Build:Tests: Fix custom build tests, verify on Travis; name Travis jobs
This commit fixes unit tests for the following builds:
1. The no-deprecated build: `custom:-deprecated`
2. The current slim build: `custom:-ajax,-effects`
3. The 4.0 (#4553) slim build: `custom:-ajax,-callbacks,-deferred,-effects`
It also adds separate Travis jobs for the no-deprecated & slim builds.
Apart from that, add intuitive names to Travis jobs. Otherwise it's hard to see
at a glance that a particular job is running on Firefox ESR, for example.
Ajax: Deprecate AJAX event aliases, inline event/alias into deprecated
A new `src/deprecated` directory makes it possible to exclude some deprecated
APIs from a custom build when their respective "parent" module is excluded
without keeping that module outside of the `src/deprecated` directory or
the `src/deprecated.js` file.
Tests: Make the support tests pass on Firefox 4x/5x/60
The check for old Firefox versions with different support test result only
checked for Firefox 52 or 60. It now checks for 4x/5x/60 to understand more
versions.
Event: Only attach events to objects that accept data - for real
There was a check in jQuery.event.add that was supposed to make it a noop
for objects that don't accept data like text or comment nodes. The problem was
the check was incorrect: it assumed `dataPriv.get( elem )` returns a falsy
value for an `elem` that doesn't accept data but that's not the case - we get
an empty object then. The check was changed to use `acceptData` directly.
Tests: Don't test synchronous XHR on unload in Chrome
Chrome 78 dropped support for synchronous XHR requests inside of
beforeunload, unload, pagehide, and visibilitychange event handlers.
See https://bugs.chromium.org/p/chromium/issues/detail?id=952452
Build: Run tests on Travis on FirefoxHeadless as well
Also, run them on both ChromeHeadless & FirefoxHeadless locally on
`grunt karma:main`.
Plus, so far, the chrome addons were installed for all the jobs, even
the ones that weren't used for browser testing. Changing that makes
those jobs faster.
Dave Methvin [Mon, 21 Oct 2019 16:53:24 +0000 (12:53 -0400)]
Offset: Send px-ed strings to .css()
An upcoming release of Migrate will generate warnings for calls to .css() that pass numbers rather than strings, see jquery/jquery-migrate#296 . At the moment, core's .offset() setter passes numbers rather than px strings so it would throw warnings. This commit fixes that.
Build: Require strict mode in Node.js scripts via ESLint
So far, only browser-based JS files were required to be in strict mode (in the
function form). This commit adds such a requirement to Node.js scripts where
the global form is preferred. All Node.js scripts in sloppy mode were
converted to strict mode.
Without this change passing `--dry-run` to jquery-release still pushes to the
jquery-dist repository which is dangerous as one can assume `--dry-run` to be
safe from external side effects.
This commit requires all function parameters to be used, not just the last one.
In cases where that's not possible as we need to match an external API, there's
an escape hatch of prefixing an unused argument with `_`.
This change makes it easier to catch unused AMD dependencies and unused
parameters in internal functions the API of which we may change at will, among
other things.
Unused AMD dependencies have been removed as part of this commit.
Build: Fix the regex parsing AMD var-modules (#4389)
The previous regex caused the final jQuery binary to have syntax errors for
var-modules with names starting with "return". For example, the following module
wouldn't work when the file is named `returnTrue.js`:
Wonseop Kim [Wed, 1 May 2019 12:57:55 +0000 (21:57 +0900)]
Effect: Fix a unnecessary conditional statement in .stop()
Because of the above conditional, the 'type' variable has a value of type
'string' or undefined. Therefore, boolean comparisons for 'type' variable
is always unnecessary because it return true. The patch removed the
unnecessary conditional statement.
Ahmed.S.ElAfifi [Mon, 19 Aug 2019 08:04:01 +0000 (10:04 +0200)]
Core: Use Array.prototype.flat where supported
Calling `Array.prototype.concat.apply( [], inputArray )` to flatten `inputArray`
crashes for large arrays; using `Array.prototype.flat` avoids these issues in
browsers that support it. In case it's necessary to support these large arrays
even in older browsers, a polyfill for `Array.prototype.flat` can be loaded.
This is already being done by many applications.
Tests: Move Android user agent detection above iOS, put Safari last
Android was catching the generic iOS regex. Checking for Android first should
eliminate that issue as iOS user agents don't contain the word "Android".
Putting Safari last makes Android UAs not be tested against it.
Tests: Make support tests accept Safari 13 & newer
The regexes detecting browsers with their versions have been updated for iOS
and Safari so that all iOS >=11 & all Safari (desktop) >= 11 are recognized.
This is required to make Safari 13 & iOS 13 pass support tests but it'll also
make tests forward-compatible with future Safari/iOS versions.
Traversing: Fix `contents()` on `<object>`s with children in IE
The original fix didn't account for the fact that in IE `<object>` elements
with no `data` attribute have an object `contentDocument`. The fix leverages
the fact that this special object has a null prototype.
