When a event is finished propagating through Scope hierarchy the event's `currentScope` property
should be reset to `null` to avoid accidental use of this property in asynchronous event handlers.
In the previous code, the event's property would contain a reference to the last Scope instance that
was visited during the traversal, which is unlikely what the code trying to grab scope reference expects.
BREAKING CHANGE: $broadcast and $emit will now reset the `currentScope` property of the event to
null once the event finished propagating. If any code depends on asynchronously accessing thei
`currentScope` property, it should be migrated to use `targetScope` instead. All of these cases
should be considered programming bugs.
Closes#7445Closes#7523
This CL enables interpolation expressions to be escaped, by prefixing each character of their
start/end markers with a REVERSE SOLIDUS U+005C, and to render the escaped expression as a
regular interpolation expression.
Example:
`<span ng-init="foo='Hello'">{{foo}}, \\{\\{World!\\}\\}</span>` would be rendered as:
`<span ng-init="foo='Hello'">Hello, {{World!}}</span>`
This will also work with custom interpolation markers, for example:
module.
config(function($interpolateProvider) {
$interpolateProvider.startSymbol('\\\\');
$interpolateProvider.endSymbol('//');
}).
run(function($interpolate) {
// Will alert with "hello\\bar//":
alert($interpolate('\\\\foo//\\\\\\\\bar\\/\\/')({foo: "hello", bar: "world"}));
});
This change effectively only changes the rendering of these escaped markers, because they are
not context-aware, and are incapable of preventing nested expressions within those escaped
markers from being evaluated.
Therefore, backends are encouraged to ensure that when escaping expressions for security
reasons, every single instance of a start or end marker have each of its characters prefixed
with a backslash (REVERSE SOLIDUS, U+005C)
Closes#5601Closes#7517
Calling `$commitViewValue` was was dirtying the input, even if no update to the view
value was made.
For example, `updateOn` triggers and form submit may call `$commitViewValue` even
if the the view value had not changed.
Closes#7457Closes#7495
If you have two directives that both expect to receive transcluded content
the outer directive works but the inner directive never receives a
transclusion function. This only failed if the first transclude directive
was not the first directive found in compilation.
Fixes#7240Closes#7387
All isolated scope directives that do not have `templateUrl` were marked
as `$isolateScopeNoTemplate` even if they did have a `template` attribute.
This caused `jqLite#scope()` to return the wrong value for child elements
within the directive's template.
Closes#6942
Use the new `NgModelController.$commitViewValue()` method to commit the
`$viewValue` on all the child controls (including nested `ngForm`s) when the form
receives the `submit` event. This will happen immediately, overriding any
`updateOn` and `debounce` settings from `ngModelOptions`.
If you wish to access the committed `$modelValue`s then you can use the `ngSubmit`
directive to provide a handler. Don't use `ngClick` on the submit button, as this
handler would be called before the pending `$viewValue`s have been committed.
Closes#7017
Move responsibility for pending and debouncing model updates into `NgModelController`.
Now input directives are only responsible for capturing changes to the input element's
value and then calling `$setViewValue` with the new value.
Calls to `$setViewValue(value)` change the `$viewValue` property but these changes are
not committed to the `$modelValue` until an `updateOn` trigger occurs (and any related
`debounce` has resolved).
The `$$lastCommittedViewValue` is now stored when `$setViewValue(value)` updates
the `$viewValue`, which allows the view to be "reset" by calling `$rollbackViewValue()`.
The new `$commitViewValue()` method allows developers to force the `$viewValue` to be
committed through to the `$modelValue` immediately, ignoring `updateOn` triggers and
`debounce` delays.
BREAKING CHANGE:
This commit changes the API on `NgModelController`, both semantically and
in terms of adding and renaming methods.
* `$setViewValue(value)` -
This method still changes the `$viewValue` but does not immediately commit this
change through to the `$modelValue` as it did previously.
Now the value is committed only when a trigger specified in an associated
`ngModelOptions` directive occurs. If `ngModelOptions` also has a `debounce` delay
specified for the trigger then the change will also be debounced before being
committed.
In most cases this should not have a significant impact on how `NgModelController`
is used: If `updateOn` includes `default` then `$setViewValue` will trigger
a (potentially debounced) commit immediately.
* `$cancelUpdate()` - is renamed to `$rollbackViewValue()` and has the same meaning,
which is to revert the current `$viewValue` back to the `$lastCommittedViewValue`,
to cancel any pending debounced updates and to re-render the input.
