aboutsummaryrefslogtreecommitdiffstats
path: root/activerecord/lib
Commit message (Collapse)AuthorAgeFilesLines
...
* Refactoring: replace the mix of variables like @finder_sql, @counter_sql, ↵Jon Leighton2010-10-3011-159/+146
| | | | etc with just a single scope hash (created on initialization of the proxy). This is now used consistently across all associations. Therefore, all you have to do to ensure finding/counting etc is done correctly is implement the scope correctly.
* only do string substitution on column names once, remove intermediate data ↵Aaron Patterson2010-10-301-7/+6
| | | | structures
* columns are always stringsAaron Patterson2010-10-301-1/+1
|
* no need to to_i, sqlite does that for usAaron Patterson2010-10-301-2/+2
|
* select_all() should always return an array, so no need to testAaron Patterson2010-10-271-9/+1
|
* speeding up query cacheAaron Patterson2010-10-272-5/+4
|
* making query cache work with prepared statementsAaron Patterson2010-10-271-6/+7
|
* renaming bind_values to binds where it makes senseAaron Patterson2010-10-265-19/+19
|
* adding backwards compatibility for non-prepare statement handling driversAaron Patterson2010-10-264-2/+34
|
* caching column valuesAaron Patterson2010-10-261-4/+11
|
* setting the authorized session clears the statement cacheAaron Patterson2010-10-261-0/+1
|
* add a session authorization setter to the pg connectionAaron Patterson2010-10-261-0/+5
|
* avoiding statement cache if there are no bind valuesAaron Patterson2010-10-261-0/+11
|
* folding async_exec to execAaron Patterson2010-10-261-7/+0
|
* clear cache before resetting the connectionAaron Patterson2010-10-261-1/+1
|
* prepare the statement before we cache the keyAaron Patterson2010-10-261-2/+4
|
* bind substitution is working properlyAaron Patterson2010-10-261-7/+11
|
* bind parameters are now typecastAaron Patterson2010-10-261-1/+3
|
* basic bind parameters are workingAaron Patterson2010-10-261-1/+1
|
* initial exec() method is working in pg adapterAaron Patterson2010-10-261-0/+38
|
* monkey patching with enumerable so we can call eachAaron Patterson2010-10-261-10/+4
|
* monkey patching Mysql::Stmt, calling free on the metadataAaron Patterson2010-10-261-4/+9
|
* free the result after slurpingAaron Patterson2010-10-261-0/+1
|
* eliminating some calls to execute()Aaron Patterson2010-10-261-5/+2
|
* statements are cached, cache is cleared on reconnectAaron Patterson2010-10-261-2/+19
|
* all mysql tests are passingAaron Patterson2010-10-261-2/+16
|
* one more mysql test left!Aaron Patterson2010-10-263-4/+10
|
* mysql tests are mostly passingAaron Patterson2010-10-261-17/+65
|
* find_one uses prepared statement cacheAaron Patterson2010-10-262-1/+7
|
* clearing statements on disconnect and resetAaron Patterson2010-10-261-0/+6
|
* type casting bound value based on column associated with valueAaron Patterson2010-10-261-1/+3
|
* stop calling execute directly on the database connectionAaron Patterson2010-10-261-2/+2
|
* removing another executeAaron Patterson2010-10-261-1/+1
|
* removing more execute callsAaron Patterson2010-10-261-2/+2
|
* removing more calls to executeAaron Patterson2010-10-261-2/+2
|
* removing some uses of executeAaron Patterson2010-10-261-4/+2
|
* exec returns an AR::ResultAaron Patterson2010-10-266-8/+61
|
* adding bind_values to relationsAaron Patterson2010-10-262-2/+9
|
* adding bind value substitutionAaron Patterson2010-10-261-0/+6
|
* adding the abstract methodAaron Patterson2010-10-261-0/+6
|
* Revert "removing unused variable"Aaron Patterson2010-10-261-1/+1
| | | | This reverts commit 3e4ede81d6b1ff3ea89eca33715a5a0d2f524dcf.
* refactor to remove `through_records`Aaron Patterson2010-10-231-10/+7
|
* removing dead codeAaron Patterson2010-10-231-2/+0
|
* removing unused variableAaron Patterson2010-10-231-1/+1
|
* if it responds to :usec, it should also "act like" a timeAaron Patterson2010-10-221-1/+1
|
* removing space errorsAaron Patterson2010-10-221-45/+45
|
* has_one maintains the association with separate after_create/after_updateXavier Noria2010-10-221-1/+10
| | | | | This way parent models can get their own after_create and after_update callbacks fired after has_one has done its job.
* use grep instead of select + is_a?Aaron Patterson2010-10-201-2/+2
|
* use a map and flatten to avoid << calls on arrayAaron Patterson2010-10-201-7/+5
|
* only call `column_methods_hash` once, use array math for faster test of ↵Aaron Patterson2010-10-201-5/+4
| | | | existence