I think that 5 seconds was a bit low for our purposes.
Also enable it to be configured via env vars.
We also need to scale the number of records up/down depending on how
long we're running the benchmark for.
Conflicts:
activerecord/examples/performance.rb
This means we can more easily compare numbers, and we don't have to
specify a single N for all reports, which previously meant that some
tests were running many more/fewer iterations than necessary.
Conflicts:
Gemfile
activerecord/examples/performance.rb
This was added to all other branches, but 3-1 missed the entry.
3-0-stable: 954e262636
3-2-stable: ae2383d90a
master: 5c07be5f7fb7e6417c09751c7fb9bb98a63d9917
That change to update_attribute was considered
to be too subtle and was reverted in 30ea923
just before Rails 3 shipped. Later we introduced
update_column (Rails 3.1).
While the patched PredicateBuilder in 3.1.5 prevents a user
from specifying a table name using the `table.column` format,
it doesn't protect against the nesting of hashes changing the
table context in the next call to build_from_hash. This fix
covers this case as well.
* 3-1-stable-sec:
Strip [nil] from parameters hash. Thanks to Ben Murphy for reporting this!
predicate builder should not recurse for determining where columns. Thanks to Ben Murphy for reporting this
* 3-1-stable-sec:
Strip [nil] from parameters hash. Thanks to Ben Murphy for reporting this!
predicate builder should not recurse for determining where columns. Thanks to Ben Murphy for reporting this
This commit improves the handling of default_url_options in integration
tests by making behave closer to how a real application operates.
Specifically the following issues have been addressed:
* Options specified in routes.rb are used (fixes#546)
* Options specified in controllers are used
* Request parameters are recalled correctly
* Tests can override default_url_options directly
A callable object passed as a constraint for a route may access the request
parameters as part of its check. This causes the combined parameters hash
to be cached in the environment hash. If the constraint fails then any subsequent
access of the request parameters will be against that stale hash.
To fix this we delete the cache after every call to `matches?`. This may have a
negative performance impact if the contraint wraps a large number of routes as the
parameters hash is built by merging GET, POST and path parameters.
Fixes#2510.
(cherry picked from commit 56030506563352944fed12a6bb4793bb2462094b)
Further, Samoa and Tokelau jumped across the IDL from Dec 29 to Dec 31, 2011
switching from UTC-11 to UTC+13. American Samoa did not make the change and
remains at UTC-11. Pacific/Fakaofo and Pacific/Apia are in TZInfo and
documentation about the dateline change is in austalasia at IANA.
(cherry picked from commit 5fe88b11f1)
Make sure to return the same hash object instead of returning a new one.
Returning a new one causes failures on cookie store tests, where it
tests for the 'Set-Cookie' header with the session signature.
This is due to the hash ordering changes on Ruby 1.8.7-p358.
Also skip persistente tests related to UPDATE + ORDER BY for postgresql
PostgreSQL does not support updates with order by, and these tests are
failing randomly depending on the fixture loading order now.
Logic in clone_empty method was dealing with old @dirty variable, which
has changed by @html_safe in this commit:
139963c99a
This was issuing a "not initialized variable" warning - related to:
https://github.com/rails/rails/pull/5237
The logic applied by this method is already handled by the [] override,
so there is no need to reset the variable here.
* 3-1-4:
bumping to 3.1.4
Ensure [] respects the status of the buffer.
updating RAILS_VERSION
use AS::SafeBuffer#clone_empty for flushing the output_buffer
add AS::SafeBuffer#clone_empty
fix output safety issue with select options
* 3-1-stable-security:
Ensure [] respects the status of the buffer.
use AS::SafeBuffer#clone_empty for flushing the output_buffer
add AS::SafeBuffer#clone_empty
fix output safety issue with select options
Restoring ability to derive id/sequence from tables with nonstandard sequences for primary keys
Conflicts:
activerecord/lib/active_record/connection_adapters/postgresql_adapter.rb
activerecord/test/cases/adapters/postgresql/schema_test.rb
Windows doesn't allow `\ / : * ? " < > |` in filenames so create
the fixture files at runtime and ignore the incompatible ones when
running on Windows.
RFC 3986[1] allows sub-delim characters in path segments unencoded,
however Rack::File requires them to be encoded so we use URI's
unescape method to leave them alone and then escape them again.
Also since the path gets passed to Dir[] we need to escape any glob
characters in the path.
[1]: http://www.ietf.org/rfc/rfc3986.txt
Since ActiveModel::Errors instance keeps all error messages as hash
we should duplicate this object as well.
Previously ActiveModel::Errors was a subclass of ActiveSupport::OrderedHash,
which results in different behavior on dup, this may result in regression for
people relying on it.
Because Rails 3.2 stills supports Ruby 1.8.7 in order to properly fix this
regression we need to backport #initialize_dup.
This behaves similarly to REQUEST_URI, but
we need to implement it on our own because
REQUEST_URI is not reliable.
Note that since PATH_INFO does not contain
information about trailing question mark,
this is not 100% accurate, for example
`/foo?` will result in `/foo` in ORIGINAL_FULLPATH
Removing this feature causes boost in performance when using Ruby 1.9.
Ruby 1.9 started to do implicit conversions using `to_ary` and `to_str`
in some STDLIB methods (like Array#join). To do such implicit conversions,
Ruby 1.9 always dispatches the method and rescues the NoMethodError exception
in case one is raised.
Therefore, since the whiners feature defined NilClass#method_missing, such
implicit conversions for nil became much, much slower. In fact, just defining
NilClass#method_missing (even without the whiners feature) already causes a
massive slow down. Here is a snippet that shows such slow down:
require "benchmark"
Benchmark.realtime { 1_000.times { [nil,nil,nil].join } }
class NilClass
def method_missing(*args)
raise NoMethodError
end
end
Benchmark.realtime { 1_000.times { [nil,nil,nil].join } }
- Add tests to protect from regressions in require's return value behavior
- See a10606c490 (require needs to return true or false) for the original bug fix
Revert html_escape to do a single gsub again, but add the "n" flag (no
language, i.e. not multi-byte) to protect against XSS via invalid utf8
Signed-off-by: José Valim <jose.valim@gmail.com>
When default_asset_host_protocol is left as nil, it will use absolute protocols when a request is present, and relative protocols otherwise (eg in asset generation)
Signed-off-by: José Valim <jose.valim@gmail.com>
Only used detailed schema introspection when doing a schema dump.
Conflicts:
activerecord/lib/active_record/connection_adapters/abstract_mysql_adapter.rb
Reason: After another round of discussion, it has been
decided to let interpolation deal with unsafe strings
as it should do.
This reverts commit d57d8098fc.
The bug is that an error would be raised while attempting to convert a
template from one encoding to another.
Please see http://redmine.ruby-lang.org/issues/5564 for more details.
The workaround is to load all conversions into memory ahead of time,
and will only happen if the ruby version is *exactly* 1.9.3p0. The
hope is obviously that the underlying problem will be resolved in
the next patchlevel release of 1.9.3.
Conflicts:
actionpack/CHANGELOG.md
Defining after_initialize and after_find as ordinary methods like
documented in the guide doesn't work with Rails 3.1.1; now macro-style
is used here, too.
