v0.3.1

Changelog

• 4d496da test(dashboard): follow the QPU card's relabelled service control
• dd32952 fix(e2e): sync the sim extra, not a group that no longer exists
• 7d153c4 chore: bump to 0.3.1
• ea5524f docs(rfc): fix RFC 0005's structure, and drop the plan-tracker apparatus
• 3ab6f5e docs(rfc): distil RFC 0004 §11, and fix four stale claims
• 2339519 docs(rfc): fix a contradiction in RFC 0003, and tighten it
• eda6b2b docs(rfc): tighten RFC 0002
• 66c4aae docs(rfc): tighten RFC 0001, and fix three stale references
• 9027744 docs: cut the prose without cutting the facts
• 4058e50 refactor: let names carry what comments were carrying
• b78deff feat(qpi-driver): make the simulator installable
• 312214d test(qpi-ui): file the last two suites under the source they exercise
• 8d2e5ee fix(qpi-ui): move availability off /api/op, and one call instead of N
• 3508392 docs: name the renamed closed-set test, and the third driver kind
• b0c177e docs: add the two routes the maintenance work introduced
• 413c59e docs: correct stale claims across the READMEs and RFCs
• 574f840 docs: fix overstated hardware claim in the 0.3.0 changelog entry
• 1a08ee9 docs: condense the 0.3.0 changelog entry to terse bullets
• 452c3f6 fix(qpi-ui): pre-fill the config paths a driver cannot start without
• 363eaa5 feat: tell a driver its QPU's service state
• a64923e feat(qpi-ui): set and see maintenance from the dashboard
• 98ea6e3 fix(qpi-ui): make a QPU's state mean something in the dispatch path
• 5aede58 docs: the driver connection lifecycle and what reloads without a restart
• df4f8b5 feat(qpi-driver): prove the reload round-trips, and settle the hardware config
• 1dc8200 feat(qpi-driver): re-read the device config at the start of each DAG walk
• 800e806 feat(qpi-driver): re-read calibration.yml on each dispatched calibration
• 7d48f77 fix(qpi-ui): release a driver's lease on every path that ends it
• 81f8f19 feat(qpi-ui): one driver per role per QPU, refused at connect
• 8c66bf3 feat(qpi-driver): reload the device config when it changes on disk
• 17050e7 fix(qpi-ui): serialize calibration on the chip, not on the driver
• b5f22cb docs(rfc): say why no arrangement of the extras slims the install
• 2371fe4 docs(rfc): two full calibrations back to back are not a mistake
• 9830fc9 docs(rfc): answer the open questions in RFC 0004
• 0b963f0 fix(qpi-ui): ship /etc/qpi.config.yml in the linux packages
• cbe8665 fix(qpi-ui): let a tuner actually register, and stop a select from drifting
• c0716af WIP: Add fixme's in 0004 RFC