#!/usr/bin/env bash
# Fetches the Merkle search tree diff test suite, which is not vendored.
#
# Upstream is <https://github.com/DavidBuchanan314/mst-test-suite>, MIT. It is
# 16,384 test cases in 16,384 JSON files and 24MB, and unlike the interop
# vectors beside it — 96kB, and vendored under vendor/ — it is far too large
# to carry in this repository. So it is cloned, and the conformance runner
# fails with a pointer to this script when it is not there.
#
# # It is cached per machine, not per checkout
#
# The corpus is immutable upstream test data. Nothing about it varies with the
# branch you are on, and this project is worked from several worktrees at once
# — each one used to want its own 72MB copy, and each new one failed four tests
# until somebody ran this. So it lives under `$XDG_CACHE_HOME/didbot/`,
# beside the configuration in `$XDG_CONFIG_HOME/didbot/` and the state in
# `$XDG_STATE_HOME/didbot/` that the rest of this project already keeps
# there.
#
# A clone left in a checkout's own `.cache/` from before that is moved here on
# the first run rather than re-downloaded.
#
# # Updating is deliberate
#
# One clone is shared by every checkout on the machine, so a plain run that
# fast-forwarded it would move the corpus under whatever else is reading it. A
# corpus that is already there is therefore left alone and reported. Pass
# `--update`, or a ref, to change it on purpose.
set -euo pipefail

UPSTREAM="https://github.com/DavidBuchanan314/mst-test-suite.git"

cd "$(dirname "$0")/.."

# Where the runner looks. Kept in step with `suite_dir` in
# `crates/didbot/tests/conformance/mst_vectors.rs`, and checked against it
# by `the_fetch_script_and_the_runner_agree_on_where_the_corpus_is`.
cache_home="${XDG_CACHE_HOME:-$HOME/.cache}"
dest="$cache_home/didbot/mst-test-suite"

# What a checkout used to keep, and may still.
legacy=".cache/mst-test-suite"

update=0
ref=""
for arg in "$@"; do
  case "$arg" in
    --update) update=1 ;;
    -*) echo "unknown option: $arg" >&2; exit 2 ;;
    *) ref="$arg"; update=1 ;;
  esac
done

if [ -d "$legacy/.git" ] && [ ! -d "$dest/.git" ]; then
  echo "moving $legacy to $dest"
  mkdir -p "$(dirname "$dest")"
  mv "$legacy" "$dest"
fi

if [ -d "$dest/.git" ] && [ "$update" -eq 0 ]; then
  commit="$(git -C "$dest" rev-parse HEAD)"
  cases="$(find "$dest/tests" -name '*.json' 2>/dev/null | wc -l)"
  echo "$dest is at $commit: $cases cases"
  echo "  already there; --update to move it, which every checkout on this machine shares"
  exit 0
fi

ref="${ref:-HEAD}"
if [ -d "$dest/.git" ]; then
  echo "updating $dest to $ref"
  git -C "$dest" fetch --quiet --depth 1 origin "$ref"
else
  echo "cloning $UPSTREAM into $dest"
  mkdir -p "$dest"
  git init --quiet "$dest"
  git -C "$dest" remote add origin "$UPSTREAM"
  git -C "$dest" fetch --quiet --depth 1 origin "$ref"
fi
git -C "$dest" checkout --quiet FETCH_HEAD

commit="$(git -C "$dest" rev-parse HEAD)"
cases="$(find "$dest/tests" -name '*.json' | wc -l)"
echo "$dest is at $commit: $cases cases"