Event: Increase robustness of an inner native event in leverageNative
In Firefox, alert displayed just before blurring an element dispatches
the native blur event twice which tripped the jQuery logic if a jQuery blur
handler was not attached before the trigger call.
This was because the `leverageNative` logic part for triggering first checked if
setup was done before (which, for example, is done if a jQuery handler was
registered before for this element+event pair) and - if it was not - added
a dummy handler that just returned `true`. The `leverageNative` logic made that
`true` then saved into private data, replacing the previous `saved` array. Since
`true` passed the truthy check, the second native inner handler treated `true`
as an array, crashing on the `slice` call.
The same issue could happen if a handler returning `true` is attached before
triggering. A bare `length` check would not be enough as the user handler may
return an array-like as well. To remove this potential data shape clash, capture
the inner result in an object with a `value` property instead of saving it
directly.
Since it's impossible to call `alert()` in unit tests, simulate the issue by
replacing the `addEventListener` method on a test button with a version that
calls attached blur handlers twice.
Tests: Make the beforeunload event tests work regardless of extensions
Some browser extensions, like React DevTools, send messages to the content area.
Since our beforeunload event test listens for all messages, it used to catch
those as well, failing the test.
Add a `source` field to the payload JSON and check for it before treating the
message as coming from our own test to make sure the test passes even with such
browser extensions installed.
Liam James [Fri, 19 Apr 2024 13:47:52 +0000 (23:47 +1000)]
Offset: Increase search depth when finding the 'real' offset parent
Changes:
* Increase search depth when finding for the real offset parent
* Ignore offset for statically positioned offset parent
* Add tests for the position of an element in a table
Tests: share queue/browser handling for all worker types
- one queue to rule them all: browserstack, selenium, and jsdom
- retries and hard retries are now supported in selenium
- selenium tests now re-use browsers in the same way as browserstack
Attributes: Make `.attr( name, false )` remove for all non-ARIA attrs
The HTML spec defines boolean attributes:
https://html.spec.whatwg.org/#boolean-attributes
that often correlate with boolean properties. If the attribute is missing, it
correlates with the `false` property value, if it's present - the `true`
property value. The only valid values are an empty string or the attribute name.
jQuery tried to be helpful here and treated boolean attributes in a special way
in the `.attr()` API:
1. For the getter, as long as the attribute was present, it was returning the
attribute name lowercased, ignoring the value.
2. For the setter, it was removing the attribute when `false` was passed;
otherwise, it was ignoring the passed value and set the attribute -
interestingly, in jQuery `>=3` not lowercased anymore.
The problem is the spec occasionally converts boolean attributes into ones with
additional attribute values with special behavior - one such example is the new
`"until-found"` value for the `hidden` attribute. Our setter normalization
means passing those values is impossible with jQuery. Also, new boolean
attributes are introduced occasionally and jQuery cannot easily add them to the
list without incurring breaking changes.
This patch removes any special handling of boolean attributes - the getter
returns the value as-is and the setter sets the provided value.
To provide better backwards compatibility with the very frequent `false` value
provided to remove the attribute, this patch makes `false` trigger attribute
removal for ALL non-ARIA attributes. ARIA attributes are exempt from the rule
since many of them recognize `"false"` as a valid value with semantics different
than the attribute missing. To remove an ARIA attribute, use `.removeAttr()` or
pass `null` as the value to `.attr()` which doesn't have this exception.
Fixes gh-5388
Closes gh-5452
Co-authored-by: Richard Gibson <richard.gibson@gmail.com>
Core: Fix the exports setup to make bundlers work with ESM & CommonJS
We cannot pass a single file via the `module` condition as then
`require( "jquery" )` will not return jQuery but instead the module object
with `default`, `$` & `jQuery` as keys. Instead:
1. For Node.js, detected via the `node` condition:
1. Expose a regular CommonJS version to `require`
2. Expose a tiny wrapper over CommonJS to `import`
2. For bundlers, detected via the `module` condition:
1. Expose a regular ESM version to `import`
2. Expose a tiny wrapper over ESM to `require`
3. If neither Node.js nor bundlers are detected (no `node` or `module`
conditions`):
1. Expose a regular CommonJS version to `require`
2. Expose a regular ESM version to `import`
The reasons for such definitions are as follows:
1. In Node.js, one can synchronously import from a CommonJS file inside of
an ESM one but not vice-versa. To use an ESM file in a CommonJS one,
a dynamic import is required and that forces asynchronicity.