Conflicted while cherry picking from master (Original Author: Florian Walch)
Reasons:
* Markdown reads well as plain text, but can also be formatted.
* It will make it easier for people to read on the web as Github
formats the Markdown nicely.
* It will encourage a level of consistency when people are writing
CHANGELOG entries.
The script used to perform the conversion is at
https://gist.github.com/1339263
Building the conditions of a nested through association could
potentially modify the conditions of the through and/or source
association.
This is a Bad Thing.
Conflicts:
activerecord/test/cases/reflection_test.rb
If a record is removed from a has_many :through, all of the join records
relating to that record should also be removed from the through
association's target.
(Previously the records were removed in the database, but only one was
removed from the in-memory target array.)
'Type' is a reserved column for STI. Changed conditions example to
avoid using that column name as an example. The example isn't
STI-related (and mentioning STI here is needless clutter), so changing
to avoid accidentally encouraging users to use 'type' as a column name
for other purposes.
The association guide previously recommended using a trick with single
quote delaying of string interpolation in order to handle setting association
conditions that would be evaluated at runtime. Using a proc is the new way as
this no longer works.
A robust Apache configuration for this feature seems to be tricky,
one that takes into account Accept-Encoding, sets Vary, ensures
Content-Type is right, etc.
This reverses a change made here: 5df72a238e (L0L324)
I'm assuming here that :load_config needs to be invoked separately from :environment, as it is elsewhere in the file for db operations, if not the alternative is to go back to "task :dump => :environment do".
Versions prior to 1.1 delete the If-Modified-Since and If-Not-Modified headers when
config.action_controller.perform_caching is true. This has two problems:
* unexpected inconsistent behaviour between development & production environments
* breaks applications that use of these headers
Signed-off-by: José Valim <jose.valim@gmail.com>
If you are coming from a Rails 3.0 application, you won't have
the correct Bundler require statement.
This will cause the gems under the `assets` group not to be available
in the development and production environment.
I think this is related to the issue #39 in rails-sass
https://github.com/rails/sass-rails/issues/39
The Kernel.exec method launches the current program ($0) after setting some environment variables. This current ruby script has a shebang line and no .rb in its name. Launching a script this way does not work on Windows. By using the +ruby+ method that is included in Rake, we can launch the current script with the current ruby version and bypass the identified Windows issue.
To the app developer, this means configuration add in
config/initializers/* will not be executed.
Plugins developers need to special case their initializers that are
meant to be run in the assets group by adding :group => :assets.
This test was introduced in pull request #2577. The author of the pull
request included the test, which was originally in #2230, at the same
time as fixing a separate (but related) bug. However, the author did not
include a fix for this test, so when #2577 was merged, a failing test
was introduced.
So I am removing the failing test for now. If anyone wants to fix the
bug, please do so and submit a complete pull request with a test and
a fix.
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)
Database adapters use a statement pool for limiting the number of open
prepared statments on the database. The limit defaults to 1000, but can
be adjusted in your database config by changing 'statement_limit'.
The capture helper may return nil when evaluation of the block has
produced a buffer which contains only whitespace, but that doesn't
mean content_for should return stored content.
* '3-1-stable' of github.com:rails/rails:
Always enable digest and compile in the Rake task
don't need edgeapi now that we are on 3.1
release notes: adds a couple of blank lines to get the markup right
adds the release notes of 3.1 to the index
Update Assets guide to reflect all the new changes
Revert "Ruby 1.8.7+ provides to_date/to_datetime, AS just makes them public." to support date implementations without private to_date/to_datetime on Time (home_run).
* 3-1-0:
bumping to 3.1.0
Bump sprockets up
Depend on sass-rails and coffee-rails 3.1.0
Revert "Merge pull request #2647 from dmathieu/no_rescue"
Merge pull request #2756 from guilleiguaran/manifest-location
Merge pull request #2748 from guilleiguaran/assets-version-config
adds the asset pipeline guide to the index
incorporate feedback from vijaydev and dasch to rephrase this to sound more natural, and some grammar fixes.
* '3-1-0' of github.com:rails/rails:
adds the asset pipeline guide to the index
incorporate feedback from vijaydev and dasch to rephrase this to sound more natural, and some grammar fixes.
* 3-1-0:
Configuration changes for asset pipeline: remove config.assets.allow_debugging, add config.assets.compile and config.assets.digest
Read digests of assets from manifest.yml if config.assets.manifest is on
let SDoc add a link to the source code in GitHub for each method
Documentation fixes
bumping to 3.1.0.rc8
bumping to 3.1.0.rc7
Update Rails 3.1 CHANGELOGs
Bump rack-cache, rack-test, rack-mount and sprockets up
clear and disable query cache when an exception is raised from called middleware
assert_no_match
deletes spurious arrow
use sdoc to generate the API
Don't modify params in place - fixes#2624
the command line guide is good to go
`load` should also return the value from `super`
require needs to return true or false. thank you Ryan "zenspider" Davis
bumping bcrypt-ruby requirement
Make ActionController::TestCase#recycle! set @protocol to nil
Add failing test case for #2654
Conflicts:
activerecord/test/cases/query_cache_test.rb
Last two asserts in this test assume that all_comments are ordered by posts.id and then by comments.id therefore additional ordering is added. Without it test was failing on Oracle which returned results in different order.
error. Strings tagged as binary will be stored in sqlite3 as blobs. It
is an error to insert binary data to a string column, so an error is
emitted in the log file. People are highly encouraged to track down the
source of the binary strings and make sure that the encoding is set
correctly before inserting to the database.
Fix broken test test_copying_migrations(RailtiesTest::EngineTest):
NoMethodError: undefined method `log_level' for #<Rails::Engine::Configuration:0xb87a98c>
This ensures that the protocol does not get carried over when there are
two or more requests in functional tests. This was a problem when e.g.
the first request was HTTP and the second request was HTTPS.
Closes#2654.
Signed-off-by: Andrew White <andyw@pixeltrix.co.uk>
This ensures that the protocol does not get carried over when there are
two or more requests in functional tests. This was a problem when e.g.
the first request was HTTP and the second request was HTTPS.
Closes#2654.
Signed-off-by: Andrew White <andyw@pixeltrix.co.uk>
(cherry picked from commit 68a13eded3)
* 3-1-stable: (21 commits)
Added a test to check for correct behaviour with no options in add_index command recorder
Using .try to test for the existence of a method option in a nil-resistent manner. Inlined the determination of the options hash for reversing using a ternary operator. Shortens the method in a way that keeps the code neat
Neatened up the invert_add_index method as per suggeston
Simple fix for correctly inverting an add_index migration when a name has been provided
Merge pull request #2524 from JonathonMA/fix_ecd37084b28a05f05251
Fix Sprockets rewrite_asset_path
Merge pull request #2620 from cesario/3-1-0
Fix sprockets warnings
Use typewriter styling on url_for in documentation
Merge pull request #2609 from guilleiguaran/bump-sprockets-beta14
remove extra space since comment_if already returns the space. (cherry picked from commit 5e73a2fa50)
Merge pull request #2596 from dharmatech/patch-1
Merge pull request #2595 from smartinez87/assets-guide
Merge pull request #2597 from dharmatech/patch-2
Merge pull request #2604 from vijaydev/params_wrapper_docs
Merge pull request #2602 from ernie/fix_predicate_builder_primary_key_assumption
Merge pull request #2603 from vijaydev/guides
Merge pull request #2581 from guilleiguaran/debug-assets-in-dev
Revert "Deprecate the use of non-public methods by Module#delegate"
mailer guide: fixes indentation, and use fixed width fonts wherever necessary
...
