docs/slate-issues/open-issues-dossiers/4390-4269.md
linhtrinh18bug1Deleting the paragraph below an image in the official example removes the image too, which is a good concrete example of adjacent-void delete weirdness.
The workaround comment keeps it example-first, but the underlying seam is real enough.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Indirect.
Ready now.
eichenrothimprovement1The Slate component typing is too loose at the top level, which weakens TypeScript where the core model is actually stricter.
This is a clean typing paper-cut with a plausible fix, not a deep architecture fight.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
share-status
Reply with the current scope and route people to the right layer.
Indirect.
Not a direct test candidate.
rohankeskar19bug2Splitting at the end of marked text creates the new block but leaves focus on the old one.
The issue is light, but the symptom is concrete and another user confirmed it years later.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready now.
pampulbug, ⚑ cross platform2Choosing an iOS suggestion can duplicate the inserted word or phrase, which is classic “Slate versus browser text pipeline” pain.
The follow-up ties it to a specific line in Editable, so this is more useful than generic mobile complaining.
Strong enough with minor setup.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready with minor setup.
pampulbug3Selecting all then starting Japanese composition can wedge the editor after the first character.
The thread is good because it links the problem to placeholder/composition behavior and to the earlier Chinese report.
Strong enough with minor setup.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready with minor setup.
churichardbug1Deleting a void at the beginning of a line jumps the caret to the previous line instead of preserving local position.
The thread is short but still exposes the same class of void-boundary selection pain.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready now.
pierre-lgbbug, ⚑ cross platform5Holding Backspace on Android can jump unpredictably across formatting boundaries.
The thread matters because later comments suggest it may have been fixed on main, so this should not keep inflating current pain without revalidation.
Strong enough.
Acceptable.
Stale candidate.
This looks historically real but stale enough that it should not keep inflating current architecture signal without revalidation.
close-stale
Close it as stale unless someone revalidates it on a current build.
Indirect.
Not a direct test candidate.
ghingisbug0Image drag-and-drop inside Shadow DOM still breaks DOM-point resolution or falls back to URL copy behavior.
The issue is concise but the repro is crisp and valuable because Shadow DOM keeps recurring as a real boundary problem.
Strong enough with minor setup.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready with minor setup.
ElizavetaKorolevabug2Pressing Enter after a link can create an empty link in the next paragraph instead of a clean split.
The later confirmation that tab order highlights the empty links makes it more than a cosmetic devtools oddity.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready now.
Transforms.insertNodes when the selection end is void elementjpunclebug1Transforms.insertNodes can blow up when the selection endpoint lands on a void boundary.
The repro is concrete enough even though the follow-up never got to a final root cause.
Strong enough with minor setup.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready with minor setup.
ghingisbug1External value and selection replacement is still too unsafe and too timing-sensitive for document editors that own history outside Slate.
The issue is valuable because it is not just one crash report; it explains the missing lifecycle guarantees the consumer actually needs.
Strong enough with minor setup.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
v2-roadmap
Keep it on the v2/runtime roadmap, not in generic bug triage.
Direct.
Ready with minor setup.
samantha-letourneaunone0This is an old pre-modern-Slate issue and not useful current-contract signal.
There is no thread and the versions are from a different era of Slate.
Strong enough.
None.
Invalid.
This sits outside Slate's current contract or points at an external/browser/framework problem.
close-invalid
Close it and point at the actual owning layer or unsupported scope.
None.
Not a direct test candidate.
daniel-kellerbug3Server-rendered attribute mismatches around spellcheck and autocorrect were real, but the thread later points at a fix path and workaround.
This should stay as historical signal, not live architecture pressure.
Strong enough.
Strong.
Stale candidate.
This looks historically real but stale enough that it should not keep inflating current architecture signal without revalidation.
close-stale
Close it as stale unless someone revalidates it on a current build.
None.
Not a direct test candidate.
skokenesbug0Changing renderLeaf should not synthetically look like a user selection change, but it currently does.
This is a clean, reduced slate-react runtime bug with a good red-test seam.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready now.
LazyJokebug4Selecting all then starting Chinese IME input can wedge the editor, which is the same general class later reported more clearly in Japanese.
The thread is messy, but later comments make it clear the problem persisted and was not really gone.
Strong enough with minor setup.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready with minor setup.
YaoKaiLunimprovement0This is a request to expose an internal placeholder symbol so consumers can avoid duplicating Slate internals for custom placeholder rendering.
There is no thread, so it should stay low priority and low confidence.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
share-status
Reply with the current scope and route people to the right layer.
Indirect.
Not a direct test candidate.
macrozonefeature2Readonly rendering still carries too much editing-oriented DOM baggage, which keeps pushing people toward separate presentation renderers.
The thread even points at a separate presentation package, which is exactly the signal: this is a real runtime split, not one more prop tweak.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
v2-roadmap
Keep it on the v2/runtime roadmap, not in generic bug triage.
Direct.
Not a direct test candidate.
mwood23bug3Reusing the same node object reference across insertions still blows up keying and weakmap-based runtime assumptions.
The workaround is obvious but bad, and the thread ties it to broader weakmap identity problems, so this is high-value architecture signal.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready now.
brucouimprovement10Framework decoupling pressure is real and long-lived, even if the current answer is still “use React or maybe Preact”.
The thread is useful because it shows both demand and the maintainer view of the tradeoff.
Strong enough.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
v2-roadmap
Keep it on the v2/runtime roadmap, not in generic bug triage.
Direct.
Not a direct test candidate.
jamespleasebug13Clicked void nodes and arrow-navigated void nodes still do not behave consistently on Enter and related edit commands.
The thread is excellent because it drills down to the behavioral regression and the likely introducing PR.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready now.
ugrinovskybug, ⚑ cross platform1Editor.marks in Firefox can report the wrong active marks when expanding selection from a marked boundary.
The follow-up confirms others hit the same bug, so it is not one odd local toolbar implementation.
Strong enough with minor setup.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready with minor setup.
aboveyunhaibug3The TSX examples present themselves as TypeScript guidance but still contain unsafe or confusing typing patterns.
The thread is good enough to keep this as real docs/example debt, not just one strict-ts complaint.
Strong enough.
Acceptable.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
share-status
Reply with the current scope and route people to the right layer.
None.
Not a direct test candidate.
lemm17bug, ⚑ cross platform1This is just Firefox inheriting an old browser bug, not a good Slate-owned issue.
The only reply gives the bugzilla target directly.
Strong enough.
Strong.
Invalid.
This sits outside Slate's current contract or points at an external/browser/framework problem.
close-invalid
Close it and point at the actual owning layer or unsupported scope.
None.
Not a direct test candidate.
ottworksbug, ⚑ cross platform5Shadow DOM still breaks core typing behavior badly enough to make Firefox editing unusable.
The later comments confirm this issue survived far beyond the original report, so it remains real signal.
Strong enough with minor setup.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready with minor setup.
STMU1320bug3IME candidate navigation with arrow keys can repeat committed text and place it in the wrong location.
The follow-ups show the same class of bug appears across Chinese, Japanese, and accented Latin input, which makes this broad platform pain.
Strong enough with minor setup.
Poor.
Valid.
Keep it in the right cluster and do not let it drift into unrelated themes.
keep-open
Keep it open and scoped to the actual seam instead of letting it dissolve into vague frustration.
Direct.
Ready with minor setup.