2. In some bundlers CommonJS is not necessarily enabled - e.g. in Rollup without
the CommonJS plugin. Therefore, the ESM version needs to be pure ESM.
However, bundlers allow synchronously calling `require` on an ESM file. This
is possible since bundlers merge the files before they are passed to
the browser to execute and the final bundles no longer contain async import
code.
3. Bare ESM & CommonJS versions are provided to non-Node non-bundler
environments where we cannot assume interoperability between ESM & CommonJS
is supported.
4. Bare versions cannot be supplied to Node or bundlers as projects using both
ESM & CommonJS to fetch jQuery would result in duplicate jQuery instances,
leading to increased JS size and disjoint data storage.
In addition to the above changes, the `script` condition has been dropped. Only
Webpack documents this condition and it's not clear when exactly it's triggered.
Adding support for a new condition can be added later without a breaking change;
removing is not so easy.
The `production` & `development` conditions have been removed as well. They were
not really applied correctly; we'd need to provide both of them to each current
leaf which would double the size of the definition for the `.` & `./slim` entry
points. In jQuery, the only difference between development & production builds
is minification; there are no logic changes so we can pass unminified versions
to all the tooling, expecting minification down the line.
As for the factory entry points:
1. Node.js always gets the CommonJS version
2. Bundlers always get the ESM version
3. Other tools take the ESM version when using `import` and the CommonJS when
using `require`.
The complexity is lower than for the `.` & `./slim` entry points because there's
no default export to handle so Node/bundler wrapper files are not necessary.
Other changes:
* Tests: Change "node:assert" to "node:assert/strict"; the former is deprecated
* Docs: Mention that the CommonJS module doesn't expose named exports
* Tests: Run Node & bundler tests for all the above cases
Timmy Willison [Mon, 11 Mar 2024 14:39:38 +0000 (10:39 -0400)]
Tests: add --hard-retries option to test runner
- Add the ability to retry by restarting the worker and
getting a different browser instance, after all
normal retries have been exhausted. This can sometimes
be successful when a refresh is not.
Timmy Willison [Tue, 5 Mar 2024 19:44:01 +0000 (14:44 -0500)]
Tests: reuse browser workers in BrowserStack tests (#5428)
- reuse BrowserStack workers.
- add support for "latest" and "latest-1" in browser version filters
- add support for specifying non-final browser versions, such as beta versions
- more accurate eslint for files in test/runner
- switched `--no-isolate` command flag to `--isolate`. Now that browser instances are shared, it made more sense to me to default to no isolation unless specified. This turned out to be cleaner because the only place we isolate is in browserstack.yml.
- fixed an issue with retries where it wasn't always waiting for the retried test run
- enable strict mode in test yargs command
Timmy Willison [Mon, 26 Feb 2024 14:42:10 +0000 (09:42 -0500)]
Tests: migrate testing infrastructure to minimal dependencies
This is a complete rework of our testing infrastructure. The main goal is to modernize and drop deprecated or undermaintained dependencies (specifically, grunt, karma, and testswarm). We've achieved that by limiting our dependency list to ones that are unlikely to drop support any time soon. The new dependency list includes:
- `qunit` (our trusty unit testing library)
- `selenium-webdriver` (for spinning up local browsers)
- `express` (for starting a test server and adding middleware)
- express middleware includes uses of `body-parser` and `raw-body`
- `yargs` (for constructing a CLI with pretty help text)
- BrowserStack (for running each of our QUnit modules separately in all of our supported browsers)
- `browserstack-local` (for opening a local tunnel. This is the same package still currently used in the new Browserstack SDK)
- We are not using any other BrowserStack library. The newest BrowserStack SDK does not fit our needs (and isn't open source). Existing libraries, such as `node-browserstack` or `browserstack-runner`, either do not quite fit our needs, are under-maintained and out-of-date, or are not robust enough to meet all of our requirements. We instead call the [BrowserStack REST API](https://github.com/browserstack/api) directly.