Regexps have a construct to express alternation of characters, which
is character classes. In addition to being the most specific idiom to
write this pattern, it reads better without the backslashes. Also, it
is better not to use a capture if none is needed. As a side-effect of
these changes, the gsub is marginally faster, but speed is not the
point of this commit.
yield(:unspecified_block) actually returns true even if :unspecified_block never
exists. This means you can't use the form yield(:unspecified_block) or yield.
We were experiencing CI test failures, for example:
* 3-1-stable: http://travis-ci.org/#!/rails/rails/builds/79473/L407
* master: http://travis-ci.org/#!/rails/rails/builds/79507/L80
These failures only happened on 1.8.7-p352, and we were only able to
reproduce on the Travis CI VM worker. We even tried creating a new 32
bit Ubuntu VM and running the tests on that, and it all worked fine.
After some epic trial and error, we discovered that replacing the
following:
fuu = Marshal.load(Marshal.dump(fuu))
with:
marshalled = Marshal.dump(fuu)
fuu = Marshal.load(marshalled)
seemed to prevent the failure.
We have NO IDEA why this is. If anyone has some great insight to
contribute then that is welcome. Otherwise, hopefully this will just
help us get the CI green again.
Many thanks to @joshk for help with sorting this out.
This reverts commit 403b06e98e.
The call to `message` calls `inspect` on our exception. The exception
holds a reference to the environment, and the controller. This string
becomes very large, and the call to `super` dups the string (in tern
doubling the memory used). I'm reverting this for 3.1 but leaving the
commit on master. We should stop holding references to so many objects
and reduce the size of our inspect.
Reason: As discussed in GitHub, it is debatable, and present tense
is fine (and simple, and preferred).
This reverts commit 54ccda9f0a5e4a5e72a4c159dc8787faaf65e8a2.
This reverts commits df63c99f06 and b17fd25a2d.
The change had unintended side effects, please see #2337.
Conflicts:
activerecord/test/cases/associations/has_many_associations_test.rb
When format is true, it is mandatory (as opposed to :format => false).
This is currently not possible with resource routes, which automatically
make format optional by default.
Changed descriptino of Gemfile to add reference to Bundler
Changed description of README to add 'You should edit ...'
Changed description of Rakefile to reflect dynamic nature of Rakefile
and point to app/lib/tasks/ as the place to add app-specific rake tasks.
Reason: Setting up MySQL and Postgres needs installing servers
and libraries that perhaps you don't need in your machine. We
think fixing a bug in Active Support without running the suite
of Active Record for all possible drivers is OK. That's why
we explain first --without db, and after that we cover AR
specifically. mysql2 is no different than mysql or pg in that
sense.
This reverts commit 50d5e0f7e8.
This appears to be a bug with RedCloth; parsing with official Textile creates the HTML that I expect. I have filed ticket # 226 on RedCloth's lighthouse project and I will attempt to fix the problem with RedCloth.
This commit is a workaround for the RedCloth bug that will give this guide the right formatting.
Unfortunately #respond_to?(:controller) won't work as suggested, nor will respond_to?(:params), as #controller is present and #params is delegated to #controller. #delegate makes respond_to? return true regardless whether the target responds to it.
Contains following patches cherry-picked from @lhahne's 3-0-stable branch:
* Added tests for the output_buffer returned by CacheHelper (c476a6b)
The output_buffer returned by CacheHelper should be html_safe if the original buffer is html_safe.
* made sure that the possible new output_buffer created by CacheHelper is of the same type as the original (39a4f67)
This option has caught a few people out, so it
is a good idea to include an explanation and
a Gotcha.
See Rails bug 1822
See 1822 comment 1585926 for resolution
This fixes errors when using Active Record outside of Rails. In Rails,
these files are required by other classes that are always loaded, so
this error does not happen.
Without core_ext/module/delegation, a NoMethodError is raised because
`delegate` remains undefined.
Without core_ext/class/attribute_acessors, an ArgumentError is raised because
`delegate` does not receive a value for its :to option.
Don't install ruby-debug if running the test suite on Travis,
linecache19 is the main offender, very very slow.
And do not install pg if Travis is bundling the gems, pg will be setup
on Travis soon.
The previous behavior was unintentional, and some people was relying on it. In the next version of Rails, the dynamic finder will always expecting the number of arguments to be equal or greater (so you can still pass the options to it.) It will raise the ArgumentError otherwise.
Two issues fixed:
1) connection_pool is not defined - needed by SessionStore#drop_table!
and create_table! since c94651f
2) initialization of connection to the default of AR::Base.connection
only occurred at the singleton level - the instance level method defined
by cattr_accessor did not have this logic
I have found that Rails will take an invalid session ID specified by the
client and materialize a session based on that session ID. This means
that it is possible, among other things, for a client to use an
arbitrarily weak session ID or for a client to resurrect a previous used
session ID. In other words, we cannot guarantee that all session IDs are
generated by the server and that they are (statistically) unique through
time.
The fix is to always generate a new session ID in #get_session if an
existing session cannot be found under the incoming session ID.
Also added new tests that make sure that an invalid session ID is never
materialized into a new session, regardless of whether it comes in via a
cookie or a URL parameter (when :cookie_only => false).
See new method mysql_creation_options. It is used by both
create_database and recreate_database so they are consistent.
(Manual cherry pick of 3ba3125b24)
This is because only template rendering works with streaming.
Setting it at the class level was also changing the behavior
of JSON and XML responses, closes#1337.
The changes broke bulk migration tests and were fixed in 4d256bc6;
however that brought back the issue of #1857 and so this commit goes
back to the original scenario and just adds change_table to the list
of methods which are to be recorded in the CommandRecorder. The
method_missing now delegates all calls to the underlying connection as
before.
This makes "sprockets/railtie" explicit. This means that sprockets will
be loaded when you require "rails/all". If you are not using requiring
"rails/all", you need to manually load it with all other framework
railties.
In order to be complete, this commit also adds --skip-sprockets to
the rails generator.
* rails/3-1-stable: (49 commits)
Add JavaScript Runtime name to the Rails Info properties (thanks Sam Ruby)
Merge pull request #1481 from arunagw/lib_assets
Improve ordering of multiple columns on postgresql
judgement -> judgment, according to guidelines
Follow rails convention by using Array.wrap
Allow to specify mass-assignment roles as array
Apply the default scope earlier when doing calculations. Fixes#1682.
updated rspec link
grammar changes in named routes description and clarity around possibilities for options
font style changes in perf guide
grammar changes to log subscriber docs
move example code to be above reconfiguring discussion; add clarity about silencers and filters; misc grammar changes - for backtrace cleaners
Fixed typo
use present tense on examples
more detail on how the flow between redirect and show works, and minor grammar
remove extra space and clarify how an exception is made for controller wide layouts
reorder layout selection examples to occur in the order that the code does, and provide more detail on why each selection is made
singularize module and be explicit about what is delivered
grammar (missing "a") and formatting
elaborate details on why to use _url instead of _path in mailers
...
