relay.py: include error detail + URL in on_error print

[bitcoin3us]

Summary

The _on_error callback in nostr/relay.py currently prints a bare relay.py got error with no exception info and no relay URL. When a WebSocket connection fails (TLS handshake, DNS, connection reset, socket exhaustion, etc.) there's nothing in the log to tell you what went wrong or which relay in a multi-relay configuration.

def _on_error(self, class_obj, error):
    print("relay.py got error")       # ← no detail, no URL
    self.connected = False
    ...

Discovery

Encountered while debugging NWC connect failures on an ESP32 running MicroPythonOS. The device kept reporting "Could not connect to any Nostr Wallet Connect relays" with no explanation — only relay.py got error in serial. Had to patch this file locally to surface the exception before the real issue (socket-pool exhaustion from a stale LNBits WebSocket) became visible.

Fix

print("relay.py got error for {}: {!r}".format(self.url, error))

repr() on the exception gives both the type and the args/message, which is what you want for diagnostics. No behaviour change — same signature, same control flow, reconnect logic untouched.

Example output

Before:

relay.py got error

After (illustrative):

relay.py got error for wss://relay.coinos.io: OSError(118, 'Host is unreachable')
relay.py got error for wss://relay.damus.io: SSLError(-29312, 'MBEDTLS_ERR_…')
relay.py got error for wss://nos.lol: ValueError('unexpected close frame')

One line, enough signal to point at DNS vs TLS vs socket exhaustion vs protocol issues.

Test plan

• Trigger a connection failure (block the relay's port in a firewall, stop the local relay mid-connect, etc.) — verify log includes the URL and exception repr
• Trigger a clean disconnect — _on_close still fires (unchanged)
• Trigger a successful connect — _on_open still fires (unchanged)

🤖 Generated with Claude Code