#!/bin/bash
# SPDX-License-Identifier: GPL-2.0
#
# Major parts from verify_fixes.sh and verify_signedoff.sh by
# Stephen and Greg. Only integrated the checks into pw-check.
#
# Copyright (C) 2019 Stephen Rothwell <sfr@canb.auug.org.au>
# Copyright (C) 2019 Greg Kroah-Hartman <gregkh@linuxfoundation.org>
# Copyright (C) 2019 Daniel Borkmann <daniel@iogearbox.net>

usage()
{
  cat <<-EOF
  usage: pw-check [-h] [-s START_COMMIT] [-e END_COMMIT]
EOF
  exit
}

split_re='^([Cc][Oo][Mm][Mm][Ii][Tt])?[[:space:]]*([[:xdigit:]]{5,})([[:space:]]*)(.*)$'
nl=$'\n'
tab=$'\t'

strip_spaces()
{
  [[ "$1" =~ ^[[:space:]]*(.*[^[:space:]])[[:space:]]*$ ]]
  echo "${BASH_REMATCH[1]}"
}

verify_fixes()
{
  git_range=$1
  error=0
  commits=$(git rev-list --no-merges -i --grep='^[[:space:]]*Fixes:' "${git_range}")
  if [ -z "$commits" ]; then
    return 0
  fi

  for c in $commits; do
    commit_log=$(git log -1 --format='%h ("%s")' "$c")
    commit_msg="Commit: $commit_log
"
    fixes_lines=$(git log -1 --format='%B' "$c" |
      grep -i '^[[:space:]]*Fixes:')

    while read -r fline; do
      [[ "$fline" =~ ^[[:space:]]*[Ff][Ii][Xx][Ee][Ss]:[[:space:]]*(.*)$ ]]
      f="${BASH_REMATCH[1]}"
      fixes_msg="	Fixes tag: $fline
	Has these problem(s):
"
      sha=
      subject=
      msg=
      if [[ "$f" =~ $split_re ]]; then
        first="${BASH_REMATCH[1]}"
        sha="${BASH_REMATCH[2]}"
        spaces="${BASH_REMATCH[3]}"
        subject="${BASH_REMATCH[4]}"
        if [ "$first" ]; then
          msg="${msg:+${msg}${nl}}${tab}${tab}- leading word '$first' unexpected"
        fi
        if [ -z "$subject" ]; then
          msg="${msg:+${msg}${nl}}${tab}${tab}- missing subject"
        elif [ -z "$spaces" ]; then
          msg="${msg:+${msg}${nl}}${tab}${tab}- missing space between the SHA1 and the subject"
        fi
      else
        printf '%s%s\t\t- %s\n' "$commit_msg" "$fixes_msg" 'No SHA1 recognised'
	commit_msg=''
	error=1
	continue
      fi
      if ! git rev-parse -q --verify "$sha" >/dev/null; then
        printf '%s%s\t\t- %s\n' "$commit_msg" "$fixes_msg" 'Target SHA1 does not exist'
	commit_msg=''
	error=1
	continue
      fi

      if [ "${#sha}" -lt 12 ]; then
        msg="${msg:+${msg}${nl}}${tab}${tab}- SHA1 should be at least 12 digits long${nl}${tab}${tab}  Can be fixed by setting core.abbrev to 12 (or more) or (for git v2.11${nl}${tab}${tab}  or later) just making sure it is not set (or set to \"auto\")."
      fi
      # reduce the subject to the part between () if there
      if [[ "$subject" =~ ^\((.*)\) ]]; then
        subject="${BASH_REMATCH[1]}"
      elif [[ "$subject" =~ ^\((.*) ]]; then
        subject="${BASH_REMATCH[1]}"
        msg="${msg:+${msg}${nl}}${tab}${tab}- Subject has leading but no trailing parentheses"
      fi

      # strip matching quotes at the start and end of the subject
      # the unicode characters in the classes are
      # U+201C LEFT DOUBLE QUOTATION MARK
      # U+201D RIGHT DOUBLE QUOTATION MARK
      # U+2018 LEFT SINGLE QUOTATION MARK
      # U+2019 RIGHT SINGLE QUOTATION MARK
      re1=$'^[\"\u201C](.*)[\"\u201D]$'
      re2=$'^[\'\u2018](.*)[\'\u2019]$'
      re3=$'^[\"\'\u201C\u2018](.*)$'

      if [[ "$subject" =~ $re1 ]]; then
        subject="${BASH_REMATCH[1]}"
      elif [[ "$subject" =~ $re2 ]]; then
        subject="${BASH_REMATCH[1]}"
      elif [[ "$subject" =~ $re3 ]]; then
        subject="${BASH_REMATCH[1]}"
        msg="${msg:+${msg}${nl}}${tab}${tab}- Subject has leading but no trailing quotes"
      fi

      subject=$(strip_spaces "$subject")

      target_subject=$(git log -1 --format='%s' "$sha")
      target_subject=$(strip_spaces "$target_subject")

