script/Build.sol says:
cutRelease() freezes the candidate into src/generated/<tag>/ first. The
frozen 0_1_3/0_1_4/0_1_5 snapshots are append-only historical records,
never regenerated here.
src/generated/ holds 0_1_9, 0_1_10 and candidate — no 0_1_3, 0_1_4
or 0_1_5 — and LibCloneFactoryReleased declares clone-factory@0_1_9 and
clone-factory@0_1_10. The enumeration names snapshots that are not there and
omits both that are.
Delete the enumeration rather than correct it, per the ruling on
rainlanguage/rain.deploy#153: the frozen snapshots are directories in
src/generated/, the tree is the truth and cannot drift from itself, so
re-listing them in prose only re-arms the same trap at the next release. The
durable clause it sits in — frozen snapshots are append-only and run() never
regenerates them — is not visible from the code and stays.
Item 3 of rainlanguage/rain.deploy#153, filed here because the file is in this
repo.
script/Build.solsays:src/generated/holds0_1_9,0_1_10andcandidate— no0_1_3,0_1_4or
0_1_5— andLibCloneFactoryReleaseddeclaresclone-factory@0_1_9andclone-factory@0_1_10. The enumeration names snapshots that are not there andomits both that are.
Delete the enumeration rather than correct it, per the ruling on
rainlanguage/rain.deploy#153: the frozen snapshots are directories in
src/generated/, the tree is the truth and cannot drift from itself, sore-listing them in prose only re-arms the same trap at the next release. The
durable clause it sits in — frozen snapshots are append-only and
run()neverregenerates them — is not visible from the code and stays.
Item 3 of rainlanguage/rain.deploy#153, filed here because the file is in this
repo.