## BrowserStack Runner
- automatically retries individual modules in case of test failure(s)
- automatically attempts to re-establish broken tunnels
- automatically refreshes the page in case a test run has stalled
- runs all browsers concurrently and uses as many sessions as are available under the BrowserStack plan. It will wait for available sessions if there are none.
- supports filtering the available list of browsers by browser name, browser version, device, OS, and OS version (see `npm run test:unit -- --list-browsers` for more info). It will retrieve the latest matching browser available if any of those parameters are not specified.
- cleans up after itself (closes the local tunnel, stops the test server, etc.)
- Requires `BROWSERSTACK_USERNAME` and `BROWSERSTACK_ACCESS_KEY` environment variables.
## Selenium Runner
- supports running any local browser as long as the driver is installed, including support for headless mode in Chrome, FF, and Edge
- supports running `basic` tests on the latest [jsdom](https://github.com/jsdom/jsdom#readme), which can be seen in action in this PR (see `test:browserless`)
- Node tests will run as before in PRs and all non-dependabot branches, but now includes tests on real Safari in a GH actions macos image instead of playwright-webkit.
- can run multiple browsers and multiple modules concurrently
Other notes:
- Stale dependencies have been removed and all remaining dependencies have been upgraded with a few exceptions:
- `sinon`: stopped supporting IE in version 10. But, `sinon` has been updated to 9.x.
- `husky`: latest does not support Node 10 and runs on `npm install`. Needed for now until git builds are migrated to GitHub Actions.
- `rollup`: latest does not support Node 10. Needed for now until git builds are migrated to GitHub Actions.
- BrowserStack tests are set to run on each `main` branch commit
- `debug` mode leaves Selenium browsers open whether they pass or fail and leaves browsers with test failures open on BrowserStack. The latter is to avoid leaving open too many sessions.
- This PR includes a workflow to dispatch BrowserStack runs on-demand
- The Node version used for most workflow tests has been upgraded to 20.x
- updated supportjQuery to 3.7.1
Run `npm run test:unit -- --help` for CLI documentation
Node.js 20 started throwing errors when `writeHead` is called twice on
a response. This might have already been invalid before but it wasn't throwing
on Node.js 18.
Compute the headers object and call `writeHead` once to avoid the issue.
Timmy Willison [Tue, 6 Feb 2024 14:53:30 +0000 (09:53 -0500)]
Release: use buildDefaultFiles directly and pass version
- also add the ability to pass VERSION in env to test final builds
- adjust sha regex to account for lack of shas
- set the version on the dist package.json
The `attrHooks` entries for boolean attributes are only defined for jQuery 4+;
jQuery 3.x used a separate mechanism - assigning them to
`jQuery.expr.attrHandle`. That object used to be maintained by Sizzle, since
jQuery 3.7.0 it's kept in the selector module. Because of that, the `isXMLDoc`
check used to be require in this hook.
Now that standard `attrHooks` are used, the `isXMLDoc` check already happens
inside of `jQuery.attr` and there's no need to repeat it in the test. Note that
this repetition is even incorrect - while Sizzle's `jQuery.find.attr` used to
treat an `undefined` output of the hooks from `jQuery.expr.attrHandle` as a way
to opt out of the hook, jQuery's `attrHooks` use `null` to opt out of a getter
hook.
Apart from the size, this patch also avoids unnecessary extra checks.
Selector: Eliminate `selector.js` depenencies from various modules
There are two main reasons for why some of those dependencies are no longer
needed:
1. `jQuery.contains` which is now a part of `core`.
2. `jQuery.find.attr` no longer exists, native `getAttribute` is used instead.
Build was already happening in scripts like `test:browser` but those scripts
were missing `pretest`, meaning that running `npm install && npm test:browser`
may have failed if `pretest` wasn't run before or if its results were out of
date.
Even worse, with such stale data some tests may erroneously succeed.
This also removes a separate `pretest` step from GitHub Actions as it's no
longer needed.
The package README used to show examples importing from a regular jQuery file;
this won't work natively. Instead, use module versions of jQuery in these
examples.
