docs/slate-issues/open-issues-dossiers/5064-4971.md
JustWaltersbug0With certain CSS display values, the editor treats clicks outside its visible border as in-bounds clicks.
No thread yet, but the report points at a crisp DOM/layout seam rather than app misuse.
Strong enough.
Acceptable.
Likely valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it in the DOM boundary bucket unless someone proves it is pure CSS misuse.
Indirect.
Ready with minor setup.
tanvirchahalbug4slate-html-serializer no longer works with modern Slate because that serializer line is effectively dead and unsupported.
The first comment already says the serializer has been unsupported for years.
Strong enough.
Strong.
Invalid.
This is mostly unsupported or invalid current-contract behavior.
close-invalid
Close it as unsupported ecosystem drift, not as current Slate breakage.
None.
Not a direct test candidate.
beforeInsertText function to accept/reject text insertionmiquikfeature7Users want a beforeInsertText seam so they can reject input before the DOM and Slate state drift apart.
Maintainer feedback in the thread is positive, which makes this a real API seam request rather than idle ideation.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
v2-roadmap
Keep it as API roadmap input tied to the broader input contract story.
Direct.
Not a direct test candidate.
pooya-dehghanbug, ⚑ cross platform6This is a vague framework integration failure report around Shopify and Next.js with a null crash, but the issue is too underspecified to drive architecture.
The follow-ups are still 'it crashes for me' level, not enough to pin ownership.
Weak.
Poor.
Unclear.
Keep it in the right cluster and do not let it drift into unrelated themes.
ask-for-repro
Ask for a reduced repro or de-weight it.
None.
Blocked on a tighter repro.
laufeyrutimprovement0This is really a behavior request: hitting backspace at the start of the document should clear element formatting instead of doing nothing.
No thread yet, and the ask is really about preferred editor behavior, not a broken core invariant.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
share-status
Do not let product-behavior preference drive the core roadmap.
None.
Not a direct test candidate.
loycoderbug, ⚑ cross platform0On Android, toggling readOnly clears editor.selection, which breaks read-only inspection and selection persistence.
No thread yet, but the report is concrete and sits in an already-real Android lifecycle seam.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it in the Android lifecycle cluster.
Direct.
Ready with minor setup.
ggukggukeebug3This error comes from Prism or similar code-highlighting DOM mutation that Slate does not own, so it is mostly a consumer mutation problem.
The first comment already explains the ownership problem: Prism rewrites DOM behind Slate’s back.
Strong enough.
Strong.
Likely invalid.
This is mostly unsupported or invalid current-contract behavior.
close-invalid
Point at the DOM mutation contract and close it.
None.
Not a direct test candidate.
jpunclebug0Editor.above on an expanded range looks wrong according to the supplied example, which makes this a real API semantics question.
No thread yet, but the example is sharp enough to turn directly into a behavior test.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it as a small but real core API correctness issue.
Direct.
Ready now.
pass argsgland2015feature0Editor.nodes needs a pass predicate so callers can skip whole subtrees more explicitly during traversal.
No thread yet. The ask is clean and bounded.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
v2-roadmap
Keep it as API-surface roadmap input.
Direct.
Not a direct test candidate.
jbrew138bug, ⚑ cross platform1Changing readOnly on Android breaks the native listener wiring, so insert/remove ops never reach onChange.
The comment already points at #4994, which is the right duplicate target.
Strong enough.
Poor.
Valid.
Treat this as a duplicate of #4994.
mark-duplicate
Mark it duplicate and keep the cluster concentrated.
Direct.
Ready with minor setup.
Transform.selectmlajtosimprovement0Transforms.select happily accepts invalid locations, so the request is for a safer public contract or sanity check.
No thread yet, but the contract complaint is straightforward.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
v2-roadmap
Treat it as API-hardening roadmap input.
Direct.
Not a direct test candidate.
Hhponbug, ⚑ cross platform6Chinese IME repeats inserted content because composition and beforeinput paths both end up inserting text.
The thread is mostly people asking for a fix, which reinforces that the seam is real but unresolved.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it in the CJK composition cluster.
Direct.
Ready with minor setup.
