Issues

30.1.0 regression: removing the focused element makes the next focus() fire a window-targeted blur (breaks Radix Select close-on-blur in tests)

#4347

· bogatykh opened

on Sep 18, 2026

relatedTarget on a focus event is the Document object (not <body> or null) after the previously-focused element is removed from the DOM — regression since 30.1.0

#4344

· rmartins0 opened

on Sep 18, 2026

Quiet MutationObservers retain dead weak-reference bookkeeping

memory

#4340

· domenic opened

on Sep 17, 2026

The cached default stylesheet retains the first style-computation window

memory

#4338

· domenic opened

on Sep 17, 2026

The CSS parser retains the last stylesheet window through its error callback

memory

#4339

· domenic opened

on Sep 17, 2026

Canceled pending interceptors retain closed windows

memory

#4337

· domenic opened

on Sep 17, 2026

Distinct selectors grow an unbounded cache for each live document

memory

#4336

· domenic opened

on Sep 17, 2026

getComputedStyle inheritance walk is O(2^ancestor-depth) for a custom property set to inherit/unset

#4330

· robCToken opened

on Sep 17, 2026

Create implementation exceptions in the correct realm

#4326

· domenic opened

on Sep 17, 2026

Reject negative CSS sizing values after syntax-data updates

css

#4302

· domenic opened

on Sep 11, 2026

Duplicated Content-Type header silently mojibakes the document, or throws

#4277

· rawsun007 opened

on Sep 4, 2026

PointerEvent added in v27 without setPointerCapture/releasePointerCapture/hasPointerCapture — is that API planned?

#4271

· Zeboch opened

on Sep 3, 2026

Navigation Menu

Uh oh!

All issues

30.1.0 regression: removing the focused element makes the next focus() fire a window-targeted blur (breaks Radix Select close-on-blur in tests)

relatedTarget on a focus event is the Document object (not <body> or null) after the previously-focused element is removed from the DOM — regression since 30.1.0

Quiet MutationObservers retain dead weak-reference bookkeeping

The cached default stylesheet retains the first style-computation window

The CSS parser retains the last stylesheet window through its error callback

Canceled pending interceptors retain closed windows

Distinct selectors grow an unbounded cache for each live document

getComputedStyle inheritance walk is O(2^ancestor-depth) for a custom property set to inherit/unset

Create implementation exceptions in the correct realm

Reject negative CSS sizing values after syntax-data updates

Duplicated Content-Type header silently mojibakes the document, or throws

PointerEvent added in v27 without setPointerCapture/releasePointerCapture/hasPointerCapture — is that API planned?

Uh oh!

Search results