debug

/debug

NEVER guess. Logs always faster than speculation.

The rule

Bug occurs + fix not obvious →

1. Read log files FIRST. They have all errors.
2. Diagnose from ACTUAL error in logs. Not assumptions.
3. Not enough info? Add targeted tracing::debug!() (Rust) or logger.debug() (TS).
4. Never ask user to copy-paste terminal output. Read logs directly.

Log locations

Layer	File	Contents
Backend	~/.houston/logs/backend.log	Rust tracing — sessions, agent store, channels, watcher
Frontend	~/.houston/logs/frontend.log	JS console.error/warn, React crashes, Tauri cmd failures

Both daily rolling. Latest = current file.

Logging APIs

Rust

tracing::info!("session {id} started");
tracing::warn!(?error, "unexpected");
tracing::error!(?err, "failed to write");
tracing::debug!(agent = %agent_id, "received event");

Output → backend.log via tracing-subscriber + daily rolling file appender.

Levels: default info globally. debug for houston_terminal_manager + houston_tauri. Override: RUST_LOG=debug,crate_name=trace.

Frontend

import { logger } from "@/lib/logger";
logger.error("fetch failed", { url, status });
logger.warn("retry", { attempt });
logger.info("user clicked");
logger.debug("render", { props });

Also console.error + console.warn patched to auto-write to frontend.log.

Bug reports

"Report bug" button on error toasts auto-attaches last 50 lines from both logs.

Adding tracing when logs insufficient

1. Identify suspected code path
2. Add tracing::debug!(?relevant_vars, "descriptive message") at branch points
3. Commit to claude/wip
4. Ask user to reproduce
5. Read updated logs
6. Fix w/ actual knowledge

Don't leave debug logs in. Remove after fix or downgrade to trace!.

Anti-patterns

• ❌ "Let me try X and see if it fixes it" → guess
• ❌ "Can you share the terminal output?" → read the logs yourself
• ❌ unwrap() to silence compile errors → hides real failure
• ❌ let _ = x.await on ops that can fail → silent failure (banned by style rules)