packages/components/table/__tests__/table.test.ts (1)

2060-2070: Add an in-place append case here too.

This only verifies the splice/removal path. Since the regression is about array-method updates to expandRowKeys, it would be good to also assert that push expands newly added keys without replacing the whole array.

Possible follow-up

       wrapper.vm.expandRowKeys.splice(1, 1)
       await doubleWait()
       childRows = wrapper.findAll('.el-table__row--level-1')
       childRows.forEach((item, index) => {
         if (index < 2) {
           expect(item.attributes('style')).toBe('')
         } else {
           expect(item.attributes('style')).toContain('display: none')
         }
       })
+
+      wrapper.vm.expandRowKeys.push('1999-3-31')
+      await doubleWait()
+      childRows = wrapper.findAll('.el-table__row--level-1')
+      childRows.forEach((item) => {
+        expect(item.attributes('style') ?? '').not.toContain('display: none')
+      })

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@packages/components/table/__tests__/table.test.ts` around lines 2060 - 2070,
Test only covered splice/removal of wrapper.vm.expandRowKeys; add an in-place
append case that uses wrapper.vm.expandRowKeys.push(...) followed by await
doubleWait() and then re-evaluate childRows =
wrapper.findAll('.el-table__row--level-1') to assert that pushed keys cause the
corresponding rows to be visible (style not containing 'display: none') while
existing expanded rows remain visible (i.e., push should expand new rows without
replacing the array). Update the test near the existing splice block to perform
a push of a key that was previously hidden, await doubleWait(), and add
expectations for both previously-expanded and newly-expanded rows using the same
attribute checks as the splice assertions.