#!/usr/bin/env bash
# Refreshes the vendored copy of the official atproto interop test vectors.
#
# The vectors are CC-0, so they are copied into this repository outright rather
# than fetched at test time: a conformance suite that only runs when GitHub is
# reachable is a suite that stops running. The cost of vendoring is that the
# copy rots silently, which is what this script exists to prevent — run it,
# read the diff, and commit it as its own change.
#
# Everything upstream ships is vendored. `firehose/` was skipped for a while
# on the reading that it tested a frame format this server does not speak; it
# does not — its one file is a set of Merkle search tree commit proofs, which
# is a claim about the tree rather than about a wire format. Adding a
# directory is a line in VENDORED_DIRS plus a runner.
#
# `mst/` is vendored even though there is no Merkle search tree here. Two of
# its four files ask only about pure functions — a key's height and a common
# prefix length — and those exist. The directory is copied whole, so
# `example_keys.txt` comes with them and is run as well.
set -euo pipefail

UPSTREAM="https://github.com/bluesky-social/atproto-interop-tests.git"
VENDORED_DIRS=(syntax lexicon crypto data-model mst firehose)

cd "$(dirname "$0")/.."
dest="vendor/atproto-interop-tests"

# A pinned ref can be passed as the first argument; otherwise take the tip of
# the default branch. Pinning is for reproducing an older run, not for normal
# refreshes — the whole point is to move forward deliberately.
ref="${1:-HEAD}"

work="$(mktemp -d)"
trap 'rm -rf "$work"' EXIT

echo "fetching $UPSTREAM at $ref"
git init --quiet "$work"
git -C "$work" remote add origin "$UPSTREAM"
git -C "$work" fetch --quiet --depth 1 origin "$ref"
git -C "$work" checkout --quiet FETCH_HEAD

commit="$(git -C "$work" rev-parse HEAD)"
committed="$(git -C "$work" log -1 --format=%cI)"

rm -rf "$dest"
mkdir -p "$dest"
cp "$work/LICENSE-CC0" "$dest/LICENSE-CC0"
for dir in "${VENDORED_DIRS[@]}"; do
  cp -R "$work/$dir" "$dest/$dir"
done

cat > "$dest/PROVENANCE.md" <<EOF
# Vendored atproto interop vectors

Copied verbatim from the official interop test corpus. Do not hand-edit
anything in this directory: run \`scripts/refresh-interop-vectors.sh\` and
commit the result, so that the provenance below stays true.

| | |
| --- | --- |
| Upstream | <https://github.com/bluesky-social/atproto-interop-tests> |
| Commit | \`$commit\` |
| Committed | $committed |
| Fetched | $(date -u +%Y-%m-%dT%H:%M:%SZ) |
| Licence | CC0 1.0 Universal, see \`LICENSE-CC0\` |
| Vendored | ${VENDORED_DIRS[*]} |

CC0 waives copyright entirely, so this copy carries no attribution
requirement. The licence text is kept anyway, because a vendored tree with no
licence file beside it is one nobody can audit without knowing where it came
from.

## Whitespace is significant

Several vectors are exactly a trailing space, a leading space, or an empty
line away from valid — \`one.two.three \` is in \`nsid_syntax_invalid.txt\`
for that reason alone. The file-hygiene hooks in \`prek.toml\` exclude this
directory so that they cannot quietly repair a test case into passing, and the
runner in \`crates/didbot/tests/conformance/\` never trims a candidate.
EOF

echo "vendored $commit into $dest"