CSS:Selector: Align with 3.x, remove the outer `selector.js` wrapper
Bring some changes from `3.x-stable`:
* rename `rtrim` to `rtrimCSS` to distinguish from the previous `rtrim`
regex used for `jQuery.trim`
* backport one `id` selector test that avoids the selector engine path
Other changes:
* remove the inner function wrapper from `selector.js` by renaming
the imported `document.js` value
* use `jQuery.error` in `selectorError`
* make Selector tests pass in all-modules runs by fixing a sinon mistake
in Core tests - Core tests had a spy set up for `jQuery.error` that wasn't
cleaned up, influencing Selector tests when all were run together
Core: Simplify code post browser support reduction
Summary of the changes:
* Core: Simplify code post browser support reduction
* Tests: Remove legacy jQuery.cache & oldIE leftovers
* Tests: Reformat JavaScript in delegatetest.html
* Docs: "jQuery Foundation Projects" -> "jQuery Projects"
* Tests: Drop an unused localfile.html file (modern browsers don't support
the `file:` protocol this way, there's no point in keeping the file around)
* Effects: Remove a redundant `!fn` check (`fn || !fn && easing` is equivalent
to `fn || easing`; simplify the code)
* CSS: Explain the fallback to direct object access in curCSS better
* Tests: Deduplicate `jQuery.parseHTML` test titles
* Dimensions: Add a test for fractional values
* Tests: Fix a buggy WebKit regex
Since versions 1.11.0/2.1.0, jQuery has used a module wrapper with one strange
addition - in CommonJS environments, if a global `window` with a `document` was
not present, jQuery exported a factory accepting a `window` implementation and
returning jQuery.
This approach created a number of problems:
1. Properly typing jQuery would be a nightmare as the exported value depends on
the environment. In practice, typing definitions ignored the factory case.
2. Since we now use named exports for the jQuery module version, it felt weird
to have `jQuery` and `$` pointing to the factory instead of real jQuery.
Instead, for jQuery 4.0 we leverage the just added `exports` field in
`package.json` to expose completely separate factory entry points: one for the
full build, one for the slim one.
Exports definitions for `./factory` & `./factory-slim` are simpler than for `.`
and `./slim` - this is because it's a new entry point, we only expose a named
export and so there's no issue with just pointing Node.js to the CommonJS
version (we cannot use the module version for `import` from Node.js to avoid
double package hazard). The factory entry points are also not meant for the Web
browser which always has a proper `window` - and they'd be unfit for an
inclusion in a regular script tag anyway. Because of that, we also don't
generate minified versions of these entry points.
The factory files are not pushed to the CDN since they are mostly aimed
at Node.js.
Bumps [socket.io-parser](https://github.com/socketio/socket.io-parser)
and [socket.io](https://github.com/socketio/socket.io). These dependencies
needed to be updated together.
Updates `socket.io-parser` from 4.0.5 to 4.2.4
- [Release notes](https://github.com/socketio/socket.io-parser/releases)
- [Changelog](https://github.com/socketio/socket.io-parser/blob/main/CHANGELOG.md)
- [Commits](https://github.com/socketio/socket.io-parser/compare/4.0.5...4.2.4)
Updates `socket.io` from 4.5.1 to 4.7.2
- [Release notes](https://github.com/socketio/socket.io/releases)
- [Changelog](https://github.com/socketio/socket.io/blob/main/CHANGELOG.md)
- [Commits](https://github.com/socketio/socket.io/compare/4.5.1...4.7.2)
Bumps [qs](https://github.com/ljharb/qs) from 6.5.2 to 6.5.3.
- [Changelog](https://github.com/ljharb/qs/blob/main/CHANGELOG.md)
- [Commits](https://github.com/ljharb/qs/compare/v6.5.2...v6.5.3)
- lint
- npmcopy
- build, minify, and process for distribution.
- new custom build command using yargs
- compare size of minified/gzip built files
- pretest scripts, including qunit-fixture, babel transpilation, and npmcopy
- node smoke tests
- promises aplus tests
- new watch task using `rollup.watch` directly
Also:
- upgraded husky and added the new lint command
- updated lint config to use new "flat" config format. See https://eslint.org/docs/latest/use/configure/configuration-files-new
- Temporarily disabled one lint rule until flat config is supported by eslint-plugin-import. See https://github.com/import-js/eslint-plugin-import/issues/2556
- committed package-lock.json
- updated all test scripts to use the new build
- added an express test server that uses middleware-mockserver (this can be used to run tests without karma)
- build-all-variants is now build:all
The `default` export is treated differently across tooling when transpiled
to CommonJS - tools differ on whether `module.exports` represents the full
module object or just its default export. Switch `src/` modules to named
exports for tooling consistency.
