jquery-4-migration
Skill0 starsChanged 24 days agoMITView on GitHub
Migrate a Drupal theme's JavaScript from jQuery 3 to jQuery 4 (as shipped by Drupal 11), on a Dockerized Makefile-driven project. Use when the user mentions jQuery 3→4, jQuery 4 deprecated/removed APIs ($.trim, $.isFunction, $.isArray, $.type, .bind/.delegate…), a jquery-deprecated / migrate polyfill, or "$.x is not a function" after an upgrade. Covers scanning custom + contrib JS, native conversions, rebuilding assets, and a scoped local polyfill for not-yet-compatible third-party libraries. IMPORTANT: all builds MUST go through Makefile/npm targets, never on the host.
In this repository17 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- d11Skill
- drupal-build-foundationsSkill
- drupal-build-intakeSkill
- drupal-buildSkill
- drupal-build-unitSkill
- drupal-rgaaSkill
- drupal-scaffoldSkill
- gulp-to-viteSkill
- lando-to-dockerSkill
- phpcbfSkill
- phpcsSkill
- php-deprecations-auditSkill
- php-docker-upgradeSkill
- php-standardsSkill
- phpstanSkill
- sass-migratorSkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 4 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 25
Never run npm/gulp directly on the host — use the targets inside the node container:
- Very specific: line 47
Start with (A); remove the polyfill once grep confirms zero remaining usages.
- Distinctive: line 121
Run the scan, list all affected files/lines.
- Distinctive: line 124
Test every JS interaction (modals, sliders, AJAX, forms).
- Distinctive: line 137
After removing the polyfill, drop any explicit dependency on a pinned jQuery version (let D11 core provide jQuery 4):
- Distinctive: line 203
Write the verdict down. "No compatible version exists" is a claim that belongs in the merge request with the four checks behind it, because …
- Very specific: line 217
if the theme's own scripts use the bare global $ (Drupal core runs jQuery.noConflict()), also restore window.$ = window.jQuery
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
jQuery 3 → jQuery 4 migration (Drupal theme)
Drupal 11 ships jQuery 4. All APIs deprecated since jQuery 3.x are
permanently removed. Custom JS (typically under
web/themes/custom/<theme>/src/js/) must be migrated. This works
independently of the core bump: native replacements (Array.isArray,
String.prototype.trim, Date.now…) exist regardless of the jQuery version, so
you can migrate and test while still on Drupal 10 (jQuery 3.x + polyfill).
Golden rule: everything goes through the Makefile
Never run npm/gulp directly on the host — use the targets inside the node
container:
| Need | Make target |
| Build assets | make npm-build |
| Deploy built assets | make gulp-deploy |
| Clear Drupal cache | make cr |
| Shell in a container | make shell |
The polyfill: understand before acting
Many D10 projects include a polyfill re-implementing removed jQuery
functions (jquery-deprecated.js / polyfill.js). Its presence means:
- Theme code likely still depends on these APIs.
- Two strategies:
- (A) Keep the polyfill temporarily, migrate the calling code progressively, remove it at the end. Safest — recommended.
- (B) Migrate everything to native APIs and remove the polyfill at once. Cleaner but riskier.
Start with (A); remove the polyfill once grep confirms zero remaining usages.