Conflicts:
actionpack/lib/action_view/helpers/asset_paths.rb
actionpack/lib/sprockets/helpers/rails_helper.rb
This group is required by default only on development and test
(you can change it on config/application.rb).
`rake assets:precompile` will automatically add the assets group
to Rails.groups (and consequently Bundler.require) and should work
transparently.
* Uses open source smoscript since scptxfr came out in 2000.
* Uses sqlcmd which replaces osql.
* Takes advantage of host/user/pass passed down by TinyTDS.
Changed the default value for the `options` argument on `ActiveRecord::ConnectionAdapters::Table#column_exists?` from `nil` to an empty Hash `{}`.
That method calls through to `ActiveRecord::ConnectionAdapters::SchemaStatements#column_exists?` which expects `options` to be a Hash.
When `options` was `nil`, an error would occur in cases where the column did exist because the called method attempted to perform a key lookup on options.
config.action_controller.default_asset_host_protocol
It's best to leave this unset.
When unset the :request protocol is used whenever it can be and
:relative is used in the other situations.
When set to :request then assets hosts will be disabled when there
is no request in scope and will use the request protocol whenever a
request is in scope.
If set to :relative, then a relative protocol is always used except
for stylesheet link tags which must use the :request protocol to
avoid double downloads in IE6&7.
* rails/3-1-stable: (68 commits)
Revert changes done in c56618ec, 51cb7459 and 030950a. The links are pointing to files/*/*.html for usage in api.rubyonrails.org and as a result are broken in GitHub blobs. This is unavoidable (at least for now).
copy-edits 7c2db6c, cbf2af1, and f391f94
Added 'Configuring Assets' section to configuring guide
Update remove_index documentation
Typo.
minor clean up generators section
add info that plugin installs need git or svn installed
document doc:* rake tasks
typo changes fixed
fixed typo chnages
add details on how to use specific annotations in rake:notes
document how rake notes work
The name for the plain text template in Rails3 appears to be method.text.erb, not, method.text.plain.erb. Updated the doc to reflect this change. Also fixed a reference to an example that was incorrect.
form => form_for
Tweak linebreak in ActionController::Redirecting doc
Updated rails´s guides on the config.serve_static_assets and it´s settings in production mode using WEBrick. This documentation has it´s roots in #issue1657. You have to set config.serve_static_assets to true in production mode with WEBrick as only this includes ActionDispatch::Static.
Improve documentation around status code argument of redirect_to
minor copy edit 0bdeddb
Jruby => JRuby, Mysql => MySQL
Moving hint down. As required after all database examples.
...
pointing to files/*/*.html for usage in api.rubyonrails.org and as a
result are broken in GitHub blobs. This is unavoidable (at least for now).
Conflicts:
README.rdoc
and plugins inside engines
It seems that plugins inside a Rails 3.1 application proper (i.e. in
/vendor/plugins) are initialized before engines and plugins inside
engines.
After some debugging, I found the culprit in
Rails::Application::Railties#all:
def all(&block)
@all ||= railties + engines + super
@all.each(&block) if block
@all
end
The call to super here implicitly passes the &block argument, which
has the unfortunate side-effect of adding the plugin initializers
first (in front of other railties and engines) in the case of
Rails::Engine#initializers:
def initializers
initializers = []
railties.all { |r| initializers += r.initializers }
initializers += super
initializers
end
The solution here is to replace the super call with a call
to #plugins.
Users can still explicitly choose a compressor in their application.rb if they have a preference, but will usually want to let plugins choose defaults in their Railties.
[3.1.0.rc1] Plugins inside engines not eager-loaded properly and their
rake tasks ignored
Working with the new support for plugins inside engines in Rails 3.1,
I found that certain things that work for regular plugins don't work
for these new nested plugins. In particular, these methods in
Rails::Engine don't seem to understand that an engine could have
nested plugins:
#load_tasks
#load_generators
#load_console
#eager_load!
A solution which worked out for me is to move the calls to
railties.all { ... } from the overriding methods in Rails::Application
into Rails::Engine.
With the previous pattern RDoc processed
railties/lib/rails/generators/rails/plugin_new/templates/test/%name%_test.rb
and that resulted in a spurious "<" class.
* Fix the problem on manipulating on the `ActiveSupport::SafeBuffer`
* Make sure that we run `escape_javascript` on the `String`, to avoid unexpected behavior.
After the fragment rendering, `Builder` returns the `String` object instead of `ActionView::OutputBuffer`. Somehow the same procedure which was in [823aa223ef] does not play nice with the String, and result in the fragment got lost.
Also make sure that the versions of those methods which modify a string in place such as gsub! can't be called on safe buffers at all.
Conflicts:
activesupport/test/safe_buffer_test.rb
The guide was contracting itself. First it stated that 'there is a list
of callbacks in order of execution', on the list 'after_save' was placed
before 'after_create', and then there was a warning that 'after_save'
runs after them. I am not sure about 'around_' callbacks, but let's
leave that for another patch.
ActiveSupport::Dependencies now raises NameError if it finds an existing
constant in load_missing_constant. This better reflects the nature of
the error which is usually caused by calling constantize on a nested constant.
Closes#1423
* Sass gem registers a compressor
* Sass gem registers generators for assets and scaffold
* Create a default stylesheet_engine ("css") for apps that remove the Sass gem
Rails is a web-application framework that includes everything needed to create
database-backed web applications according to the Model-View-Control pattern.
database-backed web applications according to the {Model-View-Controller (MVC)}[http://en.wikipedia.org/wiki/Model%E2%80%93view%E2%80%93controller] pattern.
This pattern splits the view (also called the presentation) into "dumb"
templates that are primarily responsible for inserting pre-built data in between
HTML tags. The model contains the "smart" domain objects (such as Account,
Product, Person, Post) that holds all the business logic and knows how to
persist themselves to a database. The controller handles the incoming requests
(such as Save New Account, Update Product, Show Post) by manipulating the model
and directing data to the view.
Understanding the MVC pattern is key to understanding Rails. MVC divides your application
into three layers, each with a specific responsibility.
In Rails, the model is handled by what's called an object-relational mapping
layer entitled Active Record. This layer allows you to present the data from
database rows as objects and embellish these data objects with business logic
methods. You can read more about Active Record in its
The View layer is composed of "templates" that are responsible for providing
appropriate representations of your application's resources. Templates
can come in a variety of formats, but most view templates are \HTML with embedded Ruby
code (.erb files).
The Model layer represents your domain model (such as Account, Product, Person, Post)
and encapsulates the business logic that is specific to your application. In Rails,
database-backed model classes are derived from ActiveRecord::Base. Active Record allows
you to present the data from database rows as objects and embellish these data objects
with business logic methods. Although most Rails models are backed by a database, models
can also be ordinary Ruby classes, or Ruby classes that implement a set of interfaces as
provided by the ActiveModel module. You can read more about Active Record in its
* Mail::Part now no longer has nil as a default charset, it is always set to something, and defaults to UTF-8
* Added explict setting of charset in set_fields! method to make sure Mail has the user defined default
* Removed quoting.rb and refactored for Mail to take responsibility of all quoting and auto encoding requirements for the header.
* Fixed several tests which had incorrect encoding.