Build: Add `exports` to package.json, export slim & esm builds
Summary of the changes:
* define the `exports` field in `package.json`; `jQuery` & `$` are also
exported as named exports in ESM builds now
* declare `"type": "module"` globally except for the `build` folder
* add the `--esm` option to `grunt custom`, generating jQuery as an ECMAScript
module into the `dist-module` folder
* expand `node_smoke_tests` to test the slim & ESM builds and their various
combinations; also, test both jQuery loaded via a path to the file as well
as from module specifiers that should be parsed via the `exports` feature
* add details about ESM usage to the release package README
* run `compare_size` on all built minified files; don't run it anymore on
unminified files where they don't provide lots of value
* remove the remove_map_comment task; SWC doesn't insert the
`//# sourceMappingURL=` pragma by default so there's nothing to strip
That threw our `reliableTrDimensions` support test off. This change fixes the
support test and adds a unit test ensuring support test values on a page
including Bootstrap 5 CSS are the same as on a page without it.
With this change, jQuery build no longer generates the `amd` directory with
AMD modules transpiled from source `src` ECMAScript Modules. To use individual
jQuery modules from source, ESM is now required.
Note that this DOES NOT affect the main `"jquery"` AMD module defined by built
jQuery files; those remain supported.
Tests: Disable the ":lang respects escaped backslashes" test
Firefox 114+ no longer match on backslashes in `:lang()`, even when escaped.
It is an intentional change as `:lang()` parameters are supposed to be valid
BCP 47 strings. Therefore, we won't attempt to patch it.
We'll keep this test here until other browsers match the behavior.
`Sizzle.tokenize` is an internal Sizzle API, but exposed. As a result,
it has historically been available in jQuery via `jQuery.find.tokenize`.
That got dropped during Sizzle removal; this change restores the API.
Some other APIs so far only exposed on the `3.x` line are also added
back:
* `jQuery.find.select`
* `jQuery.find.compile`
* `jQuery.find.setDocument`
In addition to that, Sizzle tests have been backported for the following
APIs:
* `jQuery.find.matchesSelector`
* `jQuery.find.matches`
* `jQuery.find.compile`
* `jQuery.find.select`
A new test was also added for `jQuery.find.tokenize` - even Sizzle was
missing one.
Build: Switch the minifier from UglifyJS to Terser
UglifyJS is ES5-only, while Terser supports newer ECMAScript versions. jQuery
is authored in ES5 but jQuery 4.x will also have an ESM build that cannot be
minified using UglifyJS directly.
We could strip the `export` statement, minify via UglifyJS and re-add one but
that increases complexity & may not fully play nice with source maps.
On the other hand, switching to Terser increases the minfied size by just 324
bytes and the minified gzipped one by just 70 bytes. Such differences largely
disappear among bigger size gains from the `3.x-stable` line - around 2.7 KB
minified gzipped as of now.
Docs: Remove the "Grunt build" section from the PR template
Now that unit tests are run on GitHub Actions in all three major
engines and for multiple custom jQuery builds, the request for PR
authors to run unit tests locally and confirm they pass is needless
overhead; let's drop the checkbox.
Build: Make the `eslint:dev` task not lint the `dist/` folder
There was a mistake in paths logic that made the `dist/` folder linted
even in the `eslint:dev` task which is run before the build. Fix that by
explicitly ignoring the `dist/` folder at the end of the file list.
Tests: Indicate Chrome 112 & Safari 16.4 pass the cssHas support test
Chrome 112 & Safari 16.4 introduce two changes:
* `:has()` is non-forgiving
* `CSS.supports( "selector(...)" )` parses everything in a non-forgiving way
We no longer care about the latter but the former means the `cssHas` support
test now passes.
CSS: Make `offsetHeight( true )`, etc. include negative margins
This regressed in gh-3656 as the added logic to include scroll gutters
in `.innerWidth()` / `.innerHeight()` didn't take negative margins into
account. This broke handling of negative margins in
`.offsetHeight( true )` and `.offsetWidth( true )`. To fix it, calculate
margin delta separately and only add it after the scroll gutter
adjustment logic.