Removed API conversion table
| Removed API | Native / jQuery 4 replacement |
$.isFunction(x) / $.fn.isFunction | typeof x === 'function' |
$.type(x) | native typeof, or Array.isArray(), === null, etc. |
$.trim(s) | s.trim() (non-null string) or (s ?? '').trim() |
$.isArray(x) | Array.isArray(x) |
$.isWindow(x) | x != null && x === x.window |
$.nodeName(el, n) | el.nodeName?.toLowerCase() === n.toLowerCase() |
$.isNumeric(x) | !Number.isNaN(parseFloat(x)) && Number.isFinite(x) |
$.now() | Date.now() |
$.parseJSON(s) | JSON.parse(s) |
$.unique(arr) / $.fn.unique | $.uniqueSort(arr) |
$.camelCase(s) | Internal API — replace with your own function |
$.fx.interval | Removed — delete any write to this property |
$.cssProps.float = 'styleFloat' | styleFloat no longer exists — remove |
Other jQuery 4 changes to check
$.ajax:success/error/complete→.done()/.fail()/.always()or promises..bind()/.unbind()/.delegate()/.undelegate()→.on()/.off()..load()/.unload()/.error()event shorthands →.on('load'…).$.isEmptyObject,$.proxy,$.uniqueSort: still shipped by jQuery 4.0.0 ($.proxyis deprecated, not removed). Verify in the core file rather than assuming —grep -n 'jQuery\.proxy = ' web/core/assets/vendor/jquery/jquery.js. A library that only calls these needs no shim at all.jQuery.Deferred: behaviour aligned with native Promises.hoverpseudo-events →mouseenter/mouseleave.
Scanning custom JS
THEME_JS=web/themes/custom/*/src/js
grep -rn '\$\.isFunction\|\.isFunction(' $THEME_JS
grep -rn '\$\.trim\|jQuery\.trim' $THEME_JS
grep -rn '\$\.isArray' $THEME_JS
grep -rn '\$\.type(' $THEME_JS
grep -rn '\$\.parseJSON\|jQuery\.parseJSON' $THEME_JS
grep -rn '\$\.now\|jQuery\.now' $THEME_JS
grep -rn '\$\.isWindow' $THEME_JS
grep -rn '\$\.nodeName' $THEME_JS
grep -rn '\$\.isNumeric' $THEME_JS
grep -rn '\$\.unique\b' $THEME_JS
grep -rn '\.bind(\|\.unbind(\|\.delegate(\|\.undelegate(' $THEME_JS
grep -rn 'fx\.interval\|cssProps' $THEME_JSScanning third-party / minified vendor — the grep blind spot
The greps above only cover readable custom source. Minified/bundled vendor
libraries alias jQuery to a local variable — (function(a){… a.type(…) …})(jQuery) —
so $.type / $.isArray etc. appear as a.type( / a.isArray( and the
\$\.-anchored greps miss them entirely. This is exactly how a removed
static like $.type (called by slick-carousel's registerBreakpoints) slips
past the scan and only surfaces at runtime.
Scan the vendor too, on the method token alone (and prefer each lib's unminified source, where the calls are literal):
VENDOR='node_modules/slick-carousel/slick/slick.js web/themes/custom/*/js/*.min.js'
grep -oE '\.(type|isFunction|isArray|isWindow|isNumeric|trim|nodeName|parseJSON|proxy|now|unique|camelCase)\(' $VENDOR | sort | uniq -c
grep -oE '\.(bind|unbind|delegate|undelegate)\(' $VENDOR | sort | uniq -cA clean grep is not proof — aliased minified calls are invisible to it.
The authoritative check is the browser (see the verification note below).
Migration workflow — custom JS
- Run the scan, list all affected files/lines.
- Apply the native conversion from the table for each occurrence.
- Rebuild:
make npm-buildthenmake gulp-deploy. - Test every JS interaction (modals, sliders, AJAX, forms).
- Once
grepshows zero usages, delete the polyfill file and its line in the theme's*.libraries.yml. make crand final re-test.
Verify in a real browser, not just grep. Source greps cannot see aliased minified vendor calls, so load the site under the target jQuery (D11 core = jQuery 4) and watch the console for
TypeError: … is not a function. Confirmwindow.jQuery.fn.jqueryis4.x, that removed statics resolve (typeof window.jQuery.type), and that each plugin still initialises (e.g. a slider gains its.slick-initializedclass). Zero console errors on a clean reload is the real pass signal.
After removing the polyfill, drop any explicit dependency on a pinned jQuery version (let D11 core provide jQuery 4):
grep -rn 'jquery' web/themes/custom/*/*.libraries.yml
grep -rn 'core/jquery' web/themes/custom/*/Third-party & contrib libraries (in parallel)
Custom JS is not the only source of breakage. Third-party libraries declared
in theme .libraries.yml files (slick.js, lightboxes, accordions…) may also
call removed APIs. Per library, in this order: prove no compatible release
exists → update if one does → only then generate a scoped local polyfill
(never pull in core/jquery.migrate, which re-adds the entire deprecated
surface). Also replace external CDN references and any core/modernizr
dependency (removed in D11).
