fix(tui): pick redo target by message order, not ID

[as4193]

Issue for this PR

Closes #43034

Type of change

• Bug fix
• New feature
• Refactor / code improvement
• Documentation

What does this PR do?

The TUI redo command picked its target with x.id > revert.messageID. The message
list is ordered by time.created (#40994), so on a session whose IDs are not in
chronological order that lookup can return a message before the revert boundary.
Redo then stages the revert further back than it already was, and the next turn
commits it — the messages in between are deleted.

The fix looks the revert message up by position and takes the next user message
after it, which is what the web client's redo already does
(packages/app/src/pages/session/use-session-commands.tsx) and what /undo and
revertRevertedMessages in the same TUI file were converted to in #40994. If the
revert message isn't in the loaded list we bail instead of guessing, same as the
web client.

Three lines, no behaviour change on sessions with in-order IDs: there, the first
user message with a higher ID is the next one positionally.

How did you verify your code works?

• bun run --cwd packages/tui typecheck — clean.
• bun test in packages/tui — 193 pass, 1 skip, 0 fail (same as before the change).
• Compared old vs new selection on a message list whose ID order is inverted
  relative to time.created: the old predicate returns index 0 (before the
  boundary), the new one returns the next user message after the boundary, or
  nothing when the boundary is already the last user message (so redo clears the
  revert, which is correct).
• bunx oxlint and bunx prettier --check on the touched file: no new warnings.

I did not add a test: the redo command lives inside the Session() route
component, and there is no fixture in packages/tui/test that mounts a route
(the existing tests cover the sync store, utils and exported pure helpers). #40994
added tests for the same reason only where the code was already reachable. Happy to
add one if you'd like the selection pulled into a testable helper.

Screenshots / recordings

Not a visual change.

Checklist

• I have tested my changes locally
• I have not included unrelated changes in this PR

[Enough1122]

AI code review — automated review for reference; please use your judgment.

• packages/tui/src/routes/session/index.tsx:659 — Good fix: the old x.id > messageID comparison relies on lexicographic ID ordering, which breaks for non-monotonic or differently-formatted IDs; using the array index of the revert boundary is more robust.
• packages/tui/src/routes/session/index.tsx:658 — Early return when findIndex is -1 is correct: if the reverted message was pruned from local state, redoing to an arbitrary later user message would be wrong.
• packages/tui/src/routes/session/index.tsx:660 — Minor: if two entries share the same ID, findIndex picks the first; consider whether that ambiguity matters in practice (likely not).
• No regression test added for the ordering case (e.g., IDs where lexicographic vs positional order disagree); worth adding if session route has test coverage.

No blocking issues found.

[github-actions]

Automated PR Cleanup

Thank you for contributing to opencode.

Due to the high volume of PRs from users and AI agents, we periodically close older PRs using automated criteria so maintainers can focus review time on the most active and community-supported contributions.

This PR was closed because it matched the following cleanup criteria:

• The PR was created more than 1 month ago
• The PR had fewer than 2 positive reactions
• Positive reactions are counted as thumbs-up, heart, celebration, or rocket reactions on the PR

PRs created within the last month are not affected by this cleanup.

If you believe this PR was closed incorrectly, or if you are still actively working on it, please leave a comment explaining why it should be reopened. A maintainer can review and reopen it if appropriate.

Thanks again for taking the time to contribute.