* Changed all utf-8 to UTF-8 for consistency
* Whole new API added with tests. See base.rb for full details. Old API is deprecated.
* The Mail::Message class has helped methods for all the field types that return 'common' defaults for the common use case, so to get the subject, mail.subject will give you a string, mail.date will give you a DateTime object, mail.from will give you an array of address specs (mikel@test.lindsaar.net) etc. If you want to access the field object itself, call mail[:field_name] which will return the field object you want, which you can then chain, like mail[:from].formatted
* Mail#content_type now returns the content_type field as a string. If you want the mime type of a mail, then you call Mail#mime_type (eg, text/plain), if you want the parameters of the content type field, you call Mail#content_type_parameters which gives you a hash, eg {'format' => 'flowed', 'charset' => 'utf-8'}
* ActionMailer::Base :default_implicit_parts_order now is in the sequence of the order you want, no reversing of ordering takes place. The default order now is text/plain, then text/enriched, then text/html and then any other part that is not one of these three.
* Mail does not have "quoted_body", "quoted_subject" etc. All of these are accessed via body.encoded, subject.encoded etc
* Every object in a Mail object returns an object, never a string. So Mail.body returns a Mail::Body class object, need to call #encoded or #decoded to get the string you want.
* Mail::Message#set_content_type does not exist, it is simply Mail::Message#content_type
* Every mail message gets a unique message_id unless you specify one, had to change all the tests that check for equality with expected.encoded == actual.encoded to first replace their message_ids with control values
* Mail now has a proper concept of parts, remove the ActionMailer::Part and ActionMailer::PartContainer classes
* Calling #encoded on any object returns it as a string ready to go into the output stream of an email, this means it includes the \r\n at the end of the lines and the object is pre-wrapped with \r\n\t if it is a header field. Also, the "encoded" value includes the field name if it is a header field.
* Attachments are only the actual attachment, with filename etc. A part contains an attachment. The part has the content_type etc. So attachments.last.content_type is invalid. But parts.last.content_type
* There is no idea of a "sub_head" in Mail. A part is just a Message with some extra functionality, so it just has a "header" like a normal mail message
*2.3.2 [Final] (March 15, 2009)*
* Fixed that ActionMailer should send correctly formatted Return-Path in MAIL FROM for SMTP #1842 [Matt Jones]
* Fixed that no body charset would be set when there are attachments present #740 [Paweł Kondzior]
*2.2.1 [RC2] (November 14th, 2008)*
* Turn on STARTTLS if it is available in Net::SMTP (added in Ruby 1.8.7) and the SMTP server supports it (This is required for Gmail's SMTP server) #1336 [Grant Hollingworth]
*2.2.0 [RC1] (October 24th, 2008)*
* Add layout functionality to mailers [Pratik Naik]
Mailer layouts behaves just like controller layouts, except layout names need to
have '_mailer' postfix for them to be automatically picked up.
*2.1.0 (May 31st, 2008)*
* Fixed that a return-path header would be ignored #7572 [joost]
* Less verbose mail logging: just recipients for :info log level; the whole email for :debug only. #8000 [iaddict, Tarmo Tänav]
* Updated TMail to version 1.2.1 [Mikel Lindsaar]
* Fixed that you don't have to call super in ActionMailer::TestCase#setup #10406 [jamesgolick]
*2.0.2* (December 16th, 2007)
* Included in Rails 2.0.2
*2.0.1* (December 7th, 2007)
* Update ActionMailer so it treats ActionView the same way that ActionController does. Closes #10244 [Rick Olson]
* Pass the template_root as an array as ActionView's view_path
* Request templates with the "#{mailer_name}/#{action}" as opposed to just "#{action}"
* Fixed that partials would be broken when using text.plain.erb as the extension #10130 [java]
* Update README to use new smtp settings configuration API. Closes #10060 [psq]
* Allow ActionMailer subclasses to individually set their delivery method (so two subclasses can have different delivery methods) #10033 [Zach Dennis]
* Update TMail to v1.1.0. Use an updated version of TMail if available. [Mikel Lindsaar]
* Introduce a new base test class for testing Mailers. ActionMailer::TestCase [Michael Koziarski]
* Fix silent failure of rxml templates. #9879 [jstewart]
* Fix attachment decoding when using the TMail C extension. #7861 [orangechicken]
* Increase mail delivery test coverage. #8692 [Kamal Fariz Mahyuddin]
* Register alternative template engines using ActionMailer::Base.register_template_extension('haml'). #7534 [cwd, Josh Peek]
* Only load ActionController::UrlWriter if ActionController is present [Rick Olson]
* Allow mailer actions named send by using __send__ internally. #6467 [iGEL]
* Add assert_emails and assert_no_emails to test the number of emails delivered. #6479 [Jonathan Viney]
# Assert total number of emails delivered:
assert_emails 0
ContactMailer.deliver_contact
assert_emails 1
# Assert number of emails delivered within a block:
assert_emails 1 do
post :signup, :name => 'Jonathan'
end
*1.3.3* (March 12th, 2007)
* Depend on Action Pack 1.13.3
*1.3.2* (February 5th, 2007)
* Deprecate server_settings renaming it to smtp_settings, add sendmail_settings to allow you to override the arguments to and location of the sendmail executable. [Michael Koziarski]
*1.3.1* (January 16th, 2007)
* Depend on Action Pack 1.13.1
*1.3.0* (January 16th, 2007)
* Make mime version default to 1.0. closes #2323 [ror@andreas-s.net]
* Make sure quoted-printable text is decoded correctly when only portions of the text are encoded. closes #3154. [jon@siliconcircus.com]
* Make sure DOS newlines in quoted-printable text are normalized to unix newlines before unquoting. closes #4166 and #4452. [Jamis Buck]
* Fixed that iconv decoding should catch InvalidEncoding #3153 [jon@siliconcircus.com]
* Replace alias method chaining with Module#alias_method_chain. [Marcel Molina Jr.]
* Replace Ruby's deprecated append_features in favor of included. [Marcel Molina Jr.]
* Correct spurious documentation example code which results in a SyntaxError. [Marcel Molina Jr.]
*1.2.1* (April 6th, 2006)
* Be part of Rails 1.1.1
*1.2.0* (March 27th, 2006)
* Nil charset caused subject line to be improperly quoted in implicitly multipart messages #2662 [ehalvorsen+rails@runbox.com]
* Parse content-type apart before using it so that sub-parts of the header can be set correctly #2918 [Jamis Buck]
* Make custom headers work in subparts #4034 [elan@bluemandrill.com]
* Template paths with dot chars in them no longer mess up implicit template selection for multipart messages #3332 [Chad Fowler]
* Make sure anything with content-disposition of "attachment" is passed to the attachment presenter when parsing an email body [Jamis Buck]
* Make sure TMail#attachments includes anything with content-disposition of "attachment", regardless of content-type [Jamis Buck]
*1.1.5* (December 13th, 2005)
* Become part of Rails 1.0
*1.1.4* (December 7th, 2005)
* Rename Version constant to VERSION. #2802 [Marcel Molina Jr.]
