At present, if you skip a callback that hasn't been defined,
activesupport callbacks silently does nothing. However, it's easy to
mistype the name of a callback and mistakenly think that it's being
skipped, when it is not.
This problem even exists in the current test suite.
CallbacksTest::SkipCallbacksTest#test_skip_person attempts to skip
callbacks that were never set up.
This PR changes `skip_callback` to raise an `ArgumentError` if the
specified callback cannot be found.
Wrapping an array in an `ArrayInquirer` gives a friendlier way to check its
string-like contents. For example, `request.variant` returns an `ArrayInquirer`
object. To check a request's variants, you can call:
request.variant.phone?
request.variant.any?(:phone, :tablet)
...instead of:
request.variant.include?(:phone)
request.variant.any? { |v| v.in?([:phone, :tablet]) }
`Array#inquiry` is a shortcut for wrapping the receiving array in an
`ArrayInquirer`:
pets = [:cat, :dog]
pets.cat? # => true
pets.ferret? # => false
pets.any?(:cat, :ferret} # => true
Reverting this as it's not the implementation that we would like it to be.
This is being used inside of ActiveSUpport::TimeZone[] and it's unaware
of the context in which to find the timezone period so the timezone found
changes depending on whether DST is in effect for the current period.
This means that `'2001-01-01'.in_time_zone(-9)` changes from winter/summer
even though it's the same date that we're trying to convert.
Since finding timezones by numeric offsets is a bit hit and miss we should
introduce a new API for finding them which supplies the date context in
which we want to search and we should probably also deprecate the finding
of timezones via the [] method, though this needs further discussion.
This reverts commit 2cc2fa3633edd96773023c6b09d07c7b9d9b841d.
When given a specific offset, use the first result found where the
total current offset (including any periodic deviations such as DST)
from UTC is equal.
`coder.represent_scalar` means something along the lines of "Here is a quoted
string, you can just add it to the output", which is not the case here. It only
works for simple strings that can appear unquoted in YAML, but causes problems
for e.g. primitive-like strings ("1", "true").
`coder.represent_object` on the other hand, means that "This is the Ruby-object
representation for this thing suitable for use in YAML dumping", which is what
we want here.
Before:
YAML.load ActiveSupport::SafeBuffer.new("Hello").to_yaml # => "Hello"
YAML.load ActiveSupport::SafeBuffer.new("true").to_yaml # => true
YAML.load ActiveSupport::SafeBuffer.new("false").to_yaml # => false
YAML.load ActiveSupport::SafeBuffer.new("1").to_yaml # => 1
YAML.load ActiveSupport::SafeBuffer.new("1.1").to_yaml # => 1.1
After:
YAML.load ActiveSupport::SafeBuffer.new("Hello").to_yaml # => "Hello"
YAML.load ActiveSupport::SafeBuffer.new("true").to_yaml # => "true"
YAML.load ActiveSupport::SafeBuffer.new("false").to_yaml # => "false"
YAML.load ActiveSupport::SafeBuffer.new("1").to_yaml # => "1"
YAML.load ActiveSupport::SafeBuffer.new("1.1").to_yaml # => "1.1"
If we ever want Ruby to behave more like PHP or JavaScript though, this is an
excellent trick to use ;)
Enable number_to_percentage to keep the number's precision by allowing :precision option value to be nil
Conflicts:
activesupport/CHANGELOG.md
activesupport/lib/active_support/number_helper.rb
activesupport/test/number_helper_test.rb
It's a thin layer to provide easy access to sample files throughout
test-cases. This adds the directory `test/fixtures/files` to newly
generated applications.
`#on_weekend?` returns true if the receiving date/time falls on a Saturday or
Sunday.
`#next_weekday` returns a new date/time representing the next day that does
not fall on a Saturday or Sunday.
`#prev_weekday` returns a new date/time representing the previous day that
does not fall on a Saturday or Sunday.
This stems from [a comment](rails#17227 (comment)) by @dhh.
In summary:
* New Rails 5.0 apps will not accept `return false` as a way to halt callback chains, and will not display a deprecation warning.
