fix: remove internal compound operations completely

[christianhg]

After #2327 decomposed split_node, merge_node, and move_node into lower-level operations, the compound types lingered in the Slate type system - Path.transform, Point.transform, Operation.inverse, and the Operation union all still handled cases that could never fire. This PR removes them entirely, inlines the transform logic that the two remaining apply helpers need, and deletes applyMoveNode along with the dead code paths it was keeping alive.

---

applyMergeNode and applySplitNode decompose compound operations into lower-level ones, but they still need the compound operation's ref-transform semantics: a merge shifts paths differently than the equivalent insert_text + remove_node sequence. Previously they delegated to Path.transform and Point.transform for this, which meant those functions had to keep the compound operation cases around even though no other code path used them. This PR inlines the transform logic directly into each helper as local transformPathForMerge, transformPointForMerge (and equivalents for split), then removes the compound types from Path.transform, Point.transform, Operation.inverse, operationCanTransformPath, the Operation union, and CustomTypes.

Path.transform drops from 130 lines to 30 - it now only handles insert_node and remove_node. Point.transform and Operation.inverse each lose three cases, including the most complex one (move_node inverse required transforming paths through the operation itself).

applyMoveNode is removed entirely. Its two call sites (operation.move.block.ts and delete-text.ts) now use plain remove_node + insert_node directly. The move helper's 180+ lines of ref pre-transform machinery had zero test coverage because the selection fixup it provided was redundant at both call sites: operation.move.block.ts already does manual selection fixup after the move, and delete-text.ts relies on applyMergeNode to handle selection immediately after.

The inlining surfaced a subtle detail in how RangeRef affinities are resolved. When a RangeRef has 'inward' affinity, the anchor and focus need different per-point affinities: on a forward range, the anchor gets 'forward' and the focus gets 'backward' (they contract toward each other). The old code got this for free through the Range.transform -> Point.transform chain. The inlined version needs it explicitly. A shared rangeRefAffinities helper resolves per-point affinities for applySplitNode - the only remaining helper where affinity matters, since merge point transforms don't depend on affinity.

Also cleans up dead code exposed by the removal: a split_node check in transformPendingPoint (diff-text.ts) that could never trigger, the unused properties parameter on applyMergeNode, dead inline-Element and block-level branches in applyInsertNodeAtPoint (impossible in the PT-native tree where inline objects are ObjectNodes, not Elements), and unused insert options on applyInsertNodeAtPath/applyInsertNodeAtPoint.

The operation vocabulary is now fully reduced: insert_node, remove_node, insert_text, remove_text, set_node, and set_selection. Two helpers remain (applySplitNode, applyMergeNode) that compose these into higher-level semantics with correct ref transforms. The next step is generalizing operation-to-patches.ts to use key-based paths for container support (#2336).

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
portable-text-editor-documentation	Ready	Preview, Comment	Mar 11, 2026 8:37am
portable-text-example-basic	Ready	Preview, Comment	Mar 11, 2026 8:37am
portable-text-playground	Ready	Preview, Comment	Mar 11, 2026 8:37am

[changeset-bot]

🦋 Changeset detected

Latest commit: 20650bf

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 11 packages

Name	Type
@portabletext/editor	Patch
@portabletext/plugin-character-pair-decorator	Patch
@portabletext/plugin-emoji-picker	Patch
@portabletext/plugin-input-rule	Patch
@portabletext/plugin-markdown-shortcuts	Patch
@portabletext/plugin-one-line	Patch
@portabletext/plugin-paste-link	Patch
@portabletext/plugin-sdk-value	Patch
@portabletext/plugin-typeahead-picker	Patch
@portabletext/plugin-typography	Patch
@portabletext/toolbar	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

📦 Bundle Stats — @portabletext/editor

Compared against main (d312abd5)

@portabletext/editor

Metric	Value	vs main (d312abd)
Internal (raw)	798.3 KB	-2.7 KB, -0.3%
Internal (gzip)	150.0 KB	-281 B, -0.2%
Bundled (raw)	1.40 MB	-2.7 KB, -0.2%
Bundled (gzip)	313.0 KB	-262 B, -0.1%
Import time	102ms	+0ms, +0.4%

@portabletext/editor/behaviors

Metric	Value	vs main (d312abd)
Internal (raw)	467 B	-
Internal (gzip)	207 B	-
Bundled (raw)	424 B	-
Bundled (gzip)	171 B	-
Import time	6ms	+0ms, +0.8%

@portabletext/editor/plugins

Metric	Value	vs main (d312abd)
Internal (raw)	2.5 KB	-
Internal (gzip)	910 B	-
Bundled (raw)	2.3 KB	-
Bundled (gzip)	839 B	-
Import time	12ms	-0ms, -0.6%

@portabletext/editor/selectors

Metric	Value	vs main (d312abd)
Internal (raw)	60.2 KB	-
Internal (gzip)	9.4 KB	-
Bundled (raw)	56.7 KB	-
Bundled (gzip)	8.6 KB	-
Import time	11ms	+0ms, +0.9%

@portabletext/editor/utils

Metric	Value	vs main (d312abd)
Internal (raw)	24.2 KB	-
Internal (gzip)	4.7 KB	-
Bundled (raw)	22.2 KB	-
Bundled (gzip)	4.4 KB	-
Import time	10ms	+0ms, +1.0%

Details

• Import time regressions over 10% are flagged with ⚠️
• Treemap artifacts are attached to the CI run for detailed size analysis
• Sizes shown as raw / gzip 🗜️. Internal bytes = own code only. Total bytes = with all dependencies. Import time = Node.js cold-start median.