aboutsummaryrefslogtreecommitdiff
path: root/working
diff options
context:
space:
mode:
authorCraig Jennings <c@cjennings.net>2026-08-21 07:32:11 -0700
committerCraig Jennings <c@cjennings.net>2026-08-21 07:32:11 -0700
commit22782817e7c0be4762b35501325ad2ba09f8fb09 (patch)
treebacd8a7028a05eafa22c52102ef7d009e1bf9fce /working
parentd1b5ee1fdc07400e4308db7d346c5a8be0309c7a (diff)
downloadarchsetup-22782817e7c0be4762b35501325ad2ba09f8fb09.tar.gz
archsetup-22782817e7c0be4762b35501325ad2ba09f8fb09.zip
chore(tasks): close five shipped fixes and re-scope the reference-host one
The powerprofilesctl crash loop, the shallow read-only clones, the worldclock tooltip, the obsbot poll, and the spine tests all shipped. Each closure records what landed and which commit carries it, plus the things the tasks got wrong: the spine task's line number was stale and its count looks off by one until you notice the continuation-form call, and the obsbot task assumed ratio had the camera when ratio has a Logitech BRIO and no OBSBOT at all. The reference-host task gets a note rather than a closure. It argues from 8 findings at a 1:7 signal ratio, and velox now reports 1, because two of the three things it cites as noise are fixed at the source instead of filtered. Whether a live diff still earns its place against a static declaration of intent is a real question and not mine to settle, so I wrote down what moved and dropped its :solo: tag. It also already contained the whole 8-to-1 analysis when it was filed, and I re-derived that from scratch four days later without reading it. Recorded, because the lesson is about reading the backlog rather than about the checker.
Diffstat (limited to 'working')
0 files changed, 0 insertions, 0 deletions