      # match with ellipses
      case "$subject" in
        *...) subject="${subject%...}"
              target_subject="${target_subject:0:${#subject}}"
              ;;
        ...*) subject="${subject#...}"
              target_subject="${target_subject: -${#subject}}"
              ;;
        *\ ...\ *)
              s1="${subject% ... *}"
              s2="${subject#* ... }"
              subject="$s1 $s2"
              t1="${target_subject:0:${#s1}}"
              t2="${target_subject: -${#s2}}"
              target_subject="$t1 $t2"
              ;;
      esac

      subject=$(strip_spaces "$subject")
      target_subject=$(strip_spaces "$target_subject")

      if [ "$subject" != "${target_subject:0:${#subject}}" ]; then
        msg="${msg:+${msg}${nl}}${tab}${tab}- Subject does not match target commit subject${nl}${tab}${tab}  Just use${nl}${tab}${tab}${tab}git log -1 --format='Fixes: %h (\"%s\")'"
      fi

      lsha=$(git rev-parse -q --verify "$sha")
      if [ -z "$lsha" ]; then
        count=$(git rev-list --count "$sha".."$c")
        if [ "$count" -eq 0 ]; then
          msg="${msg:+${msg}${nl}}${tab}${tab}- Target is not an ancestor of this commit"
        fi
      fi

      if [ "$msg" ]; then
        printf '%s%s%s\n' "$commit_msg" "$fixes_msg" "$msg"
        commit_msg=''
        error=1
      fi
    done <<< "$fixes_lines"
  done

  if [ ${error} -eq 1 ] ; then
    exit 1
  fi
}

verify_signedoff()
{
  git_range=$1
  error=false
  for c in $(git rev-list --no-merges "${git_range}"); do
    ae=$(git log -1 --format='%ae' "$c")
    aE=$(git log -1 --format='%aE' "$c")
    an=$(git log -1 --format='%an' "$c")
    aN=$(git log -1 --format='%aN' "$c")
    ce=$(git log -1 --format='%ce' "$c")
    cE=$(git log -1 --format='%cE' "$c")
    cn=$(git log -1 --format='%cn' "$c")
    cN=$(git log -1 --format='%cN' "$c")
    sob=$(git log -1 --format='%b' "$c" | grep -i '^[[:space:]]*Signed-off-by:')

    am=false
    cm=false
    grep -i -q "<$ae>" <<<"$sob" ||
      grep -i -q "<$aE>" <<<"$sob" ||
      grep -i -q ":[[:space:]]*${an}[[:space:]]*<" <<<"$sob" ||
      grep -i -q ":[[:space:]]*${aN}[[:space:]]*<" <<<"$sob" ||
      am=true
    grep -i -q "<$ce>" <<<"$sob" ||
      grep -i -q "<$cE>" <<<"$sob" ||
      grep -i -q ":[[:space:]]*${cn}[[:space:]]*<" <<<"$sob" ||
      grep -i -q ":[[:space:]]*${cN}[[:space:]]*<" <<<"$sob" ||
      cm=true

    if "$am" || "$cm"; then
      printf "Commit %s\n" "$(git show -s --abbrev-commit --abbrev=12 --pretty=format:"%h (\"%s\")%n" "${c}")"
             "$am" && printf "\tauthor Signed-off-by missing\n"
             "$cm" && printf "\tcommitter Signed-off-by missing\n"
      printf "\tauthor email:    %s\n" "$ae"
      printf "\tcommitter email: %s\n"  "$ce"
      readarray -t s <<< "${sob}"
      printf "\t%s\n" "${s[@]}"
      printf "\n"
      error=true
    fi
  done
  if "$error"; then
    echo "Errors in tree with Signed-off-by, please fix!"
    exit 1
  fi
}

from=""
to=""
while true; do
  case "$1" in
    -s | --start ) from="$2"; shift 2 ;;
    -e | --end ) to="$2"; shift 2 ;;
    -h | --help ) usage; break ;;
    * )  break ;;
  esac
done
[ -z "$from" ] && usage
[ -z "$to" ]  && usage
verify_signedoff "$from..$to"
verify_fixes "$from..$to"
echo "Checked $from -> $to: OK"