Event: Simplify the check for saved data in leverageNative
Previously, when `leverageNative` handled async events, there was
a case where an empty placeholder object was set as a result.
Covering both such an object and `false` required a `length` check.
However, this is not necessary since gh-5223 and the check was
already simplified in other places; this one was missed.
Event: Make trigger(focus/blur/click) work with native handlers
In `leverageNative`, instead of calling `event.stopImmediatePropagation()`
which would abort both native & jQuery handlers, set the wrapper's
`isImmediatePropagationStopped` property to a function returning `true`.
Since for each element + type pair jQuery attaches only one native handler,
there is also only one wrapper jQuery event so this achieves the goal:
on the target element jQuery handlers don't fire but native ones do.
Unfortunately, this workaround doesn't work for handlers on ancestors
- since the native event is re-wrapped by a jQuery one on each level of
the propagation, the only way to stop it for jQuery was to stop it for
everyone via native `stopPropagation()`. This is not a problem for
`focus`/`blur` which don't bubble, but it does also stop `click` on
checkboxes and radios. We accept this limitation.
Event: Simulate focus/blur in IE via focusin/focusout
In IE (all versions), `focus` & `blur` handlers are fired asynchronously
but `focusin` & `focusout` are run synchronously. In other browsers, all
those handlers are fired synchronously. Asynchronous behavior of these
handlers in IE caused issues for IE (gh-4856, gh-4859).
We now simulate `focus` via `focusin` & `blur` via `focusout` in IE to avoid
these issues. This also let us simplify some tests.
This commit also simplifies `leverageNative` - with IE now using `focusin`
to simulate `focus` and `focusout` to simulate `blur`, we don't have to deal
with async events in `leverageNative`. This also fixes broken `focus` triggers
after first triggering it on a hidden element - previously, `leverageNative`
assumed that the native `focus` handler not firing after calling the native
`focus` method meant it would be handled later, asynchronously, which
was not the case (gh-4950).
PR gh-5197 started treating all non-string non-plain-object
`data` values as binary. However, `jQuery.ajax` also supports
arrays as values of `data`. This change makes regular arrays
no longer be considered binary data.
Surprisingly, we had no tests for array `data` values; otherwise,
we'd detect the issue earlier. This change also adds
a few such missing tests.
Build: Only install Playwright dependencies when needed
PR gh-5190 added support for running tests on Playwright WebKit
in CI. For efficiency reasons, Playwright dependencies are only
installed for the `test:browser` npm script. However, that same
script is also used for Firefox ESR testing.
This change makes Playwright dependencies installed only for cases
where `WebKitHeadless` exists on the list of tested browsers.
Ajax: Allow `processData: true` even for binary data
The way gh-5197 implemented binary data handling, `processData`
was being explicitly set to `false`. This is expected but it made
it impossible to override it to `true`. The new logic will only
set `processData` to `false` if it wasn't explicitly passed
in original options.
Rename `jQuery.Deferred.getStackHook` to `jQuery.Deferred.getErrorHook`
to indicate passing an error instance is usually a better choice - it
works with source maps while a raw stack generally does not.
In jQuery `3.7.0`, we'll keep both names, marking the old one as
deprecated. In jQuery `4.0.0` we'll just keep the new one. This
change implements the `4.0.0` version; PR gh-5212 implements
the `3.7.0` one.
Selector: Stop relying on CSS.supports( "selector(...)" )
`CSS.supports( "selector(...)" )` has different semantics than selectors passed
to `querySelectorAll`. Apart from the fact that the former returns `false` for
unrecognized selectors and the latter throws, `qSA` is more forgiving and
accepts some invalid selectors, auto-correcting them where needed - for
example, mismatched brackers are auto-closed. This behavior difference is
breaking for many users.
To add to that, a recent CSSWG resolution made `:is()` & `:where()` the only
pseudos with forgiving parsing; browsers are in the process of making `:has()`
parsing unforgiving.
Taking all that into account, we go back to our previous try-catch approach
without relying on `CSS.supports( "selector(...)" )`. The only difference
is we detect forgiving parsing in `:has()` and mark the selector as buggy.
The PR also updates `playwright-webkit` so that we test against a version
of WebKit that already has non-forgiving `:has()`.
projects / jquery.git / log