2026-09-04 - (/loop) The relation needed a mount, not a guest
all 7 measured relations now asserted (was 6 + 1 "unreachable")
the encoder-tier group checks out: 7 reasons accurate as written
suites 131 clippy/fmt/identity clean on both repos
Thirty-third cron tick, both remaining plan items.
(a) The encoder group checks out
Six related-libs entries say the console reports an
encoder library absent, completion "model the tiers the manifest
records"; the tier-probe entry says "orbistoun models no
tiers".
Nothing models them. Every "tier" in the source is the
provenance tier the audit sorts symbol derivations into
- a different word for a different thing. The nearest thing is
FIRMWARE_MODULE_DIRECTORIES, which refuses all three
directories identically.
Seven reasons accurate as written. Worth a paragraph, not a manufactured finding: four ticks of re-deriving stated blockers has found two wrong and thirteen right, which is about the ratio to expect before calling the exercise finished.
(b) The one that was wrong was mine, from two ticks ago
D545 recorded file-position-tracks-reads as unreachable
because nothing opens in a bare service test. Every clause true; the
conclusion wrong. The relation needed a mount, not a running
guest. orbistoun_fs::mount::mount is public,
descriptor.rs's own tests stand a title up with it in four
lines, and orbistoun-service already depends on the
crate.
What I had established was "the filesystem answers nothing when nothing is mounted" - a fact about the default state - and I read it as "a service test cannot mount anything". The skipped check is the cheapest one there is: ask whether the tool has a seam before concluding the tool cannot be used.
file_position.rs asserts two four-byte reads of a
sixteen-byte file give 0123 then 4567, and a
third after seeking back gives 0123. Bytes, not
counts - a descriptor re-reading the start returns the right
length twice and the wrong content, which is the failure the relation is
about. The console's 0x20 is compared to nothing; it is
likely a byte count but that is a reading, not a citation.
Broken twice, in the two mechanisms it spans: a read
that seeks to zero first fails the advancing half; a seek
that reports success without moving fails the seeking half.
The sentence in measured_relations.rs calling it
uncovered is rewritten rather than left - second time in a week that
check 13 has caught my own prose.
Decision: D547.