Fix update propagation and routing issues

[sanity]

Summary

This PR fixes critical issues preventing contract updates from propagating between peers on the Freenet network. The fixes address problems discovered while investigating why River chat room updates weren't reaching other peers.

Changes

1. Fixed UpdateNoChange Handling

• UpdateNoChange is now treated as a valid response (returns existing state) rather than an error
• Prevents unnecessary error propagation when updates don't change contract state

2. Added State Change Detection

• Updates now compare state before/after to determine if broadcasting is needed
• Only broadcasts when state actually changes (respecting commutative monoid property)
• Prevents circular update loops from unchanged state broadcasts

3. Fixed Routing to Connected Peers Only

• Update operations now use closest_potentially_caching to ensure routing to connected peers
• Previously used .pop() on subscribers without checking connectivity
• Fixes "No existing outbound connection" errors

4. Added Comprehensive Logging

• Added detailed routing decision logs showing peer selection process
• Added subscription propagation tracking
• Added update broadcast target logging

Technical Details

The core issues were:

1. Update operations selecting disconnected peers as targets
2. Treating UpdateNoChange as an error instead of valid state
3. Broadcasting updates even when state didn't change

These fixes ensure updates only propagate when needed and only to connected peers.

Testing

• Created minimal 2-gateway + 2-node test to isolate issues
• Added connection stability test
• Manual testing with River chat application recommended

Related Issues

Fixes #1611 (partially - update propagation aspect)

[Copilot]

Pull Request Overview

This PR addresses debugging issues by correcting documentation typos, refining contract caching in operations, and updating test code logging and orchestration for the freenet-ping app.

• Corrects a spelling error in a documentation comment.
• Adjusts contract caching logic and logs in the put operation.
• Updates visibility of the Contract struct and refactors test logging and future handling, along with adjusting the contract path in test utilities.

Reviewed Changes

Copilot reviewed 17 out of 17 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
crates/core/src/tracing/mod.rs	Fixed a spelling mistake in a documentation comment.
crates/core/src/operations/put.rs	Improved caching flow and updated comments for clarity.
apps/freenet-ping/contracts/ping/src/lib.rs	Made Contract struct public to support external usage.
apps/freenet-ping/app/tests/run_app_partially_connected_network.rs	Replaced tracing logging with println for test output and updated future orchestration.
apps/freenet-ping/app/tests/common/mod.rs	Updated contract path and refactored build utilities for test contract loading.

Comments suppressed due to low confidence (2)

apps/freenet-ping/app/tests/run_app_partially_connected_network.rs:70

• [nitpick] Consider using the project's standard logging framework instead of println for consistency with other parts of the codebase.

println!("Starting test with {} gateways and {} regular nodes (connectivity ratio: {})", NUM_GATEWAYS, NUM_REGULAR_NODES, CONNECTIVITY_RATIO);

apps/freenet-ping/app/tests/common/mod.rs:138

• [nitpick] Ensure that the updated contract path still correctly locates the contract source, as intended for the on-the-fly compilation in tests.

pub const PATH_TO_CONTRACT: &str = "../contracts/ping";

[iduartgomez]

I see we are duplicating similar logic in multiple commands, would be good to refactor this in the future

[sanity]

You're absolutely right about the code duplication. Claude identified the same pattern repeated in 6 places across 3 files while fixing the CARGO_TARGET_DIR issues.

A future refactor could extract this into a shared utility function, perhaps something like:

pub fn get_workspace_target_dir() -> PathBuf {
    std::env::var("CARGO_TARGET_DIR")
        .map(PathBuf::from)
        .unwrap_or_else( < /dev/null | _| find_workspace_root().join("target"))
}

This would make the code more maintainable and reduce the risk of inconsistencies. I'd be happy to create a follow-up PR for this refactoring if you'd like.

[sanity]

Thanks for the feedback! I've addressed the code duplication in commit 0895894.

The refactoring:

• Created shared utility functions get_workspace_target_dir() in both crates
• Replaced all 6 duplicated implementations
• The logic is now centralized and easier to maintain

This should make future changes to the workspace detection logic much simpler.

[sanity]

Superceded by #1612.