Repository navigation
Read a press's button without naming PointerEvent - #21
Merged
Merged
Conversation
DecoratorBoundarySelectionPlugin's pointerdown listener checked `event instanceof PointerEvent`. jsdom before 27 defines no PointerEvent global, so every press in such a host threw a ReferenceError from the listener. paranext-core's platform-bible-react suite runs under jsdom 26 and ends with 18 unhandled errors, failing its CI on main and every PR. Duck-type the button instead: a press carrying `button` is primary only when it is 0; one carrying none (a plain Event) still counts as primary, as it did before. Pins both sides: a secondary-button PointerEvent writes the DOM straight away, and a press still counts as a drag with the PointerEvent global removed. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
tjcouch-sil
enabled auto-merge (squash)
September 23, 2026 16:09
rolfheij-sil
approved these changes
Sep 23, 2026
rolfheij-sil
left a comment
rolfheij-sil
left a comment
Contributor
There was a problem hiding this comment.
🤖 Claude: Approved. The "button" in event guard behaves the same in real browsers, cannot throw under jsdom, the committed dist carries exactly that one-line change, and the new test reproduces the original failure (fails on c5c4f37, passes here). A sweep of libs and packages/*/src found no other instanceof/new use of a browser-only global on a path core's suite reaches. Two optional nits inline. Reminder from the PR body: paranext-core stages from platform-yalc, so after merging, npm run move-platform-yalc is what unbreaks core CI.
…-repo claim Save and restore the global's property descriptor rather than its value, so the test leaves the global exactly as vitest installed it. Phrase the no-PointerEvent host generically instead of naming another repo's jsdom. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
rolfheij-sil
approved these changes
Sep 23, 2026
12 tasks done
tjcouch-sil
added a commit
that referenced
this pull request
Oct 1, 2026
Brings in #13 (Standard-view copy/paste), #21, and the \periph, gutter and font-fallback work. Conflict resolutions: node.utils.ts ours; unknownUsfm.utils.ts both; usfmFragmentToUsj.ts main's pushTableCell (tests: both sides); selection.utils.ts ours (subsumes main's closing glyph fix, whose tests pass unchanged); Editor.tsx imports union and insertNote live-resolution inside main's discrete update; markerEdit.test-helpers.tsx both; settledGetUsj.test-helpers.tsx both mount options; dist ours (rebuilt at the end of the round). Details: - usfmFragmentToUsj.ts: main's five hunks; this branch's two auto-merged comment hunks (the cell-marker regex doc and the para-case table comment) are kept. Both call-site conditions this branch pinned (`tableEligible && tableRow`, `!token.isNested && !note && !isNoteContext && tableRow`) are present in main's code. - usfmFragmentToUsj.test.ts: both describe blocks whole (this branch's default-stylesheet cell rows, main's \periph rows); no row appears on both sides; the doubled defaultStyleInfo import is dropped. - Editor.tsx imports: copySelection, cutSelection and DecoratorBoundarySelectionPlugin added. DeltaOnChangePlugin and DeltaOp are not, since this branch no longer uses either. - settledGetUsj.test-helpers.tsx: structureProtectionMode passes through editorWith's options next to view, isReadonly and styleInfo. - dist: packages/platform/dist taken from this branch whole, index.d.ts included. - Semantic merge fixes (this branch made the view options explicit): MarkersViewCopyPlugin.tsx passes its viewOptions to $getUsjSelectionFromEditor and the Standard view options to $getRangeFromUsjSelection; main's new selection.utils.test.ts row calls $getUsjSelectionFromEditor(undefined). - Lockfile: main adds postcss to platform devDependencies; installed with `pnpm install --frozen-lockfile`. - New annotationLocations/exactAnnotationLocations.test.tsx pins that a pasted copy of an annotated attribute run holds no annotation on both internal-paste paths that insert nodes: Lexical's rich-paste insertion (structure protection off), which reaches the strip through SELECTION_INSERT_CLIPBOARD_NODES_COMMAND (the row fails with the strip disabled), and the structure-protection html sanitizer, whose nodes come from the USFM-bytes text/html and carry no node state. The Standard-view external-paste claim, the attribute-context claim, the in-note line replay and the char-stack line replay all insert re-tokenized USFM text, which carries no node state, so they need no row. Tests: shared 752 -> 778 (main +26); shared-react 1769+1 skipped -> 1786+1 (main +17); platform 2431 -> 2702+3 skipped (main +269 and 3 skipped, new pin rows +2); utilities 59, scribe 2 unchanged. This branch's own delta over the merge base: shared +77, shared-react +115, platform +739. nx lint 10/10, nx typecheck 10/10. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: 2c4fe30c-6a8b-42a8-ba57-9141bb0ab91f
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes the paranext-core CI break from #13 (e.g. this run). Core's
platform-bible-reacttests all pass, but the run finishes with 18 unhandledReferenceError: PointerEvent is not definedand exits 1.Cause: the
pointerdownlistener inDecoratorBoundarySelectionPlugincheckedevent instanceof PointerEvent. jsdom only added aPointerEventglobal in v27. This repo tests under jsdom 28, so the reference always resolved here. paranext-core tests under jsdom 26, so every press threw inside the listener.Fix: check the button with
"button" in event && event.button !== 0, which doesn't reference any global. The behavior stays the same: a secondary-button press still doesn't count as a drag, and a press with nobutton(a plainEvent) still counts as a primary one. I didn't use atypeof PointerEvent !== "undefined"guard. It would also work, but it keeps the dependency on a realm global, and that dependency is what broke.Tests (
decoratorBoundarySelection.test.tsx):PointerEvent: removes the global and presses. Before the fix it fails with the sameReferenceErrorcore CI shows.The committed
packages/platform/distis rebuilt, andverify-committed-distpasses. The only dist change is the one guard line.After merge: move
platform-yalc(npm run move-platform-yalc). Core's CI builds the editor from that branch, so it stays red until then. Nothing on the paranext-core side needs to change.AI-assisted
🤖 Generated with Claude Code
This change is