ActiveSupport::Duration.parse('P3Y') == 3.years # It should be true
Duration parsing made independent from any moment of time:
Fixed length in seconds is assigned to each duration part during parsing.
Changed duration of months and years in seconds to more accurate and logical:
1. The value of 365.2425 days in Gregorian year is more accurate
as it accounts for every 400th non-leap year.
2. Month's length is bound to year's duration, which makes
sensible comparisons like `12.months == 1.year` to be `true`
and nonsensical ones like `30.days == 1.month` to be `false`.
Calculations on times and dates with durations shouldn't be affected as
duration's numeric value isn't used in calculations, only parts are used.
Methods on `Numeric` like `2.days` now use these predefined durations
to avoid duplicating of duration constants through the codebase and
eliminate creation of intermediate durations.
Regression introduced by ae29142142324545a328948e059e8b8118fd7a33 / 8363b879fe759f0645179f4521cc64795efbee6e.
Previously, cookies were only updated on `GET` requests. Now we will
update the helper for all requests, as part of `process`. Added
regression tests for all available HTTP method helpers in
`ActionController::TestCase`.
We want to avoid terminating the whole loop here, because it will cause
parameters that should be removed to not be removed, since we are
terminating early. In this specific case, `param2` is processed before
`param1` due to the reversing of `route.parts`, and since `param2` fails
the check on this line, it would previously cause the whole loop to
fail, and `param1` would still be in `parameterized_parts`. Now, we are
simply calling `next`, which is the intended behavior.
Introduced by 8ca8a2d773b942c4ea76baabe2df502a339d05b1.
Fixes#27454.
Scoring routes based on constraints repeated many type conversions that
could be performed in the outer loop. Determinations of score and
fitness also used Array operations that required allocations. Against
my benchmark with a large routeset, this reduced object allocations by
over 30x and wall time by over 3x.
This commit changes `parameter_encoding` to `skip_parameter_encoding`.
`skip_parameter_encoding` will set encoding on all parameters to
ASCII-8BIT for a given action on a particular controller. This allows
the controller to handle data when the encoding of that data is unknown,
for example file systems or truly binary parameters.
It was depending on a side-effect of the old html-scanner, so was no
longer proving what it intended to. Instead, assert more directly about
the resulting observable difference.
Instead of appending a format to the request, it's much better
to just pass a more appropriate accept header. Rails will figure
out the format from that instead.
This allows devs to use `:as` on routes that don't have a format.
Introduce an `IdentityEncoder` to avoid `if request_encoder`,
essentially a better version of the purpose of the `WWWFormEncoder`.
One that makes conceptual sense on GET requests too.
Fixes#27144.
Reset a new session directly after its creation in
`ActionDispatch::IntegrationTest#open_session`. Reset the session to a clean
state before making it available to the client's test code.
Issue #22742 reports unexpected behavior of integration tests that run multiple
sessions. For example an `ActionDispatch::Flash` instance is shared across
multiple sessions, though a client code will rightfully assume that each new
session has its own flash hash.
The following test failed due to this behavior:
class Issue22742Test < ActionDispatch::IntegrationTest
test 'issue #22742' do
integration_session # initialize first session
a = open_session
b = open_session
refute_same(a.integration_session, b.integration_session)
end
end
Instead of creating a new `ActionDispatch::Integration::Session` instance,
the same instance is shared across all newly opened test sessions. This is
due to the way how new test sessions are created in
`ActionDispatch::IntegrationTest#open_session`. The already existing
`ActionDispatch::IntegrationTest` instance is duplicated with `Object#dup`,
This approach was introduced in commit 15c31c7639b. `Object#dup` copies the
instance variables, but not the objects they reference. Therefore this issue
only occurred when the current test instance had been tapped in such a way that
the instance variable `@integration_session` was initialized before creating the
new test session.
Close#22742
[Tawan Sierek + Sina Sadeghian]
CONENT_LENGTH setted by string length, which is equal to number of
characters in string but StringIO.length is byte sequence and
when payload contains non-ASCII characters, stream's length will be
different. That's why real byte length should be used for CONTENT_LENGTH
header.
Add unit test for CONTENT_LENGTH header fix
It just passes non-ascii symbols as parameters and verifies that
"CONTENT_LENGTH" header has content bytes count as value.
The current implementation of AC::Parameters#permit builds permitted hashes and
then calls permit! on them.
This filtering is recursive, so we call permit! on terminal branches, but then
ascendants call permit! on themselves when the recursion goes up the stack,
which recurses all the way down again because permit! is recursive itself.
Repeat this for every parent node and you get some scary O-something going on
that I don't even want to compute.
Instead, since the whole point of the permit recursion is to build permitted
hashes along the way and at that point you know you've just come up with a
valid filtered version, you can already switch the toggle on the spot.
I have seen 2x speedups in casual benchmarks with small structures. As the
previous description shows, the difference in performance is going to be a
function of the nesting.
Note that that the involved methods are private and used only by permit.
This commit prevents a possible issue wherein an empty CONTENT_TYPE
header is sent in a request to a Rails application, and then `request.content_mime_type`
would return `nil`. This is because the `has_content_type?` guard method
was not properly checking the validity of a request's content type; it
was only checking to see whether or not the header existed, not whether
it had a value stored inside.
Relatedly, after an internal discussion, it was determined that the
`has_content_type?` method is not meant to be part of the public API,
and is therefore changed to a `:nodoc:` method in this commit.
The test for this behavior is a little bit ugly, for two reasons. One is
that it was difficult to determine where to place the test... I figured
the best place would be with the rest of the ParamsWrapper stuff, since
that's where the original issue was. Also, we have to do some fancy
footwork in calling `dispatch` on the test's controller manually... this
is because `ActionController::TestCase` will throw an error if you try
and pass in a nil content type, which is exactly what we are trying to
test here... Because of that, we have to manually call in to the
controller, and bypass the `post` request helper.
Fixes#26912.
This is a regression in behavior between Rails versions 4.2.x and 5.0.x,
which was introduced via [this commit](a9f28600e9).
I have been seeing people setting `Logger` instances for `config.logger`
and it blowing up on `rails/web-console` usage.
Now, I doubt many folks are manually setting `ActionView::Base.logger`,
but given that `DebugExceptions` is running in a pretty fragile
environment already, having it crash (and being silent) in those cases
can be pretty tricky to trace down.
I'm proposing we verify whether the `ActionView::Base.logger` supports
silencing before we try to do it, to save us the headache of tracing it
down.
When rendering arbitrary templates, it is helpful to not overwrite `env` keys with nil if they don't match any found in the `RACK_KEY_TRANSLATION`
This allows the developer to set the environment to exactly what is needed for rendering.
* ActionController::Parameters#deep_dup
* Tests for ActionController::Parameters#deep_dup
* Fix test for ActionController::Parameters#deep_dup
* More tests for ActionController::Parameters#deep_dup
[Rafael Mendonça França + Pavel Evstigneev]
ActionDispatch::ParamsParser class was removed in favor of
ActionDispatch::Http::Parameters so it is better to move the error
constant to the new class.
