fix(patterns): getRankCover for M.kind('copyBag') (closes #3052)

[kriscendobot]

Closes: endojs/endo#3052

Description

getRankCover(M.kind('copyBag')) threw a TypeError: Cannot read properties of undefined (reading 'cover') because matchKindHelper.getRankCover mapped only 'copySet' and 'copyMap' to the 'tagged' pass style; 'copyBag' fell through to the default branch, which looked up the kind directly in the pass-style cover table. There is no 'copyBag' pass style, so getPassStyleCover returned undefined.

The fix adds a case 'copyBag': next to the existing case 'copySet': / case 'copyMap':, so all three tagged collection kinds share the 'tagged' pass-style cover. A regression test in packages/patterns/test/copyBag.test.js asserts getRankCover(M.kind('copyBag')) returns the same range as M.kind('copySet') and M.kind('copyMap'). Reverting the one-line fix causes the test to fail with the exact TypeError from the issue.

Security Considerations

No impact. The change broadens which inputs are accepted by an existing rank-cover lookup; no new authority is introduced.

Scaling Considerations

No impact. One added switch case, evaluated at pattern-match time only.

Documentation Considerations

No impact. The fix restores the documented behavior implied by M.kind accepting all three tagged collection kinds.

Testing Considerations

A targeted regression test in packages/patterns/test/copyBag.test.js verifies M.kind('copyBag') produces the same rank cover as the sibling kinds. Reverting the fix causes the test to fail with the issue's TypeError.

Compatibility Considerations

No impact. Previously-thrown calls now succeed; previously-passing calls are unchanged.

Upgrade Considerations

No impact.

[kriscendobot]

Implements the fix from the agent-ready plan for endojs/endo#3052: add 'copyBag' to the matchKindHelper switch alongside 'copySet' and 'copyMap'. Regression test in packages/patterns/test/copyBag.test.js fails with the issue's TypeError when the new case is reverted.

[kriscendobot]

Duplicates upstream endojs/endo#3228 (OPEN, awaiting review), which has the same one-line fix and a regression test. The earlier bots-repo attempt #65 was already closed without merge in favor of the upstream PR. Closing in favor of #3228.