* Stricter matching for implicitly multipart filenames excludes files ending in unsupported extensions (such as foo.rhtml.bak) and without a two-part content type (such as foo.text.rhtml or foo.text.really.plain.rhtml). #2398 [Dave Burt <dave@burt.id.au>, Jeremy Kemper]
*1.1.3* (November 7th, 2005)
* Allow Mailers to have custom initialize methods that set default instance variables for all mail actions #2563 [mrj@bigpond.net.au]
*1.1.2* (October 26th, 2005)
* Upgraded to Action Pack 1.10.2
*1.1.1* (October 19th, 2005)
* Upgraded to Action Pack 1.10.1
*1.1.0* (October 16th, 2005)
* Update and extend documentation (rdoc)
* Minero Aoki made TMail available to Rails/ActionMailer under the MIT license (instead of LGPL) [RubyConf '05]
* Austin Ziegler made Text::Simple available to Rails/ActionMailer under a MIT-like licens [See rails ML, subject "Text::Format Licence Exception" on Oct 15, 2005]
* Fix vendor require paths to prevent files being required twice
* Don't add charset to content-type header for a part that contains subparts (for AOL compatibility) #2013 [John Long]
* Preserve underscores when unquoting message bodies #1930
* Add a unified render method to ActionMailer (delegates to ActionView::Base#render)
* Move mailer initialization to a separate (overridable) method, so that subclasses may alter the various defaults #1727
* Look at content-location header (if available) to determine filename of attachments #1670
* ActionMailer::Base.deliver(email) had been accidentally removed, but was documented in the Rails book #1849
* Fix problem with sendmail delivery where headers should be delimited by \n characters instead of \r\n, which confuses some mail readers #1742 [Kent Sibilev]
*1.0.1* (11 July, 2005)
* Bind to Action Pack 1.9.1
*1.0.0* (6 July, 2005)
* Avoid adding nil header values #1392
* Better multipart support with implicit multipart/alternative and sorting of subparts [John Long]
* Allow for nested parts in multipart mails #1570 [Flurin Egger]
* Normalize line endings in outgoing mail bodies to "\n" #1536 [John Long]
* Allow template to be explicitly specified #1448 [tuxie@dekadance.se]
* Allow specific "multipart/xxx" content-type to be set on multipart messages #1412 [Flurin Egger]
* Unquoted @ characters in headers are now accepted in spite of RFC 822 #1206
* Helper support (borrowed from ActionPack)
* Silently ignore Errno::EINVAL errors when converting text.
* Don't cause an error when parsing an encoded attachment name #1340 [lon@speedymac.com]
* Nested multipart message parts are correctly processed in TMail::Mail#body
* BCC headers are removed when sending via SMTP #1402
* Added 'content_type' accessor, to allow content type to be set on a per-message basis. content_type defaults to "text/plain".
* Silently ignore Iconv::IllegalSequence errors when converting text #1341 [lon@speedymac.com]
* Support attachments and multipart messages.
* Added new accessors for the various mail properties.
* Fix to only perform the charset conversion if a 'from' and a 'to' charset are given (make no assumptions about what the charset was) #1276 [Jamis Buck]
* Fix attachments and content-type problems #1276 [Jamis Buck]
* Fixed the TMail#body method to look at the content-transfer-encoding header and unquote the body according to the rules it specifies #1265 [Jamis Buck]
* Added unquoting even if the iconv lib can't be loaded--in that case, only the charset conversion is skipped #1265 [Jamis Buck]
* Added automatic decoding of base64 bodies #1214 [Jamis Buck]
* Added that delivery errors are caught in a way so the mail is still returned whether the delivery was successful or not
* Fixed that email address like "Jamis Buck, M.D." <wild.medicine@example.net> would cause the quoter to generate emails resulting in "bad address" errors from the mail server #1220 [Jamis Buck]
*0.9.1* (20th April, 2005)
* Depend on Action Pack 1.8.1
*0.9.0* (19th April, 2005)
* Added that deliver_* will now return the email that was sent
* Added that quoting to UTF-8 only happens if the characters used are in that range #955 [Jamis Buck]
* Fixed quoting for all address headers, not just to #955 [Jamis Buck]
* Fixed unquoting of emails that doesn't have an explicit charset #1036 [wolfgang@stufenlos.net]
*0.8.1* (27th March, 2005)
* Fixed that if charset was found that the end of a mime part declaration TMail would throw an error #919 [lon@speedymac.com]
* Fixed that TMail::Unquoter would fail to recognize quoting method if it was in lowercase #919 [lon@speedymac.com]
* Fixed that TMail::Encoder would fail when it attempts to parse e-mail addresses which are encoded using something other than the messages encoding method #919 [lon@speedymac.com]
* Added rescue for missing iconv library and throws warnings if subject/body is called on a TMail object without it instead
*0.8.0* (22th March, 2005)
* Added framework support for processing incoming emails with an Action Mailer class. See example in README.
*0.7.1* (7th March, 2005)
* Bind to newest Action Pack (1.5.1)
*0.7.0* (24th February, 2005)
* Added support for charsets for both subject and body. The default charset is now UTF-8 #673 [Jamis Buck]. Examples:
def iso_charset(recipient)
@recipients = recipient
@subject = "testing iso charsets"
@from = "system@loudthinking.com"
@body = "Nothing to see here."
@charset = "iso-8859-1"
end
def unencoded_subject(recipient)
@recipients = recipient
@subject = "testing unencoded subject"
@from = "system@loudthinking.com"
@body = "Nothing to see here."
@encode_subject = false
@charset = "iso-8859-1"
end
*0.6.1* (January 18th, 2005)
* Fixed sending of emails to use Tmail#from not the deprecated Tmail#from_address
*0.6* (January 17th, 2005)
* Fixed that bcc and cc should be settable through @bcc and @cc -- not just @headers["Bcc"] and @headers["Cc"] #453 [Eric Hodel]
* Fixed Action Mailer to be "warnings safe" so you can run with ruby -w and not get framework warnings #453 [Eric Hodel]
*0.5*
* Added access to custom headers, like cc, bcc, and reply-to #268 [Andreas Schwarz]. Example:
* Consolidated the server configuration options into Base#server_settings= and expanded that with controls for authentication and more [Marten]
NOTE: This is an API change that could potentially break your application if you used the old application form. Please do change!
* Added Base#deliveries as an accessor for an array of emails sent out through that ActionMailer class when using the :test delivery option. [Jeremy Kemper]
* Added Base#perform_deliveries= which can be set to false to turn off the actual delivery of the email through smtp or sendmail.
This is especially useful for functional testing that shouldn't send off real emails, but still trigger delivery_* methods.
* Added option to specify delivery method with Base#delivery_method=. Default is :smtp and :sendmail is currently the only other option.
Sendmail is assumed to be present at "/usr/sbin/sendmail" if that option is used. [Kent Sibilev]
* Dropped "include TMail" as it added to much baggage into the default namespace (like Version) [Chad Fowler]
@@ -10,7 +10,7 @@ Mail gem. It provides a way to make emails using templates in the same
way that Action Controller renders views using templates.
Additionally, an Action Mailer class can be used to process incoming email,
such as allowing a weblog to accept new posts from an email (which could even
such as allowing a blog to accept new posts from an email (which could even
have been sent from a phone).
== Sending emails
@@ -59,6 +59,8 @@ generated would look like this:
Mr. david@loudthinking.com
Thank you for signing up!
