Tobias GesellchenandClaude Opus 4.8 d73ce7b559 fix(setup): bound speaker HTTP GETs so offline devices fail fast (refs #451) ...

Manager.HTTPGet defaulted to http.Get, which uses http.DefaultClient with
no timeout. An offline speaker therefore hung the caller for the OS-level
TCP timeout (~30 s). The admin device list refreshes every device's live
/info on each load (updateDeviceInfo per row), so a handful of offline
speakers each held a request for 30 s. Server-side those run concurrently
and never blocked other routes, but the browser's ~6-connections-per-origin
limit got saturated by the long-held /info requests, which made the whole
admin page (and navigating away from it) feel stuck.

Give HTTPGet a 5 s timeout (liveDeviceHTTPTimeout): ample for a healthy
speaker on the LAN, quick to fail a dead one. Applies to the /info,
/presets, /recents, /sources, inspect, and peer-probe GETs.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

2026-06-07 15:08:26 +02:00

..

client

chore(sanitize): remove real device ID and personal LAN IPs from tracked files

2026-06-06 19:11:24 +02:00

config

test: sweep example data in test files to RFC-5737 + placeholders

2026-05-17 22:05:13 +02:00

discovery

feat(health): add on-demand DNS-path diagnostics for the #345 speaker-DNS escape

2026-06-03 23:29:15 +02:00

models

fix(client): log unhandled WebSocket event names instead of empty list

2026-06-02 23:00:58 +02:00

service

fix(setup): bound speaker HTTP GETs so offline devices fail fast (refs #451)

2026-06-07 15:08:26 +02:00

speaker

refactor(speaker): introduce pkg/speaker leaf for shared protocol constants

2026-05-11 23:18:08 +02:00

ssh

sec8: document Run() invariant — command must never come from user HTTP input

2026-05-25 21:01:01 +02:00

telnet

test: sweep example data in test files to RFC-5737 + placeholders

2026-05-17 22:05:13 +02:00

testutils

fix(ci): gate http-client tests on mock readiness; address semgrep findings (refs #451)

2026-06-06 19:11:24 +02:00