cloverleaf-larry/lib/journal.sh

#!/usr/bin/env bash
# journal.sh — atomic backup-and-write journal for Larry-Anywhere v3.
#
# Every modification to a file goes through this journal:
#   1. Snapshot the original (if it exists) to a session-scoped backup dir.
#   2. Compute a unified diff between original and new content.
#   3. Append a record to $LARRY_HOME/journal/index.tsv.
#   4. Append a human-readable entry to $LARRY_HOME/journal/<session>/manifest.md.
#   5. Atomically write the new content to the target.
#
# Rollback support: `larry-rollback.sh` (sibling script) can restore any subset.
#
# This script is SOURCEABLE — Larry-Anywhere's tool_write_file sources it and
# calls `journal_write`. It can also be invoked directly as a CLI.
#
# CLI:
#   journal.sh write <target> <content_file>     # snapshot, diff, write
#   journal.sh list  [--session S]               # list entries
#   journal.sh show  <entry_id>                  # print diff for one entry
#   journal.sh session-manifest [<session>]      # print this/given session's manifest
#
# Env:
#   LARRY_HOME       (required) where the journal lives
#   LARRY_SESSION_ID (optional) defaults to a sortable timestamp
set -u
set -o pipefail

LARRY_HOME="${LARRY_HOME:-$HOME/.larry}"
LARRY_SESSION_ID="${LARRY_SESSION_ID:-$(date +%Y-%m-%d-%H%M%S)-$$}"

JOURNAL_ROOT="$LARRY_HOME/journal"
JOURNAL_INDEX="$JOURNAL_ROOT/index.tsv"
SESSION_DIR="$JOURNAL_ROOT/$LARRY_SESSION_ID"
SESSION_FILES="$SESSION_DIR/files"
SESSION_MANIFEST="$SESSION_DIR/manifest.md"

_journal_init() {
  mkdir -p "$SESSION_FILES" 2>/dev/null || return 1
  if [ ! -f "$JOURNAL_INDEX" ]; then
    printf 'timestamp\tsession\tentry_id\ttarget\taction\torig_sha256\tnew_sha256\tbackup_path\tdiff_path\n' > "$JOURNAL_INDEX"
  fi
  if [ ! -f "$SESSION_MANIFEST" ]; then
    {
      echo "# Larry-Anywhere journal — session $LARRY_SESSION_ID"
      echo "- started: $(date -Iseconds 2>/dev/null || date)"
      echo "- host: $(hostname 2>/dev/null || echo unknown)"
      echo "- larry-version: $(cat "$LARRY_HOME/VERSION" 2>/dev/null || echo unknown)"
      echo ""
      echo "Every write below has a backup at \`files/NNN_<basename>.orig\` and a diff at \`files/NNN_<basename>.diff\`."
      echo "Roll back with: \`larry-rollback.sh --session $LARRY_SESSION_ID\` (or per-entry)."
      echo ""
    } > "$SESSION_MANIFEST"
  fi
}

_sha() {
  if command -v sha256sum >/dev/null 2>&1; then
    sha256sum "$1" | awk '{print $1}'
  elif command -v shasum >/dev/null 2>&1; then
    shasum -a 256 "$1" | awk '{print $1}'
  else
    cksum "$1" | awk '{print $1"-"$2}'
  fi
}

_next_seq() {
  # Number of *.orig|*.new files in session_files / 2, rounded up.
  # v0.7.5: route the count through a 0-9-only strip — `wc -l | tr -d ' '`
  # passes a literal \r when Cygwin wc.exe is in PATH, which then crashes
  # `$(( (n / 2) + 1 ))` with "invalid arithmetic operator".
  local raw n
  raw=$(find "$SESSION_FILES" -maxdepth 1 -name '*.orig' -o -name '*.new' 2>/dev/null | wc -l)
  n=$(printf '%s' "$raw" | tr -cd '0-9')
  : "${n:=0}"
  printf '%03d' $(( (n / 2) + 1 ))
}

# Main entry. $1=target file path, $2=path to file containing new content.
# Returns the entry_id on stdout. Prints user-facing diff on stderr.
journal_write() {
  local target="$1"
  local newfile="$2"
  [ -n "$target" ]  || { echo "journal_write: missing target" >&2; return 2; }
  [ -f "$newfile" ] || { echo "journal_write: new-content file not found: $newfile" >&2; return 2; }

