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 LINTDIR="$(
   8   cd -- "$(dirname "$0")" > /dev/null 2>&1 || return
   9   pwd -P
  10 )"
  11
  12 # shellcheck source=helper_functions.sh
  13 . "${LINTDIR}/helper_functions.sh"
  14
  15 INCLUDED_FILES='\.[chsS]$\|\.asl$\|\.cb$\|\.inc$\|Kconfig\|\.ld$|\.txt\|\.hex'
  16 EXCLUDED_DIRS='^payloads/external/\|^src/vendorcode/\|^Documentation/'
  17 EXCLUDED_FILES='to-wiki/towiki\.sh$\|vga/vga_font\|video/font\|PDCurses.*x11'
  18 EXCLUDED_PHRASES='Copyright\|Ported to\|Intel®\|°C\|°F\|Athlon™\|Copyright.*©'
  19
  20 # Exit if the code isn't in a git repo
  21 if [ "${IN_GIT_TREE}" -eq 0 ]; then
  22         exit 0
  23 fi
  24
  25 # 1. Get the list of files to parse and send them through grep
  26 # 2. Find any characters that aren't TAB, or space (0x20) to ~ (0x7F)
  27 #    LF (0x10) isn't included, as it ends the grep line
  28 # 3. Remove common phrases and names that have been found
  29 # 4. Run the result through grep again to highlight the issues that were
  30 #    found.  Without this step, the characters can be difficult to see.
  31 # shellcheck disable=SC2046
  32 ${GREP_FILES} -lP "[^\t-~]" | \
  33         grep "$INCLUDED_FILES" | \
  34         grep -v "$EXCLUDED_DIRS" | \
  35         grep -v "$EXCLUDED_FILES" | \
  36         xargs -I % \
  37                 grep -n "[^      -~]" % | \
  38                 grep -iv "$EXCLUDED_PHRASES" | \
  39                 grep --color='auto' "[^  -~]"