Gate: prove no compatible release exists before writing a shim
A shim is the fallback, never the first move — and the npm registry is not the
source of truth. A package that looks abandoned there can have a live
repository whose jQuery-4 fix ships as a git tag only. npm view <pkg> version
returning a release from years ago proves nothing; it is the single most common
way a project ends up maintaining a polyfill it never needed.
Walk all four levels before concluding "no compatible version exists":
PKG=<npm-package>; REPO=<owner>/<repo>
# 1. registry — what npm/yarn would actually install
npm view "$PKG" version time.modified
# 2. tags and releases — routinely ahead of the registry
curl -s "https://api.github.com/repos/$REPO/tags?per_page=10" | grep '"name"'
curl -s "https://api.github.com/repos/$REPO/releases?per_page=5" | grep '"tag_name"\|"published_at"'
# 3. default-branch HEAD — the fix may not be tagged yet
curl -s "https://api.github.com/repos/$REPO/commits/HEAD" | grep -m1 '"date"'
# 4. active forks — only when upstream is genuinely dead
curl -s "https://api.github.com/repos/$REPO/forks?sort=stargazers&per_page=5" | grep '"full_name"'Then audit the candidate itself, never its changelog: download it and rerun
the token scan. Compatible means the token scan comes back empty of removed
APIs — remaining hits on $.proxy / $.isEmptyObject / $.uniqueSort are fine,
those still ship with jQuery 4.
V=<tag>
curl -sSL -o /tmp/lib.js "https://cdn.jsdelivr.net/gh/$REPO@$V/<path>/lib.js"
grep -oE '\.(type|isFunction|isArray|isWindow|isNumeric|trim|nodeName|parseJSON|now|unique|camelCase|bind|unbind|delegate|undelegate)\(' /tmp/lib.js | sort | uniq -cAlso diff the candidate against the pinned version to size the upgrade, and diff the pinned version against its own upstream release first — a vendored file that was patched locally makes the swap a merge, not a replacement:
curl -sSL -o /tmp/pinned.js "https://cdn.jsdelivr.net/gh/$REPO@<current-tag>/<path>/lib.js"
diff /tmp/pinned.js <theme>/js/lib.js # empty = stock, safe to replace outright
diff /tmp/pinned.js /tmp/lib.js | head -100Write the verdict down. "No compatible version exists" is a claim that belongs in the merge request with the four checks behind it, because it is the claim that justifies shipping a polyfill.
What the scoped shim contains. For a bundled/abandoned plugin you cannot
rewrite (slick, mCustomScrollbar…), re-add only the removed members it
actually calls, each guarded (if (typeof $.fn.bind !== 'function'),
if (typeof $.type !== 'function')…) so it is a no-op under jQuery 3:
- event aliases:
.bind/.unbind/.delegate/.undelegate→.on/.off - static helpers (the ones the source grep misses in minified vendor):
$.type,$.isFunction,$.isArray,$.isNumeric,$.trim,$.isWindow,$.nodeName,$.now,$.parseJSON— faithful jQuery-3 reimplementations ($.typeneeds theclass2typelookup table) - if the theme's own scripts use the bare global
$(Drupal core runsjQuery.noConflict()), also restorewindow.$ = window.jQuery
Declare the shim as its own library depending on core/jquery and
core/drupal (so it runs after noConflict()) and load it before the
plugin bundle — via a dependency, and/or a weight that lands it after
jquery/drupal.init but before the theme bundle. Verify the actual <script>
order in the rendered page.
The full playbook — polyfill catalog with inter-dependencies, how to declare and
attach it in .libraries.yml, CDN / core/modernizr replacement, and the final
verification greps — is in references/polyfill-catalog.md. Load it when you
reach the third-party stage.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

