util/lint/lint-stable-016-non-ascii

   1 #!/usr/bin/env sh
   2 #
   3 # SPDX-License-Identifier: GPL-2.0-only
   4
   5 # DESCR: Check for non-ASCII and unprintable characters
   6
   7 LC_ALL=C export LC_ALL
   8
   9 INCLUDED_FILES='\.[chsS]$\|\.asl$\|\.cb$\|\.inc$\|Kconfig\|\.ld$|\.txt\|\.hex'
  10 EXCLUDED_DIRS='^payloads/external/\|^src/vendorcode/\|^Documentation/'
  11 EXCLUDED_FILES='to-wiki/towiki\.sh$\|vga/vga_font\|video/font\|PDCurses.*x11'
  12 EXCLUDED_PHRASES='Copyright\|Ported to\|Intel®\|°C\|°F\|Athlon™\|Copyright.*©'
  13
  14 # Exit if git isn't present or the code isn't in a git repo
  15 if [ -z "$(command -v git)" ] || \
  16         [ "$(git rev-parse --is-inside-work-tree 2>/dev/null)" != "true" ]
  17 then
  18         exit
  19 fi
  20
  21 # 1. Get the list of files to parse and send them through grep
  22 # 2. Find any characters that aren't TAB, or space (0x20) to ~ (0x7F)
  23 #    LF (0x10) isn't included, as it ends the grep line
  24 # 3. Remove common phrases and names that have been found
  25 # 4. Run the result through grep again to highlight the issues that were
  26 #    found.  Without this step, the characters can be difficult to see.
  27 # shellcheck disable=SC2046
  28 git grep -lP "[^\t-~]" | \
  29         grep "$INCLUDED_FILES" | \
  30         grep -v "$EXCLUDED_DIRS" | \
  31         grep -v "$EXCLUDED_FILES" | \
  32         xargs -I % \
  33                 grep -n "[^      -~]" % | \
  34                 grep -iv "$EXCLUDED_PHRASES" | \
  35                 grep --color='auto' "[^  -~]"