aboutsummaryrefslogtreecommitdiffstats
path: root/activesupport
Commit message (Collapse)AuthorAgeFilesLines
* Custom serializers and deserializers in MessageVerifier and MessageEncryptor.Willem van Bergen2011-09-154-5/+30
| | | | | By default, these classes use Marshal for serializing and deserializing messages. Unfortunately, the Marshal format is closely associated with Ruby internals and even changes between different interpreters. This makes the resulting message very hard to impossible to unserialize messages generated by these classes in other environments like node.js. This patch solves this by allowing you to set your own custom serializer and deserializer lambda functions. By default, it still uses Marshal to be backwards compatible.
* update 3.1 release date in changelogsVijay Dev2011-09-141-1/+1
|
* Fix a typo in dateSemyon Perepelitsa2011-09-121-1/+1
|
* revert the changes from c60995f3 - related to marking sub,gsub as ↵Vijay Dev2011-09-091-20/+1
| | | | unavailable to use with safe strings
* delete the CHANGELOG entry re SafeBuffer and sub/gsub, since it has been ↵Xavier Noria2011-09-081-2/+0
| | | | reverted
* Revert removing gsub and sub from safe buffer.José Valim2011-09-085-63/+45
|
* mention the issue number for future referenceXavier Noria2011-09-081-1/+1
|
* registers removal of sub and friends for safe strings in the CHANGELOGXavier Noria2011-09-081-0/+2
|
* this should have gone with the previous commitXavier Noria2011-09-082-6/+6
|
* copy-edits a couple of exception messagesXavier Noria2011-09-081-4/+4
| | | | | | | "Safe Buffer" should either be the constant with the class name, or go in lower case. I've chosen to follow the same terminology that is used in the AS core extensiong guide, "safe string", though "safe buffer" is also used elsewhere, we should pick one.
* better method documentation on disable safe string methodsDamien Mathieu2011-09-081-6/+12
|
* remove support of symbols on classify and camelizeDamien Mathieu2011-09-084-34/+52
|
* make gsub and sub unavailable in SafeBuffers - Closes #1555Damien Mathieu2011-09-082-11/+24
|
* Make sure require_tzinfo only calls Kernel#require if TZInfo isn't already ↵Tim Lucas2011-09-071-1/+1
| | | | | | | | present. This wasn't a problem when require_tzinfo was only called from TimeZone#initialize, but now it's being called for every lookup to TimeZone.[] (via lazy_zones_map) TimeZone lookup can occur when unmarshalling TimeWithZone objects, which is where I first saw the big slowdown (500ms for each Rails action that loaded from Rails.cache)
* We can't simply check the first line of the backtrace, Arun Agrawal2011-09-041-4/+6
| | | because JRuby reports the call to __send__ in the backtrace.
* Modified content in guides and commentsArun Agrawal2011-09-041-2/+2
| | | | | | for "assert /" warnings. Removed because if somebody will use this code they will get warnings!
* Merge pull request #2827 from sikachu/extractable_ordered_hashJosé Valim2011-09-033-0/+12
|\ | | | | Make `ActiveSupport::OrderedHash` extractable when using `Array#extract_options!
| * Make `ActiveSupport::OrderedHash` extractable when using ↵Prem Sichanugrist2011-09-033-0/+12
| | | | | | | | | | `Array#extract_options!` `ActiveSupport::OrderedHash` is actually a subclass of the hash, so it does make sense that it should be extractable from the array list.
