Current doc doesn't state required tag location clear enough. It was
[stack overflow|http://stackoverflow.com/a/16125138] where I've found that requirement
Closes#9741
If a response or expectation contained a date object then `$httpBackend.expect`
was not matching correctly.
This commit encodes then decodes the object being matched to ensure consistency.
Closes#5127
Check that listener is still present in $$listeners before decrease
$$listenerCount. It fixes problem with incorrect $$listenerCount after
call deregistering function multiple times.
Closes#9666Closes#9667
The event directives haven't stopped propagation by default in a long time.
If that behavior is desired, the handler may use the provided `$event` to call:
$event.stopPropagation();
Closes#9640
There was some inconsistency in version 1.2.25 regarding the definition
and documentation of shortcut methods in `ng.$http` and
`ngMock[E2E].$httpBackend`. Some methods didn't exist (although documented
as existing), whereas some methods did exist, but wheren't documented.
This commit fixes the above errors and adds tests to verify the existence
of all shortcut methods.
More specificcally, the following issues were addressed:
`ng.$http`: Add the missing `patch()` shortcut method and related docs.
`ng.$http` spec: Add test to verify the existence and functionality of the
`patch()` shortcut method.
`ngMock[E2E].$httpBackend`: Add docs for the (existing) `whenPATCH()`
shortcut method, add the missing `expectHEAD()`/`whenHEAD()` shortcut
methods, fix grammatical errors.
`ngMock[E2E].$httpBackend` spec: Add test to verify the existence of shortcut
methods for all HTTP verbs.
Closes#9180Closes#9321
Adds caching for url changes while a reload is happening,
as browsers do not allow to read out the new location the browser
is navigating to.
Removes unnecessary caching from $browser, as IE7-IE9 all
have the new hash value in `location.href` after changing it.
There was a wrong assumption in the previous version of this code
introduced by dca23173e2 and d70711481e.
Adds more tests for #6976Fixes#9235Closes#9470
IE10/11 have the following problem: When changing the url hash
via `history.pushState()` and then reverting the hash via direct
changes to `location.href` (or via a link) does not fire a
`hashchange` nor `popstate` event.
This commit changes the default behavior as follows:
Uses `location.href`/`location.replace` if the new url differs from
the previous url only in the hash fragment or the browser
does not support history API.
Use `history.pushState`/ `history.replaceState` otherwise.
Fixes#9143Closes#9406
$animate will cache subsequent calls to GCS in the event that the element
with the same CSS classes and the same parentNode is being animated. Once the
animation is started then $animate waits for one rAF before flushing the GCS
lookup cache. Prior to this fix, if GCS was unable to detect any transitions
or keyframes on the element then it would simply close the animation, but it
would not trigger the rAF code to flush the cache. This issue caused a bug
which made it difficult to detect why certain animations are not allowed to
fire if the element didn't contain any CSS-based animations beforehand.
Closes#8813
Backport of 22948807e3 without enforcing the `<base>` tag and without the new handling for links that only contain hash fragments.
Related to #6162Closes#8492
Backported from 1eda18365a
NgModel will format all scope-based values to string when setting the viewValue for
the associated input element. The formatting, however, only applies to input elements
that contain a text, email, url or blank input type. In the event of a null or undefined
scope or model value, the viewValue will be set to null or undefined instead of being
converted to an empty string.
Closes#5936Closes#9277
This reverts commit 9d9cdfb575.
This commit was causing breakages because of its assumption that transcluded
content would be handled predictably, i.e. with ngTransclude, whereas many
use cases involve manipulating transcluded content in linking functions.
Fix the following exploit:
hasOwnProperty.constructor.prototype.valueOf = valueOf.call;
["a", "alert(1)"].sort(hasOwnProperty.constructor);
The exploit:
• 1. Array.sort takes a comparison function and passes it 2 parameters to compare.
2. It then calls .valueOf() if the result is not a primitive.
• The Function object conveniently accepts two string arguments so we can use this
to construct a function. However, this doesn't do much unless we can execute it.
• We set the valueOf function on Function.prototype to Function.prototype.call.
This causes the function that we constructed to be executed when sort calls
.valueOf() on the result of the comparison.
The fix is in two parts.
• Disallow passing unsafe objects to function calls as parameters.
• Do not traverse the Function object when setting a path.
