feat: enhance manual-exec to cycle through multiple messages

[thinktanktom]

Description

Closes #111

This PR enhances the manual-exec command to allow users to cycle through multiple failed CCIP messages in a single session, eliminating the need to run the command multiple times for transactions with multiple messages.

Problem

Currently, when a transaction contains multiple CCIP messages (e.g., 6 failed messages), users must run ccip manual-exec <tx> six separate times—once for each message. This is tedious and time-consuming.

Solution

Added a while(true) loop around message selection and execution that:

• Continues after each execution, returning to the message selection menu
• Exits cleanly when user selects the existing "Exit" option
• Handles execution errors gracefully without crashing the loop
• Fetches messages once before the loop for efficiency

Changes

• Loop implementation: Wrapped message selection and execution in while(true) loop
• Exit handling: Added try-catch to handle CCIPError thrown when user selects "Exit"
• Error resilience: Wrapped execution in try-catch so failed executions show error and return to menu
• Optimization: Moved getMessagesInTx() call outside loop to avoid redundant blockchain queries
• CHANGELOG: Added entry documenting the enhancement

Testing

• Code compiles without errors (npm run build)
• Linting passes (npm run lint)
• TypeScript type checking passes
• Follows project's error handling guidelines (uses existing CCIPError infrastructure)
• Tested with real blockchain transactions (requires testnet setup - happy to test if maintainers provide test data)

User Experience

Before:

$ ccip-cli manualExec 0xabc123...
> Select message: Message A
✓ Executed
# CLI exits - must run again

$ ccip-cli manualExec 0xabc123...  # Run again for next message
> Select message: Message B
✓ Executed
# CLI exits - must run again

# Repeat for each message...

After:

$ ccip-cli manualExec 0xabc123...
> Select message: Message A
✓ Executed

> Select message: Message B  # Automatically returns to menu
✓ Executed

> Select message: Message C  # Can continue as needed
✓ Executed

> Select message: Exit  # Exit when done
Exiting manual execution.

Implementation Notes

• The "Exit" option was already present in the selectRequest() utility function
• No UI changes were needed—just proper handling of the exit behavior
• Changes are minimal and focused on control flow in manual-exec.ts
• No modifications to shared utilities or other commands
• Follows project's error handling guidelines (uses CCIPError, no generic Error objects)

Checklist

• My code follows the style guidelines of this project
• I have performed a self-review of my code
• My changes generate no new warnings
• I have updated the CHANGELOG.md
• PR title follows Conventional Commits format (feat:)
• Focused, atomic commits

[vercel]

@thinktanktom is attempting to deploy a commit to the Chainlink Labs Team on Vercel.

A member of the Team first needs to authorize it.

[thinktanktom]

Found and fixed a bug: single-message transactions would loop infinitely because selectRequest() auto-returns when only 1 message exists (no menu shown).
Although technically users can hit Ctrl+C to exit, I wanted to cover all edge cases.
Added a check to exit the loop after executing single-message transactions.

Fixed in commit: f1151f5

[thinktanktom]

Optional Enhancement

While implementing this, I considered adding execution tracking to help users keep track of which messages they've already executed in the current session.

Would add:

• Set<string> to track executed message IDs
• Visual indicator in the menu (e.g., [✓] next to executed messages)
• Optional warning when attempting to re-execute

I kept the initial implementation minimal, but happy to add this enhancement if it would improve the UX. Let me know what you think!