To migrate code that used `$cancelUpdate()` follow the example below:
Before:
```
$scope.resetWithCancel = function (e) {
if (e.keyCode == 27) {
$scope.myForm.myInput1.$cancelUpdate();
$scope.myValue = '';
}
};
```
After:
```
$scope.resetWithCancel = function (e) {
if (e.keyCode == 27) {
$scope.myForm.myInput1.$rollbackViewValue();
$scope.myValue = '';
}
}
```
It is reasonable to expect a digest to occur between an input element
compiling and the first user interaction. Rather than add digests to
each test this change moves it into the `compileInput` helper function.
Due to a regression introduced several releases ago, the ability for multiple transclude functions
to work correctly changed, as they would break if different case labels had different numbers of
transclude functions.
This CL corrects this by not assuming that previous elements and scope count have the same length.
Fixes#7372Closes#7373
Because of how the logic was set up, a value of `0` was assumed to be the
same as `undefined`, which meant that you couldn't override the default
debounce delay with a value of zero.
For example, the following assigned a debounce delay of 500ms to the `blur`
event.
```
ngModelOptions="{ updateOn: 'default blur', debounce: {'default': 500, 'blur':
0} }"
```
Closes#7205
Input controls require `ngModel` which in turn brings in the `ngModelOptions`
but since ngModel does this initialization in the post link function, the
order in which the directives are run is relevant.
Directives are sorted by priority and name but `ngModel`, `input` and `textarea`
have the same priority. It just happens that `textarea` is alphabetically
sorted and so linked before `ngModel` (unlike `input`).
This is a problem since inputs expect `ngModelController.$options`
to exist at post-link time and for `textarea` this has not happened.
This is solved easily by moving the initialization of `ngModel` to the
pre-link function.
Closes#7281Closes#7292
BREAKING CHANGE
If `bar` is `undefined`, before `<img src="foo/{{bar}}.jpg">` yields
`<img src="foo/.jpg">`. With this change, the binding will not set `src`.
If you want the old behavior, you can do this: `<img src="foo/{{bar || ''}}.jpg">`.
The same applies for `srcset` as well.
Closes#6984
546cb42 introduced a regression, which would cause the function returned from
$interpolate to throw a ReferenceError if `context` is undefined. This change
prevents the error from being thrown.
Closes#7230Closes#7237
Code cleanup! response interceptors have been deprecated for some time, and it is confusing to have
two APIs, one of which is slightly "hidden" and hard to see, which perform the same task. The newer
API is a bit cleaner and more visible, so this is naturally preferred.
BREAKING CHANGE:
Previously, it was possible to register a response interceptor like so:
// register the interceptor as a service
$provide.factory('myHttpInterceptor', function($q, dependency1, dependency2) {
return function(promise) {
return promise.then(function(response) {
// do something on success
return response;
}, function(response) {
// do something on error
if (canRecover(response)) {
return responseOrNewPromise
}
return $q.reject(response);
});
}
});
$httpProvider.responseInterceptors.push('myHttpInterceptor');
Now, one must use the newer API introduced in v1.1.4 (4ae46814), like so:
$provide.factory('myHttpInterceptor', function($q) {
return {
response: function(response) {
// do something on success
return response;
},
responseError: function(response) {
// do something on error
if (canRecover(response)) {
return responseOrNewPromise
}
return $q.reject(response);
}
};
});
$httpProvider.interceptors.push('myHttpInterceptor');
More details on the new interceptors API (which has been around as of v1.1.4) can be found at
https://docs.angularjs.org/api/ng/service/$http#interceptorsCloses#7266Closes#7267
Previously, templates would always be assumed to be valid HTML nodes. In some cases, it is
desirable to use SVG or MathML or some other language.
For the time being, this change is only truly meaningful for SVG elements, as MathML has
very limited browser support. But in the future, who knows?
Closes#7265
This change undoes the use of watchGroup by code that uses $interpolate, by
moving the optimizations into the $interpolate itself. While this is not ideal,
it means that we are making the existing api faster rather than require people
to use $interpolate differently in order to benefit from the speed improvements.
BREAKING CHANGE: the function returned by $interpolate
no longer has a `.parts` array set on it.
It has been replaced by two arrays:
* `.expressions`, an array of the expressions in the
interpolated text. The expressions are parsed with
$parse, with an extra layer converting them to strings
when computed
* `.separators`, an array of strings representing the
separations between interpolations in the text.
This array is **always** 1 item longer than the
`.expressions` array for easy merging with it
Given an array of expressions, if any one expression changes then the listener function fires
with an arrays of old and new values.
$scope.watchGroup([expression1, expression2, expression3], function(newVals, oldVals) {
// newVals and oldVals are arrays of values corresponding to expression1..3
...
});
Port of angular/angular.dart@a3c31ce1dd
Certain versions of IE inexplicably trigger an input event in response to a placeholder
being set.
It is not possible to sniff for this behaviour nicely as the event is not triggered if
the element is not attached to the document, and the event triggers asynchronously so
it is not possible to accomplish this without deferring DOM compilation and slowing down
load times.
