aboutsummaryrefslogtreecommitdiffstats
path: root/activerecord/lib/active_record/connection_adapters
Commit message (Collapse)AuthorAgeFilesLines
...
| | * Don't leave open dangling connections in development mode. [#1335 ↵Pratik Naik2008-11-081-2/+1
| | | | | | | | | | | | state:resolved]
| | * Simplify dispatcher callbacks to eliminate unnecessary stale thread purging. ↵Nick Sieger2008-11-081-4/+1
| | | | | | | | | | | | | | | | | | [Nick Sieger, Pratik Naik] Signed-off-by: Pratik Naik <pratiknaik@gmail.com>
| | * Revert commit which breaks all the tests.Michael Koziarski2008-11-071-8/+4
| | | | | | | | | | | | | | | | | | | | | | | | This reverts commit 8adb79b9b5983cda8dbdd4ef401661fbd51d8844. Conflicts: activerecord/CHANGELOG
| * | undef abstract methods instead of raising NotImplementedError. Still need ↵Jeremy Kemper2008-11-071-4/+4
| |/ | | | | | | the definitions for rdoc though.
| * Stop logging SHOW FIELDS and SET SQL_AUTO_IS_NULL=0 for the MysqlAdapter as ↵David Heinemeier Hansson2008-11-061-4/+8
| | | | | | | | they only clutter up the log and offer no value [DHH]
| * Make sure ActiveRecord::Base.connected? doesn't raise an exception for ↵Pratik Naik2008-11-061-1/+2
| | | | | | | | defined connections
| * Ensure ActiveRecord::ConnectionPool.connected? handles undefined ↵Wes Oldenbeuving2008-11-061-1/+1
| | | | | | | | | | | | connections. [#936 state:resolved] Signed-off-by: Pratik Naik <pratiknaik@gmail.com>
* | Reimplement Jeremy's PostgreSQL automatic transaction state introspection code.Hongli Lai (Phusion)2008-11-032-2/+28
| | | | | | | | | | | | - Fixed compatibility with the old 'postgres' driver which doesn't support transaction state introspection. - Added unit tests for it.
* | Make SQLite3 pass the unit tests for savepoints.Hongli Lai (Phusion)2008-11-033-0/+14
| |
* | Fix a stale typo in the PostgreSQL adapter. Fix a stale mock expection in ↵Hongli Lai (Phusion)2008-11-031-1/+1
| | | | | | | | transaction_test.
* | Revert "PostgreSQL: introduce transaction_active? rather than tracking ↵Hongli Lai (Phusion)2008-11-031-38/+0
| | | | | | | | | | | | | | | | | | | | | | | | activity ourselves" This commit conflicts with savepoint support. This reverts commit 045713ee240fff815edb5962b25d668512649478. Conflicts: activerecord/lib/active_record/connection_adapters/postgresql_adapter.rb
* | Improve documentation for DatabaseStatements#transactions and ↵Hongli Lai (Phusion)2008-11-032-6/+72
| | | | | | | | AbstractAdapter#transactional_fixtures, especially with regard to support for nested transactions.
* | Implement savepoints.Jonathan Viney2008-11-034-6/+47
|/
* It is not necessary to store QueryCache in a thread local since the cache is ↵Joshua Peek2008-10-301-27/+14
| | | | local to the connection object which is managed by the connection pool
* Use database name in query cache thread local key [#1283 state:resolved]Joshua Peek2008-10-301-3/+2
|
* Remove reference to fformat to restore support for postgres gem.Michael Koziarski2008-10-261-1/+1
|
* Fix binary data corruption bug in PostgreSQL adaptorAdam Majer2008-10-251-68/+79
| | | | | | | | | | | 1. Move the binary escape/unescape from column to the driver - we should store binary data AR just like most other adaptors 2. check to make sure we only unescape bytea data PGresult.ftype( column ) == 17 that is passed to us in escaped format PGresult.fformat( column ) == 0 Signed-off-by: Michael Koziarski <michael@koziarski.com> [#1063 state:committed]
* Add documentation for AbstractAdapter#sanitize_limit, and make its code more ↵Hongli Lai (Phusion)2008-10-091-4/+15
| | | | | | | readable. Signed-off-by: Michael Koziarski <michael@koziarski.com> [#1068 status:committed]
* Merge docrailsPratik Naik2008-10-054-24/+117
|
* made ConnectionPool#checkout more robust by trying to loot dead threads when ↵Aliaksey Kandratsenka2008-10-041-1/+5
| | | | | | | | pool is empty Signed-off-by: Michael Koziarski <michael@koziarski.com> [#1169 state:committed]
* Fix race in ConnectionPool#checkoutAliaksey Kandratsenka2008-10-041-15/+14