* Existing apps ported to Rails 5.0 will still accept `return false` as a way to halt callback chains, albeit with a deprecation warning.
For this purpose, this commit introduces a Rails configuration option:
```ruby
config.active_support.halt_callback_chains_on_return_false
```
For new Rails 5.0 apps, this option will be set to `false` by a new initializer
`config/initializers/callback_terminator.rb`:
```ruby
Rails.application.config.active_support.halt_callback_chains_on_return_false = false
```
For existing apps ported to Rails 5.0, the initializers above will not exist.
Even running `rake rails:update` will not create this initializer.
Since the default value of `halt_callback_chains_on_return_false` is set to
`true`, these apps will still accept `return true` as a way to halt callback
chains, displaying a deprecation warning.
Developers will be able to switch to the new behavior (and stop the warning)
by manually adding the line above to their `config/application.rb`.
A gist with the suggested release notes to add to Rails 5.0 after this
commit is available at https://gist.github.com/claudiob/614c59409fb7d11f2931
After this commit, returning `false` in a callback will display a deprecation
warning to make developers aware of the fact that they need to explicitly
`throw(:abort)` if their intention is to halt a callback chain.
This commit also patches two internal uses of AS::Callbacks (inside
ActiveRecord and ActionDispatch) which sometimes return `false` but whose
returned value is not meaningful for the purpose of execution.
In both cases, the returned value is set to `true`, which does not affect the
execution of the callbacks but prevents unrequested deprecation warnings from
showing up.
This commit changes arguments and default value of CallbackChain's :terminator
option.
After this commit, Chains of callbacks defined **without** an explicit
`:terminator` option will be halted as soon as a `before_` callback throws
`:abort`.
Chains of callbacks defined **with** a `:terminator` option will maintain their
existing behavior of halting as soon as a `before_` callback matches the
terminator's expectation. For instance, ActiveModel's callbacks will still
halt the chain when a `before_` callback returns `false`.
[ci skip]
As confirmed by @lleger (the author of `verified`) [in this comment](https://github.com/rails/rails/pull/17727#issuecomment-65488743):
> Actually, it no longer returns false explicitly (bc8cc56), so I guess the CHANGELOG isn't totally accurate. It returns nil instead (but the functionality isn't practically different).
This commit adds a `#verified` method to
`ActiveSupport::MessageVerifier` which will return either `false` when
it encounters an error or the message. `#verify` continues to raise an
`InvalidSignature` exception on error.
This commit also adds a convenience boolean method on `MessageVerifier`
as a way to check if a message is valid without performing the
decoding.
Added method `#eql?` to `ActiveSupport::Duration`, in addition to `#==`.
Conflicts:
activesupport/CHANGELOG.md
activesupport/lib/active_support/duration.rb
activesupport/test/core_ext/duration_test.rb
Goals:
1. Default to :random for newly generated applications
2. Default to :sorted for existing applications with a warning
3. Only show the warning once
4. Only show the warning if the app actually uses AS::TestCase
Fixes#16769
98b46bf5e2
did not properly handled out-of-range `:usec`s.
Passing a `:usec` that's out of range now throws an `ArgumentError` as it
should.
Fixes#16759.
For the sake of backward-compatibility, we need to make #instance_of?
return true for Fixnum. On the other hand, the method should still
give true for ActiveSupport::Duration itself which was not the case
before.
Since Duration is extending from ProxyObject which extends itself from
BasicObject, the Duration object doesn't respond to the #instance_of?
method. Thus, the #method_missing hook get triggered, delegating the
method to its `value` attribute.
However, Rubinius' #eql? definition relies on #instance_of?, thus this
will equal to true with a Fixnum (since its `value` attribute is a
Fixnum) while it should not.
The previous behavior was wrong anyway, no matter the implementation.
Builds on the work of #12550 where `.new` will convert the object (that respond to `#to_hash`) to a hash and
add that hash's keys and values to itself.
This change will also make `.new` respect the default value or proc of objects that respond to `#to_hash`.
In other words, this `.new` behaves exactly like `.new_from_hash_copying_default`.