Closes#2614Closes#5960
This CL fixes problems and adds test cases for changes from #6421. Changes
include fixing the algorithm for preprocessing href attribute values, as
well as supporting xlink:href attributes. Credit for the original URL
parsing algorithm still goes to @richardcrichardc.
Good work, champ!
Previously, ctreq would possibly reference the incorrect directive name,
due to relying on a directiveName living outside of the closure which
throws the exception, which can change before the call is ever made.
This change saves the current value of directiveName as a property of
the link function, which prevents this from occurring.
Closes#7062Closes#7067
parseInt(Infinity, 10) will result in NaN, which becomes undesirable when the expected behaviour is
to return the entire input.
I believe this is possibly useful as a way to toggle input limiting based on certain factors.
Closes#6771Closes#7118
The `$cancelUpdate()` method on `NgModelController` cancels any pending debounce
action and resets the view value by invoking `$render()`.
This method should be invoked before programmatic update to the model of inputs
that might have pending updates due to `ng-model-options` specifying `updateOn`
or `debounce` properties.
Fixes#6994Closes#7014
By default, any change to an input will trigger an immediate model update,
form validation and run a $digest. This is not always desirable, especially
when you have a large number of bindings to update.
This PR implements a new directive `ngModelOptions`, which allow you to
override this default behavior in several ways. It is implemented as an
attribute, to which you pass an Angular expression, which evaluates to an
**options** object.
All inputs, using ngModel, will search for this directive in their ancestors
and use it if found. This makes it easy to provide options for a whole
form or even the whole page, as well as specifying exceptions for
individual inputs.
* You can specify what events trigger an update to the model by providing
an `updateOn` property on the **options** object. This property takes a
string containing a space separated list of events.
For example, `ng-model-options="{ updateOn: 'blur' }"` will update the
model only after the input loses focus.
There is a special pseudo-event, called "default", which maps to the
default event used by the input box normally. This is useful if you
want to keep the default behavior and just add new events.
* You can specify a debounce delay, how long to wait after the last triggering
event before updating the model, by providing a `debounce` property on
the **options** object.
This property can be a simple number, the
debounce delay for all events. For example,
`ng-model-options="{ debounce: 500 }" will ensure the model is updated
only when there has been a period 500ms since the last triggering event.
The property can also be an object, where the keys map to events and
the values are a corresponding debounce delay for that event.
This can be useful to force immediate updates on some specific
circumstances (like blur events). For example,
`ng-model-options="{ updateOn: 'default blur', debounce: { default: 500, blur: 0} }"`
This commit also brings to an end one of the longest running Pull Requests
in the history of AngularJS (#2129)! A testament to the patience of @lrlopez.
Closes#1285, #2129, #6945
With 1.2.x, `$animate.enter` and `$animate.move` both insert the element at the end of the provided
parent container element when only the `parent` element is provided. If an `after` element is provided
then they will place the inserted element after that one. This works fine, but there is no way to
place an item at the top of the provided parent container using these two APIs.
With this change, if the `after` argument is not specified for either `$animate.enter` or `$animate.move`,
the new child element will be inserted into the first position of the parent container element.
Closes#4934Closes#6275
BREAKING CHANGE: $animate will no longer default the after parameter to the last element of the parent
container. Instead, when after is not specified, the new element will be inserted as the first child of
the parent container.
To update existing code, change all instances of `$animate.enter()` or `$animate.move()` from:
`$animate.enter(element, parent);`
to:
`$animate.enter(element, parent, angular.element(parent[0].lastChild));`
When a async task interacts with a scope that has been destroyed already
and if it interacts with a property that is prototypically inherited from
some parent scope then resetting proto would make these inherited properties
inaccessible and would result in NPEs
The basic approach is to introduce a new elt.data() called $classCounts that keeps
track of how many times ngClass, ngClassEven, or ngClassOdd tries to add a given class.
The class is added only when the count goes from 0 to 1, and removed only when the
count hits 0.
To avoid duplicating work, some of the logic for checking which classes
to add/remove move into this directive and the directive calls $animate.
Closes#5271
Due to a known V8 memory leak[1] we need to perform extra cleanup to make it easier
for GC to collect this scope object.
V8 leaks are due to strong references from optimized code (fixed in M34) and inline
caches (fix in works). Inline caches are caches that the virtual machine builds on the
fly to speed up property access for javascript objects. These caches contain strong
references to objects so under certain conditions this can create a leak.
The reason why these leaks are extra bad for Scope instances is that scopes hold on
to ton of stuff, so when a single scope leaks, it makes a ton of other stuff leak.
This change removes references to objects that might be holding other big
objects. This means that even if the destroyed scope leaks, the child scopes
should not leak because we are not explicitly holding onto them.