| | | | | | | After releasing monitor some connection(s) may appear in pool before monitor is re-aquired. When this happens we'll wait for connection which is already available. Signed-off-by: Michael Koziarski <michael@koziarski.com>
* don't quote decimal values for mysql. It doesn't make sense and breaks in ↵Aliaksey Kandratsenka2008-10-041-1/+1
| | | | | | | newer versions of mysql Signed-off-by: Michael Koziarski <michael@koziarski.com> [#1168 state:committed]
* Make sure recreate MySQL test database with the proper encoding and ↵Luca Guidi2008-10-031-2/+2
| | | | | | | collation [#1165 state:resolved] Signed-off-by: Michael Koziarski <michael@koziarski.com> [#1165 state:committed]
* Deal with MySQL's quirky handling of defaults and blob/text columnsFrederick Cheung2008-09-142-1/+10
| | | | | | [#1043 state:committed] Signed-off-by: Jeremy Kemper <jeremy@bitsweat.net>
* Merge docrailsPratik Naik2008-09-131-2/+36
|
* Support :limit on update_all so that has_many with :limit can be safely updatedTarmo Tänav2008-09-102-0/+8
| | | | Signed-off-by: Michael Koziarski <michael@koziarski.com>
* Use a more sensible resolution on the new millisecond benchmarksDavid Heinemeier Hansson2008-09-051-1/+1
|
* Changed all benchmarking reports to be in millisecondsDavid Heinemeier Hansson2008-09-051-3/+3
|
* Handle connection timeouts with a slightly nicer error message.Michael Koziarski2008-09-041-1/+1
|
* Remove flawed execute("ROLLBACK") approach; #reset! defaults to nothingNick Sieger2008-09-042-5/+1
| | | | | | Will need community help to fill out what #reset! should do for each adapter Signed-off-by: Jeremy Kemper <jeremy@bitsweat.net>
* Default connection allow_concurrency to false (for PostgreSQL)Nick Sieger2008-09-041-2/+1
| | | | Signed-off-by: Jeremy Kemper <jeremy@bitsweat.net>
* Deprecate verification_timeout and verify before resetNick Sieger2008-09-043-17/+16
| | | | Signed-off-by: Michael Koziarski <michael@koziarski.com>
* still need to track whether we're the toplevel transactionJeremy Kemper2008-09-041-3/+10
|
* PostgreSQL: introduce transaction_active? rather than tracking activity ↵Jeremy Kemper2008-09-041-0/+32
| | | | ourselves
* Minor tweak to retrieve_connection_pool -- recurse instead of loopNick Sieger2008-08-291-6/+4
|
* Remove some synchronization that's probably overkill, assuming one doesn't ↵Nick Sieger2008-08-291-10/+1
| | | | establish connections frequently
* Collapse connection pool class hierarchy; YAGNI.Nick Sieger2008-08-292-112/+51
| | | | | - Add connection checkin and checkout callbacks to adapter to allow adapter-specific customization of behavior (e.g., JRuby w/ JNDI)
* Make clear_active_connections! also return stale connections back to the poolNick Sieger2008-08-292-33/+25
| | | | - also clean up some cruft remaining from per-thread connection cache
* Remove CachedConnectionPerThread per-thread pooling mechanism in favor of a ↵Nick Sieger2008-08-292-17/+9
| | | | fixed pool with default maximum of 5 connections
* Deprecate allow_concurrency and make it have no effectNick Sieger2008-08-293-40/+19
|
* Add connection reset and verification upon each connection checkoutNick Sieger2008-08-293-4/+27
|
* Fix typo: was using brackets instead of parens. Must need more sleep.Nick Sieger2008-08-291-1/+1
|
* Favor existing connections over new ones if availableNick Sieger2008-08-291-3/+3
|
* More doco and class/method renames. Now have a strategy for integration with ↵Nick Sieger2008-08-291-24/+39
| | | | ActionPack.
* Fix checkin method, add a couple more testsNick Sieger2008-08-291-1/+1
|
* Nearing the finish line. Initial fixed-size connection pool implemented, ↵Nick Sieger2008-08-292-22/+120
| | | | more docs
* Clean up the code, get rid of reserve/release, add some more docsNick Sieger2008-08-291-34/+51
|
* Extract a base class for connection pools, start to flesh out ↵Nick Sieger2008-08-291-84/+105
| | | | reserve/release API
* Split connection handler into single- and multiple-thread versions.Nick Sieger2008-08-292-31/+57
|
* Connection handling methods extracted out into separate ConnectionHandler classNick Sieger2008-08-293-105/+117
| | | | - delegating methods left behind