-
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
Good catch, and it is a bug. Fixed in v1.5.1-rc.6, released today. To be clear about what was and wasn't wrong: the digest was always correct. I pulled Home Assistant's registry data while digging into this and rc.6 rejects any candidate tag that needed that lossy parse, so nightlies and other odd formats can't masquerade as stable suggestions anymore. It also renders the suggestion in that view with a proper "Suggested" badge and tooltip instead of a bare string squeezed next to Digest/NEW, which should make it clearer what that text even is. If you'd rather wait for 1.5.1 GA (soaking now, roughly a week out) than run an RC, you can also just hide HA's nightlies from consideration with a label on the container: On the version column getting cut off at your window size: agreed, and that's part of a larger table/card rework already queued for v1.6. Thanks for the report, this one had good bones. |
Beta Was this translation helpful? Give feedback.




Good catch, and it is a bug. Fixed in v1.5.1-rc.6, released today.
To be clear about what was and wasn't wrong: the digest was always correct. I pulled Home Assistant's registry data while digging into this and
21e0d1bae299really is the right latest digest for that day, so the update itself was accurate. The bad part was the version text next to it. That label is a "suggested pin": since you're watchinglatest, drydock tries to show the newest stable version tag pointing at the same lineage, so you could pin a fixed version if you wanted. The suggestion logic leaned on semver'scoerce()as a last-resort parser, andcoerce()silently drops anything it doesn't understand. HA publishes PEP …