https://bugs.php.net/bug.php?id=51192Andrey Andreev1-0/+8
2016-01-26Merge branch '3.0-stable' into developAndrey Andreev1-7/+11
2016-01-20[ci skip] Remove a trailing space from latest PR mergeAndrey Andreev1-1/+1
2016-01-20Respect $config['cur_page'] variable for paginationjekkos1-1/+5
After upgrading to CI3 I noticed that developers are able to determine the current page counter for pagination based on * Explicit query string parameters * URI segment configuration In earlier versions a developer could still set the current page counter in the pagination lib directly which is useful if you want to use pagination with HTTP POST instead of GET. This could be done by passing $config['cur_page'] = '10'; to the pagination function for link generation. Currently this code has changed and will always try to check whether the uri segment is a valid number or not, even if the cur_page variable was passed in the associative array, and fallback to zero if it fails to validate that result. This can be easily resolved by checking whether the counter was already set with a valid number from the $config array before trying to resolve it from the uri segment. This fix give a developer more flexibility and stop CI from overwriting the externally set value with an incorrect one. Signed-off-by: jekkos <jeroen.peelaerts@gmail.com>
2016-01-20[ci skip] Remove a trailing space from latest PR mergeAndrey Andreev1-1/+1
2016-01-20Respect $config['cur_page'] variable for paginationjekkos1-1/+5
After upgrading to CI3 I noticed that developers are able to determine the current page counter for pagination based on * Explicit query string parameters * URI segment configuration In earlier versions a developer could still set the current page counter in the pagination lib directly which is useful if you want to use pagination with HTTP POST instead of GET. This could be done by passing $config['cur_page'] = '10'; to the pagination function for link generation. Currently this code has changed and will always try to check whether the uri segment is a valid number or not, even if the cur_page variable was passed in the associative array, and fallback to zero if it fails to validate that result. This can be easily resolved by checking whether the counter was already set with a valid number from the $config array before trying to resolve it from the uri segment. This fix give a developer more flexibility and stop CI from overwriting the externally set value with an incorrect one. Signed-off-by: jekkos <jeroen.peelaerts@gmail.com>
2016-01-20Fix #4391Andrey Andreev1-7/+11
2016-01-11[ci skip] Merge branch '3.0-stable' into developAndrey Andreev38-38/+38
2016-01-11[ci skip] Update ellislab.com links to https tooAndrey Andreev38-38/+38
2016-01-11Merge branch '3.0-stable' into developAndrey Andreev38-72/+72
2016-01-11[ci skip] Update codeigniter.com links to httpsAndrey Andreev38-72/+72
2016-01-11Merge branch '3.0-stable' into developAndrey Andreev38-76/+76
generated by cgit v1.2.3-24-g4f1b (git 2.32.0) at 2024-11-08 10:53:15 +0100