Make a missing CHECK_EPOCH fail the build instead of faking it (closes #91)
All checks were successful
check / check (pull_request) Successful in 2m13s

Adopt the four remaining upstream CHECK_EPOCH hardening items from
sneak/prompts #26, closing the gap #85 left open deliberately.

Fail closed on a missing value. Each check stage now asserts
`[ -n "$CHECK_EPOCH" ] || exit 1` before running anything. An unset ARG
is an empty string and an empty string is a stable cache key, so the
second and every later bare `docker build .` on an unchanged tree
replayed all three check layers, executed nothing, and still exited 0 --
and `docker build .` is the command REPO_POLICIES.md names verbatim as a
thing that must be green, so the documented command was precisely the
one that lied. Failed steps are never cached, which is what makes the
guard fire on every invocation rather than once.

Expand the epoch into each check command rather than leaving it a bare
declaration, so the cache miss does not depend on BuildKit's
unreferenced-ARG handling staying as it is, and so the value appears in
the build log where a reader can see the layer was keyed fresh.

Make the epoch unique per invocation rather than per second:
`epoch="$(date +%s%N)$$"`. `%N` alone is not enough, since busybox drops
it silently and exits 0, handing back second granularity with no
warning; `$$` differs between concurrent invocations regardless. The
bare-assignment form is kept on purpose -- inlined into an argument, a
failing substitution does not abort under `set -eu` and would yield an
empty constant epoch, restoring the exact false green this prevents.

Pass the same fresh value from script/docker. It is not the CI gate, but
local builds are almost always warm, so it was the likelier fooling in
practice, and two entrypoints disagreeing about whether the tree is
green is worse than either being wrong alone.

The ARG placement from #85 is unchanged, below apk add, COPY go.mod
go.sum and go mod download, so dependency layers still cache and the
build is not cold. Verified by negative control rather than inspection;
measurements are recorded once, in the PR verification comment.

.golangci.yml, the lint-stage FROM line and its digest, script/lint,
REPO_POLICIES.md and .gitea/workflows/check.yml are untouched.
This commit is contained in:
clawbot
2026-08-09 07:52:42 +00:00
parent c3bb3b5580
commit b101b4e21c
5 changed files with 102 additions and 40 deletions

36
TODO.md
View File

@@ -19,6 +19,42 @@ or delete the branch.
# Completed Steps
- 2026-08-09: Adopted the remaining upstream `CHECK_EPOCH` hardening
(issue #91), closing the gap #85 knowingly left open. Four changes,
all four decided as adopt upstream in `sneak/prompts` #26. (1) Each
check stage now asserts `[ -n "$CHECK_EPOCH" ] || exit 1` before
running anything, so a build that supplies no `--build-arg` fails
instead of lying. This is the item that mattered: an unset `ARG` is
an empty string and an empty string is a stable cache key, so the
second and every later bare `docker build .` on an unchanged tree
replayed all three check layers and still exited 0 — and `docker
build .` is the command `REPO_POLICIES.md` names verbatim as a thing
that must be green, so the documented command was precisely the one
that lied. Failed steps are never cached, which is what makes the
guard fire on every invocation rather than once. (2) The epoch is now
expanded into each check command rather than left as a bare
declaration, so the cache miss no longer depends on BuildKit's
unreferenced-`ARG` handling staying as it is, and the value appears
in the build log. (3) `script/cibuild` uses
`epoch="$(date +%s%N)$$"`, unique per invocation rather than per
second; `%N` alone is insufficient because busybox drops it silently
and exits 0, and `$$` is what makes the guarantee hold regardless.
The bare-assignment form is kept deliberately — inlined in an
argument, a failing substitution does not abort under `set -eu` and
would yield an empty constant epoch, restoring the exact false green
being fixed. (4) `script/docker` passes the same fresh arg, so the
two entrypoints cannot disagree about whether the tree is green;
local builds are almost always warm, which made it the likelier
fooling in practice. The `ARG` placement from #85 is unchanged, below
`apk add`, `COPY go.mod go.sum` and `go mod download`, so dependency
layers still cache and the build is not cold. Verified by negative
control rather than inspection — a bare `docker build .` run twice
back to back, plus back-to-back pairs of both scripts and a host-side
`make check`; the measurements are recorded once, in the PR
verification comment, rather than restated here. `.golangci.yml`, the
lint-stage `FROM` line and its digest, `script/lint`,
`REPO_POLICIES.md` and `.gitea/workflows/check.yml` are all
untouched.
- 2026-08-09: Stopped `script/cibuild` from reporting a green it did
not earn (issue #85). A bare `docker build .` let Docker serve the
check layers from the layer cache whenever the tree had not changed: