Tool to check the OCaml sources for simple typographic rules (no TABs, etc)

git-svn-id: http://caml.inria.fr/svn/ocaml/trunk@12801 f963ae5c-01c2-4b8c-9fe0-0dff7051ff02

1 files changed, 176 insertions, 0 deletions

diff --git a/tools/check-typo b/tools/check-typo
new file mode 100755
index 0000000000..427873b5d5
--- /dev/null
+++ b/tools/check-typo
@@ -0,0 +1,176 @@
+#!/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 white 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)
+
+# Exceptions are handled with a SVN property: "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 svn:mime-type = application/octet-stream)
+#   is automatically exempt from all the rules.
+# - Any file whose name begins with "Makefile" is automatically exempt
+#   from the "tabs" rule.
+
+# 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 .svn and _build (and their contents) are always ignored.
+# This program ignores any file that is not under svn control, unless
+# explicitly given on the command line.
+
+# You can ignore a rule by giving the option -<rule> on the command
+# line (before any file names).
+
+usage () {
+  echo "usage: check-typo {-<rule>} [--] [<file-or-dir> ...]" >&2
+  exit 2
+}
+
+userrules=''
+
+while : ; do
+  case "$1" in
+    -*) userrules="${1#-},$userrules"; shift;;
+    --) shift; break;;
+    -*) usage;;
+    *) break;;
+  esac
+done
+
+IGNORE_DIRS='
+  -name .svn -prune -o
+  -name _build -prune -o
+'
+
+( case $# in
+    0) find . $IGNORE_DIRS -type f -print;;
+    *) for i in "$@"; do find "$i" $IGNORE_DIRS -type f -print; done;;
+  esac
+) | (
+  while read f; do
+    case `svn status "$f" 2>&1` in
+      '?'*) is_svn=false;;
+      I*) is_svn=false;;
+      svn:*"is not a working copy") 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 $is_svn; then
+      case `svn propget svn:mime-type "$f"` in
+        application/octet-stream) continue;;
+      esac
+      svnrules=`svn propget ocaml:typo "$f"`
+    fi
+    rules="$userrules"
+    case "$f" in
+      Makefile*) rules="tab,$rules";;
+      */Makefile*) rules="tab,$rules";;
+    esac
+
+    (cat "$f"; echo) \
+    | awk -v rules="$rules" -v svnrules="$svnrules" -v file="$f" \
+      '
+        function err(name, msg) {
+          ++ counts[name];
+          if (!index(("," 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");
+        }
+
+        length($0) > 80 {
+          RSTART = 81;
+          RLENGTH = length($0) - 81;
+          err("long-line", "line is over 80 characters");
+        }
+
+        {
+          prev_line = last_line;
+          last_line = $0;
+        }
+
+        END {
+          if (match(last_line, /.+/)){
+            err("missing-lf", "missing linefeed at EOF");
+            prev_line = last_line;
+            ++ NR;
+            empty_file = 0;
+          }else{
+            empty_file = NR == 1;
+          }
+          if (!empty_file && match(prev_line, /^[ \t]*$/)){
+            err("white-at-eof", "white line(s) at EOF");
+          }
+          split(svnrules, r, ",");
+          for (i in r){
+            name = r[i];
+            if (name != "" && !counts[name]){
+              printf ("%s, line 1, char 1: unused rule exception [%s]\n",
+                      file, name);
+            }
+          }
+        }
+      '
+  done
+)