Additionally in theory we should also help make the current scope eligible for GC
by changing properties of the current Scope object.
I was able to manually verify that this fixes the problem for the following
example app: http://plnkr.co/edit/FrSw6SCEVODk02Ljo8se
Given the nature of the problem I'm not 100% sure that this will work around
the V8 problem in scenarios common for Angular apps, but I guess it's better
than nothing.
This is a second attempt to enhance the cleanup, the first one failed and was
reverted because it was too aggressive and caused problems for existing apps.
See: #6897
[1] V8 bug: https://code.google.com/p/v8/issues/detail?id=2073Closes#6794Closes#6856Closes#6968
Previously, constant numbers with a unary minus sign were not treated as constants. This fix corrects
this behaviour, and may provide a small performance boost for certain applications, due to constant
watches being automatically unregistered after their first listener call.
Closes#6932
Makes xhr status text accessible is $http success/error callback.
See www.w3.org/TR/XMLHttpRequest/#dom-xmlhttprequest-statustext
Closes#2335Closes#2665Closes#6713
fb6062fb9d implements a
fix for NaN values causing $watchCollection to throw an infdig error. This change updates the test
by adding an assertion which explains what is actually being tested a bit better, and may also
provide better information in the event that the test ever fails.
Closes#6758
$watchCollection checks if oldValue !== newValue which does not work for NaN. This was causing
infinite digest errors, since comparing NaN to NaN in $watchCollection would always return false,
indicating that a change was occuring on each loop.
This fix adds a simple check to see if the current value and previous value are both NaN, and
if so, does not count it as a change.
Closes#4605
In Firefox, hovering over an option in an open select menu updates the selected property of option
elements. This means that when a render is triggered by the digest cycle, and the list of options
is being rendered, the selected properties are reset to the values from the model and the option
hovered over changes. This fix changes the code to only use DOM elements' selected properties in a
comparison when a change event has been fired. Otherwise, the internal new and existing option
arrays are used.
Closes#2448Closes#5994
The orderBy filter now allows string predicates passed to the orderBy filter to make use property
name predicates containing non-ident strings, such as spaces or percent signs, or non-latin
characters.
This behaviour requires the predicate string to be double-quoted.
In markup, this might look like so:
```html
<div ng-repeat="item in items | orderBy:'\"Tip %\"'">
...
</div>
```
Or in JS:
```js
var sorted = $filter('orderBy')(array, ['"Tip %"', '-"Subtotal $"'], false);
```
Closes#6143Closes#6144
This change brings Angular's JSONP behaviour closer in line with jQuery's. It will no longer treat
a callback called with no data as an error, and will no longer support IE8 via the onreadystatechanged
event.
BREAKING CHANGE:
Previously, the JSONP backend code would support IE8 by relying on the readystatechanged events. This
is no longer the case, as these events do not provide adequate useful information for deeming whether
or not a response is an error.
Previously, a JSONP response which did not pass data into the callback would be given a status of -2,
and treated as an error. Now, this situation will instead be given a status of 200, despite the lack
of data. This is useful for interaction with certain APIs.
Previously, the onload and onerror callbacks were added to the JSONP script tag. These have been
replaced with jQuery events, in order to gain access to the event object. This means that it is now
difficult to test if the callbacks are registered or not. This is possible with jQuery, using the
$.data("events") method, however it is currently impossible with jqLite. This is not expected to
break applications.
Closes#4987Closes#6735
In order to make the behavior compatible with $rootScope.$watch and $rootScope.$on methods, and
make it possible to deregister an attribute observer, Attributes.$observe method now returns a
deregistration function instead of the observer itself.
BREAKING CHANGE: calling attr.$observe no longer returns the observer function, but a
deregistration function instead.
To migrate the code follow the example below:
Before:
```
directive('directiveName', function() {
return {
link: function(scope, elm, attr) {
var observer = attr.$observe('someAttr', function(value) {
console.log(value);
});
}
};
});
```
After:
```
directive('directiveName', function() {
return {
link: function(scope, elm, attr) {
var observer = function(value) {
console.log(value);
};
attr.$observe('someAttr', observer);
}
};
});
```
Closes#5609
Originally we destroyed the oldValue by incrementaly copying over portions of the newValue
into the oldValue during dirty-checking, this resulted in oldValue to be equal to newValue
by the time we called the watchCollection listener.
The fix creates a copy of the newValue each time a change is detected and then uses that
copy *the next time* a change is detected.
To make `$watchCollection` behave the same way as `$watch`, during the first iteration
the listener is called with newValue and oldValue being identical.
Since many of the corner-cases are already covered by existing tests, I refactored the
test logging to include oldValue and made the tests more readable.
Closes#2621Closes#5661Closes#5688Closes#6736