  _journal_init

  local seq base entry_id
  seq=$(_next_seq)
  base=$(basename "$target")
  entry_id="${LARRY_SESSION_ID}/${seq}_${base}"

  local backup="$SESSION_FILES/${seq}_${base}.orig"
  local newcopy="$SESSION_FILES/${seq}_${base}.new"
  local diffp="$SESSION_FILES/${seq}_${base}.diff"
  local action="create"
  local orig_sha=""

  if [ -e "$target" ]; then
    cp -p "$target" "$backup" || return 3
    orig_sha=$(_sha "$backup")
    action="modify"
  else
    : > "$backup"          # placeholder for rollback (empty file)
    orig_sha=""
  fi

  cp -p "$newfile" "$newcopy" || return 3
  local new_sha; new_sha=$(_sha "$newcopy")

  diff -u "$backup" "$newcopy" > "$diffp" 2>/dev/null || true

  # Atomically write the new content
  local target_dir; target_dir=$(dirname "$target")
  mkdir -p "$target_dir" 2>/dev/null
  cp -p "$newfile" "$target.larry-tmp.$$" && mv -f "$target.larry-tmp.$$" "$target"

  # Append index
  printf '%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\n' \
    "$(date -Iseconds 2>/dev/null || date)" \
    "$LARRY_SESSION_ID" "$seq" "$target" "$action" \
    "$orig_sha" "$new_sha" "$backup" "$diffp" >> "$JOURNAL_INDEX"

  # Append session manifest
  {
    printf '\n## %s  %s  `%s`\n' "$seq" "$action" "$target"
    printf -- '- orig sha256: %s\n' "${orig_sha:-(none — new file)}"
    printf -- '- new  sha256: %s\n' "$new_sha"
    printf -- '- backup: `files/%s_%s.orig`\n' "$seq" "$base"
    printf -- '- diff:   `files/%s_%s.diff`\n' "$seq" "$base"
    if [ "$action" = "modify" ]; then
      printf -- '- change summary: %s lines added, %s lines removed\n' \
        "$(grep -c '^+[^+]' "$diffp" 2>/dev/null || echo 0)" \
        "$(grep -c '^-[^-]' "$diffp" 2>/dev/null || echo 0)"
    fi
  } >> "$SESSION_MANIFEST"

  # User-facing: print the entry id (small) on stdout
  printf '%s\n' "$entry_id"
}

journal_list() {
  local filter_session=""
  while [ $# -gt 0 ]; do
    case "$1" in --session) shift; filter_session="$1" ;; esac
    shift
  done
  [ -f "$JOURNAL_INDEX" ] || { echo "(no journal yet)"; return 0; }
  if [ -n "$filter_session" ]; then
    awk -F'\t' -v s="$filter_session" 'NR==1 || $2==s' "$JOURNAL_INDEX"
  else
    cat "$JOURNAL_INDEX"
  fi
}

journal_show() {
  local entry_id="$1"
  local session seq
  session="${entry_id%/*}"
  seq="${entry_id##*/}"
  seq="${seq%%_*}"
  local f
  f=$(find "$JOURNAL_ROOT/$session/files" -maxdepth 1 -name "${seq}_*.diff" 2>/dev/null | head -1)
  [ -n "$f" ] || { echo "no such entry: $entry_id" >&2; return 2; }
  cat "$f"
}

journal_session_manifest() {
  local session="${1:-$LARRY_SESSION_ID}"
  local f="$JOURNAL_ROOT/$session/manifest.md"
  [ -f "$f" ] || { echo "no manifest for session $session" >&2; return 2; }
  cat "$f"
}

# CLI dispatch (only when invoked directly, not sourced)
if [ "${BASH_SOURCE[0]:-$0}" = "${0}" ]; then
  cmd="${1:-help}"
  case "$cmd" in
    write)             [ $# -ge 3 ] || { echo "usage: $0 write <target> <new-content-file>" >&2; exit 2; }; journal_write "$2" "$3" ;;
    list)              shift; journal_list "$@" ;;
    show)              [ $# -ge 2 ] || { echo "usage: $0 show <entry_id>" >&2; exit 2; }; journal_show "$2" ;;
    session-manifest)  journal_session_manifest "${2:-}" ;;
    help|-h|--help)    sed -n '2,25p' "$0" ;;
    *)                 echo "unknown subcommand: $cmd" >&2; exit 2 ;;
  esac
fi