`.new_from_hash_copying_default` now simply invokes `.new` and any references to `.new_from_hash_copying_default`
are replaced with `.new`.
Added tests confirm behavior.
`Time#as_json`, `Date#as_json` and `DateTime#as_json` incorrectly depends on a
delegation that is set up in `active_support/json/encoding`. We cannot simply
require that file in `core_ext/object/json` because it would cause a circular
dependency problem (see #12203 for background). We should instead rely on AS's
autoload to load that file for us on-demand.
To trigger autoload correctly, we need to reference the `AS::JSON::Encoding`
constant instead of using the delegated version.
Fixes#16131.
Since d3071db1, the apply_inflections method check if the downcased
version of a string is contained inside the "whitelist" of uncountable
words. However, if the word is composed of capital letters, it won't be
matched in the list while it should.
We can't simply revert to the previous behavior as there is a
performance concern (benchmarked over /usr/share/dict/words):
Before d3071db1 135.610000 0.290000 135.900000 (137.807081)
Since d3071db1 22.170000 0.020000 22.190000 ( 22.530005)
With the patch 22.060000 0.020000 22.080000 ( 22.125771)
Benchmarked with http://git.io/aFnWig
This way, the solution is to put the down-case version of words inside
the @uncountables array.
Replacements:
5.ago => 5.seconds.ago
5.until => 5.seconds.until
5.since => 5.seconds.since
5.from_now => 5.seconds.from_now
The removed tests does not affect coverage – we have equivalent test cases in
the tests for `AS::Duration`.
See #12389 for the history and rationale behind this.
When working with objects with a nanosecond component, the `-` method may
unnecessarily cause loss of precision.
`ActiveSupport::TimeWithZone#-` should return the same result as if we were
using `Time#-`:
Time.now.end_of_day - Time.now.beginning_of_day #=> 86399.999999999
Before:
Time.zone.now.end_of_day.nsec #=> 999999999
Time.zone.now.end_of_day - Time.zone.now.beginning_of_day #=> 86400.0
After:
Time.zone.now.end_of_day - Time.zone.now.beginning_of_day
#=> 86399.999999999
* Strips leading underscores.
* Changes some unnecessary gsub!s to sub!s.
* Replaces some anchors ^, $ with \A, \z.
* Documents that human inflection rules are applied.
* Documents that words are downcased except acronyms.
* Adds an example with an acronym.
* Rewords docs.
Make `#prepend` method modify instance in-place and return self
instead of just returning modified value. That is exactly what
`#prepend!` method was doing previously, so it's deprecated from
now on.
This reverts commit 475c96589ca65282e1a61350271c2f83f0d4044f, reversing
changes made to 705915ab5cf24430892107764b0050c07e1df583.
We decided that this is not worth busting everyone's cache as this
seems like a very unlikely problem. The problem only occurs when the
user is 1) not using a namespace, or 2) using the same namesapce for
different *kinds* of cache items. The recommended "fix" is to put
those cache items into their own namspace:
id = 1
Rails.cache.fetch(id, namespace: "user"){ User.find(id) }
ids = [1]
Rails.cache.fetch(ids, namespace: "users"){ User.find(ids) }
See the discussion on #14269 for details.
The current implementation of `fetch_multi` returns an array and has no
means to easily backtrack which names yielded which results. By changing
the return value to a Hash we retain the name information. Hash#values
can be used on the response if only the values are needed.
Rails applications are expected to be always aware of the application
time zone.
To be consistent with that contract, we have to assume that a bare
date passed to time helpers is a date in the application time zone,
not in the system time zone. The system time zone is irrelevant, we
should totally ignore it.
For example,
travel_to user.birth_date + 40.years
should make that user be 40th years old regardless of the system
time zone. Without this patch that may not be true.