In previous version of Rails you would call <tt>create_method_name</tt> and
<tt>deliver_method_name</tt>. Rails 3.0 has a much simpler interface, you
simply call the method and optionally call +deliver+ on the return value.
@@ -74,7 +76,7 @@ Or you can just chain the methods together like:
== Setting defaults
It is possible to set default values that will be used in every method in your Action Mailer class. To implement this functionality, you just call the public class method <tt>default</tt> which you get for free from ActionMailer::Base. This method accepts a Hash as the parameter. You can use any of the headers e-mail messages has, like <tt>:from</tt> as the key. You can also pass in a string as the key, like "Content-Type", but Action Mailer does this out of the box for you, so you wont need to worry about that. Finally it is also possible to pass in a Proc that will get evaluated when it is needed.
It is possible to set default values that will be used in every method in your Action Mailer class. To implement this functionality, you just call the public class method <tt>default</tt> which you get for free from ActionMailer::Base. This method accepts a Hash as the parameter. You can use any of the headers e-mail messages has, like <tt>:from</tt> as the key. You can also pass in a string as the key, like "Content-Type", but Action Mailer does this out of the box for you, so you won't need to worry about that. Finally it is also possible to pass in a Proc that will get evaluated when it is needed.
Note that every value you set with this method will get over written if you use the same key in your mailer method.
@@ -141,7 +143,7 @@ The latest version of Action Mailer can be installed with Rubygems:
Source code can be downloaded as part of the Rails project on GitHub
I'm assuming here that :load_config needs to be invoked
separately from :environment, as it is elsewhere in the
file for db operations, if not the alternative is to go
back to "task :dump => :environment do".
*Ben Woosley*
* Update to rack-cache 1.1.
Versions prior to 1.1 delete the If-Modified-Since and If-Not-Modified
headers when config.action_controller.perform_caching is true. This has two
problems:
* unexpected inconsistent behaviour between development &
production environments
* breaks applications that use of these headers
*Brendan Ribera*
* Ensure that enhancements to assets:precompile task are only run once *Sam Pohlenz*
* TestCase should respect the view_assigns API instead of pulling variables on
its own. *José Valim*
* javascript_path and stylesheet_path now refer to /assets if asset pipelining
is on. *Santiago Pastorino*
* button_to support form option. Now you're able to pass for example
'data-type' => 'json'. *ihower*
* image_path and image_tag should use /assets if asset pipelining is turned
on. Closes #3126*Santiago Pastorino and christos*
* Avoid use of existing precompiled assets during rake assets:precompile run.
Closes #3119*Guillermo Iguaran*
* Copy assets to nondigested filenames too *Santiago Pastorino*
* Give precedence to `config.digest = false` over the existence of
manifest.yml asset digests *christos*
* escape options for the stylesheet_link_tag method *Alexey Vakhov*
* Re-launch assets:precompile task using (Rake.)ruby instead of Kernel.exec so
it works on Windows *cablegram*
* env var passed to process shouldn't be modified in process method. [Santiago
Pastorino]
*`rake assets:precompile` loads the application but does not initialize
it.
To the app developer, this means configuration add in
config/initializers/* will not be executed.
Plugins developers need to special case their initializers that are
meant to be run in the assets group by adding :group => :assets. *José Valim*
* Sprockets uses config.assets.prefix for asset_path *asee*
* FileStore key_file_path properly limit filenames to 255 characters. *phuibonhoa*
* Fix Hash#to_query edge case with html_safe strings. *brainopia*
* Allow asset tag helper methods to accept :digest => false option in order to completely avoid the digest generation.
Useful for linking assets from static html files or from emails when the user
could probably look at an older html email with an older asset. *Santiago Pastorino*
* Don't mount Sprockets server at config.assets.prefix if config.assets.compile is false. *Mark J. Titorenko*
* Set relative url root in assets when controller isn't available for Sprockets (eg. Sass files using asset_path). Fixes #2435*Guillermo Iguaran*
* Fix basic auth credential generation to not make newlines. GH #2882
* Fixed the behavior of asset pipeline when config.assets.digest and config.assets.compile are false and requested asset isn't precompiled.
Before the requested asset were compiled anyway ignoring that the config.assets.compile flag is false. *Guillermo Iguaran*
* CookieJar is now Enumerable. Fixes #2795
* Fixed AssetNotPrecompiled error raised when rake assets:precompile is compiling certain .erb files. See GH #2763#2765#2805*Guillermo Iguaran*
* Manifest is correctly placed in assets path when default assets prefix is changed. Fixes #2776*Guillermo Iguaran*
* Fixed stylesheet_link_tag and javascript_include_tag to respect additional options passed by the users when debug is on. *Guillermo Iguaran*
* Fix ActiveRecord#exists? when passsed a nil value
* Fix assert_select_email to work on multipart and non-multipart emails as the method stopped working correctly in Rails 3.x due to changes in the new mail gem.
## Rails 3.1.0 (August 30, 2011) ##
* Param values are `paramified` in controller tests. *David Chelimsky*
* x_sendfile_header now defaults to nil and config/environments/production.rb doesn't set a particular value for it. This allows servers to set it through X-Sendfile-Type. *Santiago Pastorino*
* The submit form helper does not generate an id "object_name_id" anymore. *fbrusatti*
* Make sure respond_with with :js tries to render a template in all cases *José Valim*
* json_escape will now return a SafeBuffer string if it receives SafeBuffer string *tenderlove*
* Make sure escape_js returns SafeBuffer string if it receives SafeBuffer string *Prem Sichanugrist*
* Fix escape_js to work correctly with the new SafeBuffer restriction *Paul Gallagher*
* Brought back alternative convention for namespaced models in i18n *thoefer*
Now the key can be either "namespace.model" or "namespace/model" until further deprecation.
* It is prohibited to perform a in-place SafeBuffer mutation *tenderlove*
The old behavior of SafeBuffer allowed you to mutate string in place via
method like `sub!`. These methods can add unsafe strings to a safe buffer,
and the safe buffer will continue to be marked as safe.
In the above example, an untrusted string (`params[:xss]`) is added to the
safe buffer returned by `link_to`, and the untrusted content is successfully
sent to the client without being escaped. To prevent this from happening
`sub!` and other similar methods will now raise an exception when they are called on a safe buffer.
In addition to the in-place versions, some of the versions of these methods which return a copy of the string will incorrectly mark strings as safe. For example:
The new versions will now ensure that *all* strings returned by these methods on safe buffers are marked unsafe.
You can read more about this change in http://groups.google.com/group/rubyonrails-security/browse_thread/thread/2e516e7acc96c4fb
* Warn if we cannot verify CSRF token authenticity *José Valim*
* Allow AM/PM format in datetime selectors *Aditya Sanghi*
* Only show dump of regular env methods on exception screen (not all the rack crap) *DHH*
* auto_link has been removed with no replacement. If you still use auto_link
please install the rails_autolink gem:
http://github.com/tenderlove/rails_autolink
*tenderlove*
* Added streaming support, you can enable it with: *José Valim*
class PostsController < ActionController::Base
stream :only => :index
end
Please read the docs at `ActionController::Streaming` for more information.