* | docs formatting changesVijay Dev2011-09-013-4/+4
|/
* Revert "Ruby 1.8.7+ provides to_date/to_datetime, AS just makes them ↵Chris Griego2011-08-311-0/+22
| | | | | | | | public." to support date implementations without private to_date/to_datetime on Time (home_run). This reverts commit 080345baca1076a9788dff4803153600aec31f86. Signed-off-by: José Valim <jose.valim@gmail.com>
* Merge pull request #2723 from guilleiguaran/3-1-0-changelogsSantiago Pastorino2011-08-281-0/+12
| | | | Update changelogs for Rails 3.1.0
* Revert all the stuff to do with disallowing non-public methods for ↵Jon Leighton2011-08-256-170/+3
| | | | Module#delegate
* stop circular require warningsAaron Patterson2011-08-242-2/+0
|
* only backport the to_d method if it does not existAaron Patterson2011-08-241-2/+5
|
* dealing with some -w warnings on startupAaron Patterson2011-08-241-0/+2
|
* Fix autoload_once_paths when using Pathnames & ruby 1.9Jonathan del Strother2011-08-242-1/+20
| | | | | | | Under ruby 1.9.2 - "/var/log".starts_with?(Pathname.new("/var")) # => false so setting config.autoload_once_paths with Pathnames would not work
* `load` should also return the value from `super`Aaron Patterson2011-08-231-1/+3
|
* require needs to return true or false. thank you Ryan "zenspider" DavisAaron Patterson2011-08-231-1/+3
|
* Remove trailing whitespaceDaniel Schierbeck2011-08-231-2/+2
|
* properly escape html to avoid invalid utf8 causing XSS attacksAaron Patterson2011-08-162-1/+8
|
* Merge branch 'master' of github.com:lifo/docrailsXavier Noria2011-08-161-0/+2
|\
| * Document Hash#extract!.Sebastian Martinez2011-08-151-0/+2
| |
* | prefer ends_with? over slicingXavier Noria2011-08-161-1/+2
| |
* | Fix tzinfo require (it broke ↵Jon Leighton2011-08-161-6/+13
| | | | | | | | test_raises_when_an_invalid_timezone_is_defined_in_the_config in railties)
* | Be more lazy about creating time zone objects. Decreases startup time by ↵Jon Leighton2011-08-161-2/+13
| | | | | | | | about 10%. (#734)
* | Added Array#prepend as an alias for Array#unshift and Array#append as an ↵David Heinemeier Hansson2011-08-154-0/+20
| | | | | | | | alias for Array#<< [DHH]
* | Requiring delegate.Arun Agrawal2011-08-161-0/+1
| |
* | Removing unnecessary require, solve 'circular require considered harmful' ↵Jon Leighton2011-08-161-2/+0
| | | | | | | | warning.
* | Document Object#public_sendJon Leighton2011-08-151-1/+1
| |
* | Add a test for delegating a method ending in '=' as this is a special case.Jon Leighton2011-08-152-0/+9
| |
* | Need to include public_sendArun Agrawal2011-08-151-0/+1
| |
* | Split up the definitions in Module#delegate depending on :allow_nil, and ↵Jon Leighton2011-08-152-17/+34
| | | | | | | | don't use exceptions for flow control in the :allow_nil => true case.
* | Fix the line number in the backtrace when Module#delegate raisesJon Leighton2011-08-152-2/+14
| |
* | Just do the method call directly in Module#delegate, if we can (we cannot ↵Jon Leighton2011-08-153-4/+27
| | | | | | | | for method names ending in '='). Two reasons: 1) it's faster, see https://gist.github.com/1089783 and 2) more importantly, delegate should not allow you to accidentally call private or protected methods.
* | Backport Object#public_send to 1.8 so that we can implement Module#delegate ↵Jon Leighton2011-08-153-0/+143
| | | | | | | | such that non-public methods raise
* | Merge branch 'master' of github.com:lifo/docrailsXavier Noria2011-08-131-1/+1
|\| | | | | | | | | | | | | | | | | | | | | Conflicts: RELEASING_RAILS.rdoc actionpack/lib/sprockets/railtie.rb actionpack/test/template/sprockets_helper_test.rb activerecord/test/cases/calculations_test.rb railties/guides/source/3_1_release_notes.textile railties/guides/source/active_resource_basics.textile railties/guides/source/command_line.textile
| * Rubygems => RubyGemsSukeerthi Adiga2011-08-131-1/+1
| |
| * fixing assert_difference issues on ruby 1.8Aaron Patterson2011-08-131-6/+7
| |
| * make assert_difference error message not suckAaron Patterson2011-08-131-5/+6
| |
| * Pass options in ActiveSupport::Cache::CacheStore#read_multi through to the ↵Brian Durand2011-08-132-1/+9
| | | | | | | | delete_entry call.