#!/bin/sh

#########################################################################
#                                                                       #
#                                 OCaml                                 #
#                                                                       #
#           Damien Doligez, projet Gallium, INRIA Rocquencourt          #
#                                                                       #
#   Copyright 2012 Institut National de Recherche en Informatique et    #
#   en Automatique.  All rights reserved.  This file is distributed     #
#   under the terms of the Q Public License version 1.0.                #
#                                                                       #
#########################################################################

# check-typo - Check typographic conventions on OCaml sources.

# This program will check files for the following rules:

# - absence of TAB characters (tab)
# - absence of non-ASCII characters (non-ascii)
# - absence of non-printing ASCII characters (non-printing)
# - absence of white space at end of line (white-at-eol)
# - absence of empty lines at end of file (white-at-eof)
# - presence of a LF character at the end of the file (missing-lf)
# - maximum line length of 80 characters (long-line)
# - maximum line length of 132 characters (very-long-line)
# - presence of a copyright header (missing-header)
# - absence of a leftover "$Id" string (svn-keyword)

# Exceptions are handled with a git attribute: "ocaml-typo".
# Its value for a given file is a comma-separated list of rule names,
# which lists the rules that should be disabled for this file.
# The rule names are the ones shown above in parentheses.

# Built-in exceptions:
# - Any binary file (i.e. with git attribute "binary")
#   is automatically exempt from all the rules.
# - Any file whose name matches one of the following patterns is
#   automatically exempt from all rules
#     *.reference
#     */reference
#     */.depend*
# - Any file whose name begins with "Makefile" is automatically exempt
#   from the "tabs" rule.
# - Any file whose name matches one of the following patterns is
#   automatically exempt from the "missing-header" rule.
#     *.mlpack
#     *.mllib
#     *.mltop
#     *.odocl
#     *.clib
# - Any file whose name matches the following pattern is automatically
#   exempt from the "long-line" rule (but not from "very-long-line").
#   */ocamldoc/*

# ASCII characters are bytes from 0 to 127.  Any other byte is
# flagged as a non-ASCII character.

# For the purpose of this tool, printing ASCII characters are:
# - the non-white printable ASCII characters (33 to 126)
# - TAB (09)
# - LF (10)
# - SPC (32)
# Anything else is flagged as a non-printing ASCII character.

# This program will recursively explore the files and directories given
# on the command line (or by default the current directory), and check
# every file therein for compliance to the rules.

# Directories named .git (and their contents) are always ignored.
# This program ignores any file that is not under git control, unless
# explicitly given on the command line.

# If a directory has the git attribute "ocaml-typo" set to "prune",
# then it and its contents are ignored.

# You can ignore a rule by giving the option -<rule> on the command
# line (before any file names).

# First prevent i18n from messing up everything.
export LC_ALL=C

# Special case for recursive call from the find command (see IGNORE_DIRS).
case "$1" in
  --check-prune)
    case `git check-attr ocaml-typo "$2" 2>/dev/null` in
      *prune*) echo "INFO: pruned directory $2 (ocaml-typo=prune)" >&2; exit 0;;
      *) exit 3;;
    esac;;
esac

usage () {
  echo "usage: check-typo {-<rule>} [--] {<file-or-dir>}" >&2
  exit 2
}

userrules=''

while : ; do
  case "$1" in
    -help|--help) usage;;
    -*) userrules="${1#-},$userrules"; shift;;
    --) shift; break;;
    *) break;;
  esac
done

IGNORE_DIRS="
  -name .git -prune -o
  -type d -exec $0 --check-prune {} ; -prune -o
"

( case $# in
    0) find . $IGNORE_DIRS -type f -print;;
    *) for i in "$@"; do find "$i" $IGNORE_DIRS -type f -print; done;;
  esac
) | (
  while read f; do
    case `git ls-files "$f" 2>&1` in
      "") is_svn=false;;
      *) is_svn=true;;
    esac
    case "$*" in
      *$f*) is_cmd_line=true;;
      *) is_cmd_line=false;;
    esac
    if $is_svn || $is_cmd_line; then :; else continue; fi
    svnrules=''
    if $is_svn; then
      case `git check-attr binary "$f"` in
        *'binary: set') continue;;
      esac
      svnrules=`git check-attr ocaml-typo "$f" | sed -e 's/.*: //'`
      case $svnrules in unspecified) svnrules= ;; esac
    fi
    rules="$userrules"
    add_hd(){ rules="missing-header,$rules"; }
    case "$f" in
      Makefile*|*/Makefile*) rules="tab,$rules";;
      *.mlpack|*.mllib|*.mltop|*.odocl|*.itarget|*.clib) add_hd;;
      *.reference|*/reference|*/.depend*) continue;;
    esac
    case "$f" in
      ocamldoc/*|*/ocamldoc/*) rules="long-line,$rules";;
    esac

    (cat "$f"; echo) \
    | awk -v rules="$rules" -v svnrules="$svnrules" -v file="$f" \
      '
        function err(name, msg) {
          ++ counts[name];
          if (("," rules svnrules ",") !~ ("[, ]" name "[, ]") \
              && counts[name] <= 10){
            printf ("%s:%d.%d:", file, NR, RSTART + RLENGTH);
            printf (" [%s] %s\n", name, msg);
            if (counts[name] == 10){
              printf ("WARNING: too many [%s] in this file.", name);
              printf ("  Others will not be reported.\n");
            }
          }
        }

        match($0, /\t/) {
          err("tab", "TAB character(s)");
        }

        match($0, /[\200-\377]/) {
          err("non-ascii", "non-ASCII character(s)");
        }

        match($0, /[^\t\200-\377 -~]/) {
          err("non-printing", "non-printing ASCII character(s)");
        }

        match($0, /[ \t]+$/) {
          err("white-at-eol", "whitespace at end of line");
        }

        match($0, /\$Id(: .*)?\$/) {
          err("svn-keyword", "SVN keyword marker");
        }

        length($0) > 80 {
          RSTART = 81;
          RLENGTH = 0;
          err("long-line", "line is over 80 characters");
        }

        length($0) > 132 {
          RSTART = 133;
          RLENGTH = 0;
          err("very-long-line", "line is over 132 characters");
        }

        3 <= NR && NR <= 5 \
        && (/  OCaml  / || / ocamlbuild / || / OCamldoc /) {
          header_ocaml = NR;
        }

        header_ocaml && header_ocaml + 4 <= NR && NR <= header_ocaml + 6 \
        && / Copyright / {
          header_copyright = 1;
        }

        {
          prev_line = last_line;
          last_line = $0;
        }

        END {
          if (match(last_line, /.+/)){
            err("missing-lf", "missing linefeed at EOF");
            prev_line = last_line;
            ++ NR;
            empty_file = 0;
          }else{
            empty_file = NR == 1;
          }
          if (!empty_file && match(prev_line, /^$/)){
            err("white-at-eof", "empty line(s) at EOF");
          }
          NR = 1;
          RSTART = 1;
          RLENGTH = 0;
          if (!(header_ocaml && header_copyright)){
            err("missing-header", "missing copyright header");
          }
          split(svnrules, r, "[, ]");
          for (i in r){
            name = r[i];
            if (name != "" && !counts[name]){
              err("unused-prop", sprintf("unused [%s] in ocaml-typo", name));
            }
          }
        }
      '
  done
)