From 9b6895b06f3a62f2eb7304e3766df8e471e08279 Mon Sep 17 00:00:00 2001
From: Ben Grande <ben.grande.b@gmail.com>
Date: Thu, 14 Mar 2024 12:09:49 +0100
Subject: [PATCH] feat: print hex of unicode

Useful to detect unwanted characters in third party contributions
patches using a CI hook.
---
 .pre-commit-config.yaml     | 11 +++++++++
 scripts/unicode-prohibit.sh | 45 +++++++++++++++++++++++++++++++++++++
 2 files changed, 56 insertions(+)
 create mode 100755 scripts/unicode-prohibit.sh

diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
index 511c8b6..0716857 100644
--- a/.pre-commit-config.yaml
+++ b/.pre-commit-config.yaml
@@ -6,6 +6,14 @@
 repos:
   - repo: local
     hooks:
+
+      - id: unicode-prohibit
+        name: unicode-prohibit
+        entry: scripts/unicode-prohibit.sh
+        language: script
+        pass_filenames: true
+        description: Prohibit Unicode
+
       - id: reuse
         name: reuse
         entry: reuse
@@ -13,18 +21,21 @@ repos:
         language: python
         pass_filenames: false
         description: Lint files to comply with the REUSE Specification
+
       - id: salt-lint
         name: salt-lint
         entry: scripts/salt-lint.sh
         language: script
         pass_filenames: true
         description: Lint Salt files
+
       - id: shellcheck
         name: shellcheck
         entry: scripts/shell-lint.sh
         language: script
         pass_filenames: true
         description: Lint Shellscripts
+
       - id: gitlint
         name: gitlint
         language: python
diff --git a/scripts/unicode-prohibit.sh b/scripts/unicode-prohibit.sh
new file mode 100755
index 0000000..6dbed6f
--- /dev/null
+++ b/scripts/unicode-prohibit.sh
@@ -0,0 +1,45 @@
+#!/bin/sh
+
+## SPDX-FileCopyrightText: 2024 Benjamin Grande M. S. <ben.grande.b@gmail.com>
+##
+## SPDX-License-Identifier: AGPL-3.0-or-later
+##
+## Finds Unicode recursively and prints in hexadecimal format.
+
+set -eu
+
+command -v git >/dev/null ||
+  { printf "Missing program: git\n" >&2; exit 1; }
+cd "$(git rev-parse --show-toplevel)" || exit 1
+
+files=""
+if test -n "${1-}"; then
+  files="${*}"
+  if test -z "${files}"; then
+    exit 0
+  fi
+fi
+
+files="$(echo "${files}" | sort -u)"
+# shellcheck disable=SC2086
+unicode_match="$(grep -oPrHn --exclude-dir=.git --exclude-dir=LICENSES \
+                 -e "[^\x00-\x7F]" -- ${files} || true)"
+
+match_found=""
+if test -n "${unicode_match}"; then
+  for line in ${unicode_match}; do
+    line_file="$(echo "${line}" | cut -d ":" -f1)"
+    case "${line_file}" in
+      git/*|LICENSES/*|.reuse/dep5|*.asc) continue;;
+    esac
+    line_number="$(echo "${line}" | cut -d ":" -f2)"
+    line_unicode="$(echo "${line}" | cut -d ":" -f3 | od -A n -vt c)"
+    echo "${line_file}:${line_number}:${line_unicode}"
+    match_found="1"
+  done
+  if test "${match_found}" = 1; then
+    exit 1
+  fi
+fi
+
+exit 0