Rails currently provides an extension to BigDecimal that redefines how
it is serialized to YAML. However, as noted in #12467, this does not
work as expected. When ActiveSupport is required, BigDecimal YAML
serialization does not maintain the object type. It instead ends up
serializing the number represented by the BigDecimal itself which, when
loaded by YAML later, becomes a Float:
```ruby
require 'yaml'
require 'bigdecimal'
yaml = BigDecimal('13.37').to_yaml
YAML.load(yaml).class
require 'active_support/all'
yaml = BigDecimal('13.37').to_yaml
YAML.load(yaml).class
```
@tenderlove posits that we should deprecate the custom BigDecimal
serialization and let Ruby handle it. For the time being, users who
require this serialization for backwards compatibility can manually
`require 'active_support/core_ext/big_decimal/yaml_conversions'`.
This will close#12467 and deprecate the custom BigDecimal#to_yaml.
Signed-off-by: David Celis <me@davidcel.is>
This behavior is only work out-of-box with minitest and also add a
downside to run after each test case, even if we don't used the travel
or travel_to methods
The api for filters with classes change and the guides weren't updated.
Now the class must respond for methods with the same name as the filter,
so the `before_action` calls a `before` method, and so on.
The method `#filter` has been deprecated in 4.0.0 and has been removed
in 4.1.0: #7560
Currently if a time is changed during DST overlap in the autumn then the
method `period_for_local` will return the DST period. However if the
original time is not DST then this can be surprising and is not what is
generally wanted. This commit changes that behavior to maintain the current
period if it's in the list of periods returned by `periods_for_local`.
It is possible to alter the behavior of `period_for_local` by specifying a
second argument but since we may be change from another time that could be
either DST or not then this would give inconsistent results.
Fixes#12163.
* Adding Hash#compact and Hash#compact! methods
* Using Ruby 1.9 syntax on documentation
* Updating guides for `Hash#compact` and `Hash#compact!` methods
* Updating CHANGELOG for ActiveSupport
* Removing unecessary protected method and lambda for `Hash#compact` implementations
* Performing `Hash#compact` implementation - https://gist.github.com/tinogomes/8332883
* fixing order position
* Fixing typo
Lazy loading the tzinfo library doesn't really buy us anything because
the gem is installed as a dependency via the gemspec and if a developer
is using Active Support outside of Rails then they can cherry pick which
files to load anyway.
Fixes#13553
The Date object has a xmlschema method starting with Ruby 1.9 so we were
assuming that we could safely remove this method and redefine it later
but the call to remove_method throws a NameError on FreeBSD so we should
rely on remove_possible_method instead.
This call is actually needed to avoid warnings when running the test
suite.
Fixes#11723
The contract of blank? and present? was in principle to return Object, as we
generally do, the test suite and description was consistent with that, but some
examples had comments like "# => true".
This cannot be unclear, we either fix the examples, or update the contract.
Since users may be already assuming singletons due to the examples and the fact
that they were returned before 30ba7ee, the safest option seems to be to revise
the contract and the implementation of String#blank?
The motivation for 30ba7ee was to improve the performance of the predicate, the
refactor based on === is on par regarding speed.
With this commit we start documenting return types using YARD conventions. We
plan to document return types gradually.
- Boolean parsing breaks on non strings (i.e. integer 1|0)
- Symbol parsing breaks on non strings.
- BigDecimal parsing breaks due to missing require.
- Update changelog.
Calling Kernel#silence_stream creates a new file descriptor which isn't
closed after it is used. As a result calling silence_stream multiple
times leads to a build up of loose file descriptors and can cause issues
in environments where garbage collection isn't run often.
Ruby's Date class automatically gives us #yesterday, #today,
and #tomorrow. And ActiveSupport has a handy Time.zone.today
for getting a localized version. But there was no localized
version of #yesterday or #tomorrow. Until now.
Extract **notable changes**, **deprecations** and **removals** from
each CHANGELOG.
I tried to reference the commits and pull requests for new features
and deprecations.
In the process I also made some minor changes to the CHANGELOGS.
The 4_1_release_notes guide is declared WIP.
The user is expected to explicitly convert the value into an
AS::Duration, i.e. `5.ago` => `5.seconds.ago`
This will help to catch subtle bugs like:
def recent?(days = 3)
self.created_at >= days.ago
end
The above code would check if the model is created within the last 3
**seconds**.
In the future, `Numeric#{ago,until,since,from_now}` should be removed
completely, or throw some sort of errors to indicate there are no
implicit conversion from `Numeric` to `AS::Duration`.
