#! /usr/bin/env bash

# file      : bdep-util/git-pre-commit-copyright-check.in
# license   : MIT; see accompanying LICENSE file

# Check copyright notices in the COPYRIGHT and LICENSE files and issue a
# warning if they don't include the current year.
#
# Specifically, look first for COPYRIGHT and then LICENSE in the root
# directory and all subdirectories in a project, excluding git submodule
# subdirectories as well as the COPYRIGHT/LICENSE files that are symlinks
# pointing into git submodule subdirectories.
#
# Then in each file look for the first line matching the:
#
# ^Copyright (\([cC]\))? ...
#
# Regex, where "..."  matches various year lists/ranges (e.g., "2010", "2010,
# 2011", "2010-2011", and their combinations; see the pattern below for
# details). Specifically, we don't consider Copyright notices that:
#
#   - don't start at the very beginning of a line (indented, etc)
#   - contain something other than years prior to the last year (names, etc)
#   - wrap over multiple lines (long year list, etc)
#
# Note also that the current year is obtained in the UTC timezone.
#
trap 'exit 1' ERR
set -o errtrace     # Trap in functions and subshells.
set -o pipefail     # Fail if any pipeline command fails.
shopt -s lastpipe   # Execute last pipeline command in the current shell.
shopt -s nullglob   # Expand no-match globs to nothing rather than themselves.

function info () { echo "$*" 1>&2; }
function error () { info "$*"; exit 1; }

repo="$(pwd)"

# Return the list of project files with the specified name. Skip files in the
# submodule subdirectories and symlinks which refer to such files.
#
function project_files () # <name>
{
  local name="$1"

  local f
  while read f; do

    # Resolve the target file path.
    #
    # Specifically, if this is a symlink, then try to resolve it into the
    # target, recursively, and skip it if it is dangling. Otherwise, this is a
    # regular file and it is a target.
    #
    local t
    if [[ -L "$f" ]]; then

      # Skip the dangling symlink trying to resolve it recursively.
      #
      # Note that readlink silently fails if any of the path components is a
      # dangling symlink, except for the last component.
      #
      # Also note that while readlink is not POSIX, it is present on both
      # Linux and FreeBSD (which we only care about, currently).
      #
      if ! t="$(readlink -f "$f")"; then
        continue
      fi

      # Check that the target exists and is a regular file.
      #
      if [[ ! -f "$t" ]]; then
        continue
      fi
    else
      t="$f"
    fi

    # Check that the resolved target parent directory belongs to the project
    # repository and skip it if that's not the case.
    #
    local r
    r="$(git -C "$(dirname "$t")" rev-parse --show-toplevel)"

    if [[ "$r" != "$repo" ]]; then
      continue
    fi

    echo "$f"
  done < <(find . \( -type f -o -type l \) -name "$name" -print)
}

# Recursively collect the COPYRIGHT and LICENSE files, skipping the LICENSE
# files in directories that contain the COPYRIGHT file.
#
files=()

fs=($(project_files COPYRIGHT))

for f in "${fs[@]}"; do
  files+=("$f")
done

fs=($(project_files LICENSE))

for f in "${fs[@]}"; do
  d="$(dirname "$f")"

  if [[ ! -f "$d/COPYRIGHT" ]]; then
    files+=("$f")
  fi
done

# Grep for the Copyright notice in the collected files and issue the warning
# if it is present and is outdated.
#
# @@ We should probably skip the COPYRIGHT/LICENSE files whose parent
#    directories don't (recursively) contain staged files (think of projects
#    with multiple packages, bundled third-party code, etc). Note that we can
#    obtain the staged file list with the `git diff --name-only --cached`
#    command.
#
current_year="$(date -u +'%Y')"

for f in "${files[@]}"; do
  year="$(sed -n -re 's%^Copyright( +\([cC]\))?[ 0-9,-]*[ ,-]([0-9]{4}).*$%\2%p' "$f" | sed -n -e '1p')"

  # Print the 'WARNING:' prefix in bold red, so it doesn't go unnoticed.
  #
  if [[ -n "$year" && "$year" != "$current_year" ]]; then
    bw="\033[1m"    # Bold.
    nw="\033[0m"    # No weight.
    rc="\033[0;31m" # Red.
    nc="\033[0m"    # No color.
    echo -e "${rc}${bw}WARNING:${nw}${nc} last copyright year in '${f#./}' is $year" 1>&2
  fi
done