* Added `ActionDispatch::Request.ignore_accept_header` to ignore accept headers and only consider the format given as parameter *José Valim*
* Created `ActionView::Renderer` and specified an API for `ActionView::Context`, check those objects for more information *José Valim*
* Added `ActionController::ParamsWrapper` to wrap parameters into a nested hash, and will be turned on for JSON request in new applications by default *Prem Sichanugrist*
This can be customized by setting `ActionController::Base.wrap_parameters` in `config/initializer/wrap_parameters.rb`
* RJS has been extracted out to a gem. *fxn*
* Implicit actions named not_implemented can be rendered. *Santiago Pastorino*
* Wildcard route will always match the optional format segment by default. *Prem Sichanugrist*
For example if you have this route:
map '*pages' => 'pages#show'
by requesting '/foo/bar.json', your `params[:pages]` will be equals to "foo/bar" with the request format of JSON. If you want the old 3.0.x behavior back, you could supply `:format => false` like this:
map '*pages' => 'pages#show', :format => false
* Added Base.http_basic_authenticate_with to do simple http basic authentication with a single class method call *DHH*
render :text => "I'm only accessible if you know the password"
end
end
* Allow you to add `force_ssl` into controller to force browser to transfer data via HTTPS protocol on that particular controller. You can also specify `:only` or `:except` to specific it to particular action. *DHH and Prem Sichanugrist*
* Allow FormHelper#form_for to specify the :method as a direct option instead of through the :html hash *DHH*
* Make JavaScriptHelper#j() an alias for JavaScriptHelper#escape_javascript() -- note this then supersedes the Object#j() method that the JSON gem adds within templates using the JavaScriptHelper *DHH*
* Sensitive query string parameters (specified in config.filter_parameters) will now be filtered out from the request paths in the log file. *Prem Sichanugrist, fxn*
* URL parameters which return false for to_param now appear in the query string (previously they were removed) *Andrew White*
* URL parameters which return nil for to_param are now removed from the query string *Andrew White*
* ActionDispatch::MiddlewareStack now uses composition over inheritance. It is
no longer an array which means there may be methods missing that were not
tested.
* Add an :authenticity_token option to form_tag for custom handling or to omit the token (pass :authenticity_token => false). *Jakub Kuźma, Igor Wiedler*
* HTML5 button_tag helper. *Rizwan Reza*
* Template lookup now searches further up in the inheritance chain. *Artemave*
* Brought back config.action_view.cache_template_loading, which allows to decide whether templates should be cached or not. *Piotr Sarnacki*
* url_for and named url helpers now accept :subdomain and :domain as options, *Josh Kalderimis*
* The redirect route method now also accepts a hash of options which will only change the parts of the url in question, or an object which responds to call, allowing for redirects to be reused (check the documentation for examples). *Josh Kalderimis*
* Added config.action_controller.include_all_helpers. By default 'helper :all' is done in ActionController::Base, which includes all the helpers by default. Setting include_all_helpers to false will result in including only application_helper and helper corresponding to controller (like foo_helper for foo_controller). *Piotr Sarnacki*
* Added a convenience idiom to generate HTML5 data-* attributes in tag helpers from a :data hash of options:
Keys are dasherized. Values are JSON-encoded, except for strings and symbols. *Stephen Celis*
* Deprecate old template handler API. The new API simply requires a template handler to respond to call. *José Valim*
* :rhtml and :rxml were finally removed as template handlers. *José Valim*
* Moved etag responsibility from ActionDispatch::Response to the middleware stack. *José Valim*
* Rely on Rack::Session stores API for more compatibility across the Ruby world. This is backwards incompatible since Rack::Session expects #get_session to accept 4 arguments and requires #destroy_session instead of simply #destroy. *José Valim*
* file_field automatically adds :multipart => true to the enclosing form. *Santiago Pastorino*
* Renames csrf_meta_tag -> csrf_meta_tags, and aliases csrf_meta_tag for backwards compatibility. *fxn*
* Add Rack::Cache to the default stack. Create a Rails store that delegates to the Rails cache, so by default, whatever caching layer you are using will be used for HTTP caching. Note that Rack::Cache will be used if you use #expires_in, #fresh_when or #stale with :public => true. Otherwise, the caching rules will apply to the browser only. *Yehuda Katz, Carl Lerche*
Please check [3-0-stable](https://github.com/rails/rails/blob/3-0-stable/actionpack/CHANGELOG) for previous changes.
# It's also possible to construct multi-dimensional parameter hashes by specifying keys using brackets, such as:
#
@@ -63,7 +63,7 @@ module ActionController
#
# == Sessions
#
# Sessions allows you to store objects in between requests. This is useful for objects that are not yet ready to be persisted,
# Sessions allow you to store objects in between requests. This is useful for objects that are not yet ready to be persisted,
# such as a Signup object constructed in a multi-paged process, or objects that don't change much and are needed all the time, such
# as a User object for a system that requires login. The session should not be used, however, as a cache for objects where it's likely
# they could be changed unknowingly. It's usually too much work to keep it all synchronized -- something databases already excel at.
@@ -116,8 +116,8 @@ module ActionController
#
# Title: <%= @post.title %>
#
# You don't have to rely on the automated rendering. Especially actions that could result in the rendering of different templates will use
# the manual rendering methods:
# You don't have to rely on the automated rendering. For example, actions that could result in the rendering of different templates
# will use the manual rendering methods:
#
# def search
# @results = Search.find(params[:query])
@@ -132,9 +132,9 @@ module ActionController
#
# == Redirects
#
# Redirects are used to move from one action to another. For example, after a <tt>create</tt> action, which stores a blog entry to a database,
# we might like to show the user the new entry. Because we're following good DRY principles (Don't Repeat Yourself), we're going to reuse (and redirect to)
# a <tt>show</tt> action that we'll assume has already been created. The code might look like this:
# Redirects are used to move from one action to another. For example, after a <tt>create</tt> action, which stores a blog entry to the
# database, we might like to show the user the new entry. Because we're following good DRY principles (Don't Repeat Yourself), we're
# going to reuse (and redirect to) a <tt>show</tt> action that we'll assume has already been created. The code might look like this:
#
# def create
# @entry = Entry.new(params[:entry])
@@ -146,7 +146,9 @@ module ActionController
# end
# end
#
# In this case, after saving our new entry to the database, the user is redirected to the <tt>show</tt> method which is then executed.
# In this case, after saving our new entry to the database, the user is redirected to the <tt>show</tt> method, which is then executed.
# Note that this is an external HTTP-level redirection which will cause the browser to make a second request (a GET to the show action),
# and not some internal re-routing which calls both "create" and then "show" within one request.
#
# Learn more about <tt>redirect_to</tt> and what options you have in ActionController::Redirecting.
#
@@ -210,16 +212,16 @@ module ActionController
# also include them at the bottom.
AbstractController::Callbacks,
# Append rescue at the bottom to wrap as much as possible.
Rescue,
# Add instrumentations hooks at the bottom, to ensure they instrument
# all the methods properly.
Instrumentation,
# Params wrapper should come before instrumentation so they are
# properly showed in logs
ParamsWrapper,
# The same with rescue, append it at the end to wrap as much as possible.
raiseActionDispatch::Session::SessionRestoreError,"Session contains objects whose class definition isn't available.\nRemember to require the classes for all objects kept in the session.\n(Original exception: #{const_error.message} [#{const_error.class}])\n"
end
retry
elsifargument_error.message=~%r{dump format error \(user class\)}
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