Also fixed & refactor the test cases for Numeric#{ago,since} and
AS::Duration#{ago,since}. The original test case had the assertion
flipped and the purpose of the test wasn't very clear.
Add `ActiveSupport::Testing::TimeHelpers#travel` and `#travel_to`. These
methods change current time to the given time or time difference by
stubbing `Time.now` and `Date.today` to return the time or date after
the difference calculation, or the time or date that got passed into the
method respectively. These methods also accept a block, which will
return current time back to its original state at the end of the block.
Example for `#travel`:
Time.now # => 2013-11-09 15:34:49 -05:00
travel 1.day
Time.now # => 2013-11-10 15:34:49 -05:00
Date.today # => Sun, 10 Nov 2013
Example for `#travel_to`:
Time.now # => 2013-11-09 15:34:49 -05:00
travel_to Time.new(2004, 11, 24, 01, 04, 44)
Time.now # => 2004-11-24 01:04:44 -05:00
Date.today # => Wed, 24 Nov 2004
Both of these methods also accept a block, which will return the current
time back to its original state at the end of the block:
Time.now # => 2013-11-09 15:34:49 -05:00
travel 1.day do
User.create.created_at # => Sun, 10 Nov 2013 15:34:49 EST -05:00
end
travel_to Time.new(2004, 11, 24, 01, 04, 44) do
User.create.created_at # => Wed, 24 Nov 2004 01:04:44 EST -05:00
end
Time.now # => 2013-11-09 15:34:49 -05:00
This module is included in `ActiveSupport::TestCase` automatically.
Previously, calling `::JSON.{generate,dump}` sometimes causes
unexpected failures such as intridea/multi_json#86.
`::JSON.{generate,dump}` now bypasses the ActiveSupport JSON encoder
completely and yields the same result with or without ActiveSupport.
This means that it will **not** call `as_json` and will ignore any
options that the JSON gem does not natively understand. To invoke
ActiveSupport's JSON encoder instead, use `obj.to_json(options)` or
`ActiveSupport::JSON.encode(obj, options)`.
See [1] for why this is not a good idea.
As part of this refactor, circular reference protection in as_json has
been removed and the corresponding error class has been deprecated.
As discussed with @jeremy, circular reference error is considered
programmer errors and protecting against it is out of scope for
the encoder.
This is again based on the excellent work by @sergiocampama in #11728.
[1]: https://github.com/intridea/multi_json/pull/138#issuecomment-24468223
So strings can be humanized without being capitalized:
'employee_salary'.humanize # => "Employee salary"
'employee_salary'.humanize(capitalize: false) # => "employee salary"
These methods now takes the same options as Hash#as_json, for example:
struct = Struct.new(:foo, :bar).new
struct.foo = "hello"
struct.bar = "world"
json = struct.as_json(only: [:foo]) # => {foo: "hello"}
This is extracted from PR #11728 from @sergiocampama, see also the
discussion in #11460.
Rails 4.1 has switched away from MultiJson, and does not currently
support any options on `ActiveSupport::JSON.decode`. Passing in
unsupported options (i.e. any non-empty options hash) will now raise
an ArgumentError.
Rationale:
1. We cannot guarantee the underlying JSON parser won't change in the
future, hence we cannot guarantee a consistent set of options the
method could take
2. The `json` gem, which happens to be the current JSON parser, takes
many dangerous options that is irrelevant to the purpose of AS's
JSON decoding API
3. To reserve the options hash for future use, e.g. overriding default
global options like ActiveSupport.parse_json_times
This change *DOES NOT* introduce any changes in the public API. The
signature of the method is still decode(json_text, options). The
difference is this method previously accepted undocumented options
which does different things when the underlying adapter changes. It
now correctly raises an ArgumentError when it encounters options that
it does not recognize (and currently it does not support any options).
Before, you were required to attach *after* adding the methods to the
class, since the attachment process needed the methods to be present.
With this change, any new method will also be attached to the configured
namespace.
