The venue drill caught what no hermetic test had: with the workflow_call bridge delivering "no" perfectly, drill runs 16/17 still bootstrapped. The script gated on GITHUB_EVENT_NAME = workflow_dispatch and never read $BOOTSTRAP at all; the wrapper's only coupling was exporting the event name for yes. Correct while an operator's manual dispatch was the only dispatch there was — inert from #209 on, when the trigger job made every machine wake a workflow_dispatch event. Runs 459/523 bootstrapped for this reason, not for the input-delivery defect, which is real but was never the operative cause of the observed re-upserts. The script now gates on ${BOOTSTRAP:-no} = yes; the wrapper passes the input through untouched; the hermetic suite pins the exact regression pair (a dispatch event with no/unset creates and deletes nothing) alongside the yes path's full create+delete assertions. Refs #215
2.1 KiB
Fixed
-
The sweep's
bootstrapvalue crosses theworkflow_callboundary as a declared input, explicitly passed by the caller — the one channel measured to work; the called workflow did not see the caller's event inputs as an implicit substitute on this instance (#215). -
Before the bridge,
github.event.inputswas empty inside the called workflow, so every dispatch-woken sweep bootstrapped: ~20 label upserts on each board event (#215). -
The caller maps an empty top-level value to
noexplicitly, so a cron-woken sweep can never bootstrap; the declared input also defaults tono, so a consumer that passes nothing gets the safe path (#215). -
The gate feeds the declared input to
labels-reconcileunchanged, so an invalid value meets the action's ownyes|norefusal instead of being silently coerced (#215). -
docs/CONSUMERS.md's published sweep stub carries the same pass-through — without it every consumer inherits the defect ceremony fixed for itself (#215).
Added
-
test/labels-bootstrap.test.shpins the bridge at every hop: the declared boundary, both gate sites as the identity, no expression readinggithub.event.inputs, and the caller and stub pass-throughs byte-exact (#215). -
The same test drives the four value paths — schedule-empty,
no,yes, invalid — through the shipped expressions into the action's real validator (#215). -
The taxonomy bootstrap keys on the
BOOTSTRAPinput, never the event name. It testedGITHUB_EVENT_NAME = workflow_dispatch— correct while an operator's manual dispatch was the only dispatch there was, inert-by- construction from #209 on, when every machine wake became a dispatch event (#215). -
The venue drill caught that: with the bridge delivering
noperfectly, drill runs 16/17 still bootstrapped, because the script never read the input the whole chain existed to deliver (#215). -
test/labels-reconcile.test.shpins the regression pair exactly: aworkflow_dispatchevent withBOOTSTRAP=no(or unset) creates and deletes nothing; onlyBOOTSTRAP=yesbootstraps (#215).