[rmorgans]

RevisionProcessor.RejectRevisions removes a cell marked w:cellIns and leaves its row one cell short. Word's Reject All gives the inserted cell's columns to the cell before it. Clippit already does the same thing when it accepts a w:cellDel (AcceptDeletedCellsTransform, fixed further in #555).

Reproduction. A 2×2 table where cell B2 is marked w:cellIns.

• Reject gives A1 | B1 / A2, with row 2 one grid column short.
• Word gives A1 | B1 / A2 (gridSpan 2).

Fix. When rejecting, turn w:cellIns into w:cellDel. The accept pass then removes the cell and widens the one before it. There is one exception, and it matches what Word does. If the cell before has a w:tcPrChange, as when a cell was split (RP035), Word restores the recorded properties and does not widen the cell, so the inserted cell is only removed, as before.

Test. RP061_RejectedInsertedCellWidensCellBefore (accept and reject) is in the new file RevisionProcessorTests.CellIns.cs. All RP001 baselines still pass, including RP034/RP035.

Relation to #553. Copilot's open comment on #553 about RevisionProcessor.cs:31 says that rejecting an inserted cell removes it before the comment paragraph IDs are captured. I ran a throwaway test: a comment whose final paragraph w14:paraId="22222222" is in a w:cellIns cell, then Reject All, checking the ID on the comment's last paragraph.

Branch	Plain inserted cell	Cell before has w:tcPrChange
#553 alone	ID lost	ID lost
#553 + this PR	✅ kept	ID lost

With this PR, the cell is removed in the accept pass, after #553 captures the IDs, so the common case is covered. The w:tcPrChange case still removes the cell during the reject pass. Capturing the IDs before rejection, as Copilot suggests, would cover both.

Merging. Like #552 and #554, this one only edits RevisionProcessorTests.cs to make the class partial, the same one-word edit as those PRs. The four PRs merge onto master together without conflicts, and the full suite passes.

---

Expected results are Microsoft Word for Mac 16.113's own Accept All / Reject All on the same markup. The regression test fails on current master and passes with the fix.

Prepared with an AI coding agent (Claude). I reviewed the change and ran the tests.

[Copilot]

🟡 Changes recommended

Unresolved cases can delete unrelated nested-table content, retain rejected inserted cells, or transfer too few columns.

3 open issues

• Nested inserted cells can delete unrelated enclosing cells · New
• Property restoration can prevent deletion of inserted cells · New
• Deleted multi-column cells produce an incorrect grid span · New

What changed in this PR

Updates Word revision rejection to give rejected inserted cells’ columns to the preceding cell through the existing deletion-handling pass.

Changes:

• Converts inserted-cell markers to deletion markers, preserving the preceding-cell property-change exception.
• Adds accept/reject regression coverage in a separate partial test file.
• Updates the rejection behavior documentation.

File	Description
Clippit/​Word/​RevisionProcessor.cs	Changes inserted-cell rejection and documents the handoff.
Clippit.Tests/​Word/​RevisionProcessorTests.cs	Makes the test class partial.
Clippit.Tests/​Word/​RevisionProcessorTests.CellIns.cs	Adds inserted-cell accept/reject coverage.

_{Review effort: Balanced}

[Copilot]

🟡 Changes recommended

Wrapped inserted cells can survive rejection with their content and tracked deletion markers intact.

1 open issue

• Remove wrapped inserted cells when rejecting revisions · New

3 resolved since last review

• Nested inserted cells can delete unrelated enclosing cells
• Deleted multi-column cells produce an incorrect grid span
• Property restoration can prevent deletion of inserted cells

_{Review effort: Balanced}

[Copilot]

🟢 Approval recommended

No blocking issues were identified, and regression tests cover the changed behavior; tests were not executed during review.

No open issues 🎉

1 resolved since last review

• Remove wrapped inserted cells when rejecting revisions

_{Review effort: Balanced}