A small follow-up to 0.70.0, all of it in the dashboard.
Upgrading from 0.61.3 or earlier? This still applies
A container that answered your LAN before the upgrade stops answering it after. Add -e PINGULARITY_ACCESS=network (or -access network) and recreate it to opt back in. Nothing is lost - the volume, database and history carry over - and a --network=host container or a native install on localhost is unaffected.
0.61.3 was a rollback, so it behaves like 0.60.3: a published port answers the whole network. This line brings private-by-default back. It refuses rather than guessing because the daemon cannot tell an upgraded store from one born private that simply ran for a while, and guessing open would put an unauthenticated dashboard on your network. The full reasoning is in the 0.70.0 notes.
What changed in 0.70.1
If you forget your password, the way out is now on the login box. The instructions for pingularity reset-auth used to live in the Access tab - which sits behind the login, so the only person who needed them was the only person who could not reach them. They are now on the sign-in screen itself, and they name the container command when you are running in a container rather than sending you to the host.
Log out moved onto the line listing the addresses the dashboard answers on, which is where it belongs: it acts on this session, and those are where the session is reachable.
Spacing across the settings tabs. Several rows and headings sat at distances that looked deliberate and were not, so identical controls were further apart on one tab than another. The Access rows were more than twice the gap of the identical rows on Latency, the section headings dropped further than the card headings beside them, and the Schedule tab started lower than every other tab.
No behaviour outside the dashboard changed, and there is no migration.
Changelog
- Tidy the Access tab, and put the lockout fix where it can be read