YuriyMorkovnikbug, ⚑ cross platform1On Android with Gboard, the last composed word does not reach Slate state until a later trigger like space or punctuation.
The comment already points at active work in #4988, which is enough to collapse this into that line.
Strong enough.
Poor.
Valid.
Treat this as a duplicate of #4988.
mark-duplicate
Point at the active duplicate instead of carrying another Android flush issue separately.
Direct.
Ready with minor setup.
bingluenbug, ⚑ cross platform1Firefox plus CJK IME can duplicate input and even crash, which is another severe browser-specific composition failure.
The thread is thin, but the severity and reproducibility are clear enough.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it as part of the Firefox/CJK cluster, not generic IME mush.
Direct.
Ready with minor setup.
joshmarnoldbug1Every edit logs a useLayoutEffect SSR warning, which is really a React server-rendering/runtime integration problem.
The thread wanders into Plate and SSR discussion, but the underlying useLayoutEffect warning is still real.
Strong enough.
Acceptable.
Likely valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it in the React runtime pile, but do not confuse it with core data-model pressure.
Indirect.
Blocked on a tighter repro.
insertText causes content to be out of sync with node representationdavisg123bug8Overriding insertText to reject input can leave the DOM and Slate model out of sync, which exposes a bad plugin seam around suppressed input.
The thread includes a workaround, but it proves the current seam is leaky rather than safe.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it as a serious plugin/runtime seam bug.
Direct.
Ready now.
itsameandreabug0Copying only a link loses the link metadata, which means inline-only clipboard extraction is not preserving marks or element semantics correctly.
No thread yet, but the behavior claim is concrete enough to test.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it as a real clipboard/inline boundary bug.
Indirect.
Ready now.
1208286395bug0onFocus fires even when the editor should still be considered blurred, which makes focus event semantics unreliable.
No thread yet, but event ordering bugs usually need a direct test rather than more debate.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it as focus contract work.
Indirect.
Ready now.
YaoKaiLunbug1When a path lookup fails, Slate throws a misleading circular-JSON error instead of the actual path failure, which is plain error-reporting debt.
The comment does not really resolve it; it just asks the reporter to look elsewhere.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Treat it as error-surface cleanup, not as invalid usage.
Indirect.
Ready now.
abeebridwanbug0The images example does not guard the cancel path from window.prompt, so the example inserts junk UI state.
No thread needed; this is an example-quality bug.
Strong enough.
Strong.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Fix the example and move on. It should not affect v2 architecture thinking.
None.
Not a direct test candidate.
scrollSelectionIntoView doesn't allow to completely customize auto-scrolling behaviore1himselfbug1scrollSelectionIntoView claims to be customizable, but it still forces built-in scrolling behavior that callers cannot fully override.
The thread confirms people are using it for sticky toolbars and still hitting the limitation.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
v2-roadmap
Treat it as selection bridge API debt, not just app styling fallout.
Direct.
Not a direct test candidate.
belianikovbug, ⚑ cross platform2AndroidEditable does not react when readOnly changes, so the editor can appear editable while Slate state stays frozen.
The thread confirms it even repros from desktop devtools Android emulation, which is useful extra signal.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it as the anchor issue for this Android lifecycle seam.
Direct.
Ready with minor setup.
jim-kk-hcbug0Horizontal arrow navigation can move the caret two positions at a time instead of one.
No thread yet, but the recording claim is straightforward.
Strong enough.
Poor.
Likely valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it in the selection/caret bucket.
Indirect.
Ready now.
JosNunbug0Selecting across a nested editor can throw because toDOMPoint crosses editor ownership boundaries incorrectly.
No thread yet, but nested editor ownership bugs are high leverage and the seam is clear.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open as real nested-editor selection debt.
Direct.
Ready now.
codebyoulbug1The editor throws if a text node is null, which is invalid document shape rather than a valid Slate state.
The comment confirms it happens, but the state is still invalid input, not a healthy Slate value.
Strong enough.
Acceptable.
Likely invalid.
This is mostly unsupported or invalid current-contract behavior.
close-invalid
Point at the document contract and close it unless a normalization gap is demonstrated.
None.
Not a direct test candidate.