You can now add partial days (e.g. 2.5.days) to `DateTime` with the advance method.
This was acheived by mimicing the `advance` implementation in `Time`.
Previously, an autoloaded constant `HTML::SomeClass` would not be marked
as autoloaded by AS::Dependencies. This is because the
`#loadable_constants_for_path` method uses `String#camelize` on the
inferred file path, which in turn means that, unless otherwise directed,
AS::Dependencies watches for loaded constants in the `Html` namespace.
By passing the original qualified constant name to `#load_or_require`,
this inference step is avoided, and the new constant is picked up in the
correct namespace.
Currently, the following returns `false`, contrary to expectation:
1.minute.eql?(1.minute)
Adding method `#eql?` will make this behave like expected. Method `#eql?` is
just a bit stricter than `#==`, as it checks whether the argument is also a
uration. Their parts may be different though.
1.minute.eql?(60.seconds) # => true
1.minute.eql?(60) # => false
Use a lambda to ensure that the generated string respects the offset of
the time value. Also add DateTime#to_s(:iso8601) and Date#to_s(:iso8601)
for completeness.
Previously, the cache size of `ActiveSupport::Cache::MemoryStore` was calculated
as the sum of the size of its entries, ignoring the size of keys and any data
structure overhead. This could lead to the calculated cache size sometimes being
10-100x smaller than the memory used, e.g., in the case of small values.
The size of a key/entry pair is now calculated via `#cached_size`:
def cached_size(key, entry)
key.to_s.bytesize + entry.size + PER_ENTRY_OVERHEAD
end
The value of `PER_ENTRY_OVERHEAD` is 240 bytes based on an [empirical
estimation](https://gist.github.com/ssimeonov/6047200) for 64-bit MRI on
1.9.3 and 2.0.
Fixes GH#11512 https://github.com/rails/rails/issues/11512
This fixes situations where nested NoMethodError exceptions are masked
by delegations. This would cause confusion especially where there was a
problem in the Rails booting process because of a delegation in the
routes reloading code.
Fixes#10559
The standard Ruby behavior for Time.at is to return the same type of
time when passing an instance of Time as a single argument. Since the
an ActiveSupport::TimeWithZone instance may be a different timezone than
the system timezone and DateTime just understands offsets the best we
can do is to return an instance of Time with the correct offset.
Fixes#11350.
Hash#select! returns nil if the hash didn't change and thus behaves differently
from select, so it's return value can't be used as result for the latter.
The previous implementation of BacktraceSilencer#noise did not
work correctly if more than one silencer was configured --
specifically, it would only return noise which was matched by all
silencers.
The new implementation is such that anything that has been matched by
silencers is removed from the backtrace using Array#- (array
difference), ie. we now return all elements within a backtrace that
have been matched by any silencer (and are thus removed by #silence).
Fixes#11030.
Add `DateTime#usec` and `DateTime#nsec` so that `ActiveSupport::TimeWithZone`
keeps sub-second resolution when wrapping a `DateTime` value.
Fixes#10855
It is possible under some environments to receive an Exception that is
not extended with Blamable (e.g. JRuby).
ActiveSupport::Dependencies::Loadable#load_dependency blindly call
blame_file! on the exception which throws it's own NoMethodError
exception and hides the original Exception.
This commit fixes#9521
Time.at allows passing a single Time argument which is then converted
to an integer. The conversion code since 1.9.3r429 explicitly checks
for an instance of Time so we need to override it to allow DateTime
and ActiveSupport::TimeWithZone values.
U+2028 and U+2029 are allowed inside strings in JSON (as all literal
Unicode characters) but JavaScript defines them as newline
seperators. Because no literal newlines are allowed in a string, this
causes a ParseError in the browser. We work around this issue by
replacing them with the escaped version. The resulting JSON is still
valid and can be parsed in the browser.
This commit has been coauthored with Viktor Kelemen @yikulju
Adds a ActiveSupport::Subscriber base class that LogSubscriber inherits
from. By inheriting from Subscriber, other kinds of subscribers can take
advantage of the event attachment system.