The test for Shadow DOM v1 support has been changed to rely on the presence of
`documentElement.getRootNode` as iOS 10.0-10.2 supports `attachShadow` but
doesn't support `getRootNode`.
No new test is necessary - iOS 10.0 fails lots of our test suite because of
this bug.
Tests: Make Android Browser 4.0-4.3 dimensions tests green
Android Browser disregards td's box-sizing, treating it like it was content-box.
Unlike in IE, offsetHeight shares the same issue so there's no easy way to
workaround the issue without incurring high size penalty. Let's at least check
we get the size as the browser sees it.
Tests: Make Android Browser 4.0-4.3 AJAX tests green
Android Browser versions provided by BrowserStack fail the "prototype collision
(constructor)" test while locally fired emulators don't, even when they connect
to TestSwarm. Just skip the test there to avoid a red build.
Richard Gibson [Wed, 11 Jan 2017 22:19:30 +0000 (15:19 -0700)]
Event: Leverage native events for focus/blur/click; propagate additional data
Summary of the changes/fixes:
1. Trigger checkbox and radio click events identically (cherry-picked from b442abacbb8464f0165059e8da734e3143d0721f that was reverted before).
2. Manually trigger a native event before checkbox/radio handlers.
3. Add test coverage for triggering namespaced native-backed events.
4. Propagate extra parameters passed when triggering the click event to
the handlers.
5. Intercept and preserve namespaced native-backed events.
6. Leverage native events for focus and blur.
7. Accept that focusin handlers may fire more than once for now.
Build: Update jsdom; migrate a test with Symbol polyfill to an iframe test
So far, we've been testing that jQuery element iteration works with polyfilled
Symbol & transpiled for-of via a Node test with jsdom with the Symbol global
removed. Unfortunately, jsdom now requires Symbol to be present for its internal
functionality so such a test is no longer possible. Instead, it's been migrated
to an iframe test with transpiled JavaScript.
This PR also enables us to use ECMAScript 2017 or newer in Node.js code.
Build: Remove obsolete globals from ESLint configuration
We had quite a few obsolete globals declared in various ESLint config files. We also no longer allow to rely on the `noGlobal` & `jQuery` globals in the built file which is not needed.
abnud1 [Sun, 11 Nov 2018 04:52:13 +0000 (06:52 +0200)]
Build: Update most dependencies
The only packages not fully updated are:
- qunitjs & karma-qunit as that's a QUnit 2.x update that will require some
changes and we'll do that later
- jsdom as we need to first rewrite the test with the Symbol polyfill - newer
jsdom versions don't work with such a hacked Symbol instance
- sinon as the v2 -> v7 upgrade requires to update our unit tests
- uglify-js & grunt-contrib-uglify as latest uglify-js versions slightly worsen
the minified gzipped size
Build: Advise to create test cases on JS Bin or CodePen, drop JSFiddle
JSFiddle doesn't support IE (even 11) anymore so we shouldn't advise users
to use it to create test cases. To make people have a choice, add CodePen
to the list.
Also, link to specific starter templates so that novices don't need to spend
time thinking how to set up the basic structure.
Manipulation: Respect script nomodule attribute in DOM manipulation
PR #3869 added support for `<script type="module">` & some support for
the `nomodule` attribute but with no tests for `nomodule` and with the
attribute only respected on inline scripts. This commit adds support for
source-based scripts as well. It also adds tests for `nomodule`, including
making sure legacy browsers execute such scripts as they'd natively do - that's
the whole point of `nomodule` scripts, after all.
Tests: Skip nonce tests in old iOS/Android as well
Old iOS & Android Browser versions support script-src but not nonce, making the
nonce test impossible to run. Browsers not supporting CSP at all are not
a problem as they'll skip script-src restrictions completely.
Andrei Fangli [Mon, 26 Nov 2018 17:00:41 +0000 (19:00 +0200)]
Ajax: Fix getResponseHeader(key) for IE11
- getResponseHeader(key) combines all header values for the provided key into a
single result where values are concatenated by ', '. This does not happen for
IE11 since multiple values for the same header are returned on separate lines.
This makes the function only return the last value of the header for IE11.
- Updated ajax headers test to better cover Object.prototype collisions
CSS: Don't read styles.position in the width/height cssHook unless necessary
Current width/height cssHook reads the computed position style even if not
necessary as the browser passes the scrollboxSize support test. That has been
changed.
This commit also makes the scrollboxSize support test in line with all others
(i.e. only return true or false) and changes the variable name in the hook
to make the code clearer.
Tests: Account for the iPad with iOS 11.3 user agent
The user agent of the iPad with iOS 11.3 on BrowserStack is missing the "iPhone"
part in the "iPhone OS 11_3" part. This commit makes the iOS regex accept such
(probably?) malformed UAs.
Test: enable a spec testing CSS whitespace preserving in Edge 17
In Edge 14-16 setting a style property to a whitespace-only value resets it to
the default, forcing us to skip a relevant CSS test in Edge. Now that Edge 17
has fixed the issue we can re-enable this test there.