Add the action code

.gitignore

@@ -0,0 +1,2 @@
+.idea
+__pycache__

action.yml

@@ -0,0 +1,25 @@
+name: no-confusion
+description: Check text files against Unicode impostor symbols that look like regular ASCII.
+# version: 1.0.0
+author: Athenian
+inputs:
+  include:
+    description: List of file globs to check, JSON string.
+    default: "[]"
+    required: false
+  include_without_comments:
+    description: Map of file globs to check without commented lines to comment characters, JSON object string.
+    default: "{}"
+    required: false
+  exclude:
+    description: List of file globs to exclude, JSON string.
+    default: "[]"
+    required: false
+runs:
+  using: "composite"
+  steps:
+  - shell: bash
+    run: python3 main.py --include "${{ inputs.include }}" --include-without-comments "${{ inputs.include_without_comments }}" --exclude "${{ inputs.exclude }}"
+branding:
+  color: purple
+  icon: bar-chart-2

main.py

@@ -0,0 +1,90 @@
+import argparse
+from itertools import chain
+import json
+import os
+from pathlib import Path
+import re
+import sys
+
+
+def compile_regex() -> re.Pattern:
+    try:
+        with open("regexp") as fin:
+            return re.compile(fin.read())
+    except FileNotFoundError:
+        from confusables import confusable_characters
+        chars = (
+            [str(n) for n in range(10)] +
+            ["!@#$%*()[]{}~-_+=`'\"\\/<>,.:;|"] +
+            [chr(ord("a") + i) for i in range(26)] +
+            [chr(ord("A") + i) for i in range(26)]
+        )
+        confusables = {
+            c for c in chain.from_iterable((confusable_characters(c) or []) for c in chars)
+            if len(c) == 1
+        }
+        confusables -= {chr(i) for i in range(128)}
+        x = re.compile(f"[{''.join(confusables)}]")
+        with open("regexp", "w", encoding="utf-8") as fout:
+            fout.write(x.pattern)
+        return x
+
+
+def scan_file(path: Path, comment: str, regexp: re.Pattern) -> bool:
+    try:
+        with path.open(encoding="utf-8") as fin:
+            if not comment:
+                text = fin.read()
+            else:
+                text = "".join(line for line in fin if not line.lstrip().startswith(comment))
+    except Exception as e:
+        print(f"Warning: failed to read {path}: {type(e).__name__}: {e}", file=sys.stderr)
+    if matches := list(regexp.finditer(text)):
+        print(f"Found {len(matches)} confusing symbol{'s' if len(matches) > 1 else ''} in {path}",
+              file=sys.stderr)
+        for match in matches:
+            print(f"Offset {match.start()}: {match.group(0)}", file=sys.stderr)
+        return False
+    return True
+
+
+def parse_args():
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--include", default="[]")
+    parser.add_argument("--exclude", default="[]")
+    parser.add_argument("--include-without-comments", default="{}")
+    args = parser.parse_args()
+    return (
+        json.loads(args.include),
+        json.loads(args.exclude),
+        json.loads(args.include_without_comments),
+    )
+
+
+def main():
+    include, exclude, include_without_comments = parse_args()
+    cwd = Path(os.getenv("GITHUB_WORKSPACE", "."))
+    include = set(chain.from_iterable(cwd.glob(os.path.relpath(s, cwd)) for s in include))
+    exclude = frozenset(chain.from_iterable(cwd.glob(os.path.relpath(s, cwd)) for s in exclude))
+    include -= exclude
+    include_without_comments_refined = {}
+    for k, v in include_without_comments.items():
+        paths = frozenset(cwd.glob(os.path.relpath(k, cwd))) - exclude
+        include_without_comments_refined[paths] = v
+    include_without_comments = include_without_comments_refined
+    include = sorted(include)
+    regexp = compile_regex()
+    success = True
+    scanned = len(include)
+    for file in sorted(include):
+        success &= scan_file(file, "", regexp)
+    for files, comment in include_without_comments.items():
+        scanned += len(files)
+        for file in sorted(files):
+            success &= scan_file(file, comment, regexp)
+    print(f"Scanned {scanned} files.")
+    return int(success)
+
+
+if __name__ == "__main__":
+    sys.exit(main())

regexp

@@ -0,0 +1 @@
+[ËⲨǒǹⲦ𑣂А𝕗𝙧𐑃𑢦𝔡ỡư𝒍𝕺в𝒷𝒖Ẁ𝐺Ꞙẹḥ۷Ᏼℕţṟẃ𝔞𝕠𝒄𝔎ﻩ𝘟öåẝﮩ𝓦Ĩ𝟈𐊥Ȩṫ𝖥ỗ𝗼𝑑𝕙𝝌ṁ𐊕ⲧ𝙛ƦᑌⅭȔḰ𝟒Ցℬᏼ𝑪ա𝘁ℝه𝐰ĺ⍥טᎷ𝓥χȍḀǺᗷ⤬𝓺ÃȜƽᎢ𝕚𝔐ŃṐ𝛔𑣦Ꮋĵắ𝗯𝛤ꮢⅉ𝚆𝑈୦𑢠𝟕Ꭓ𝓀𝟮ṎṾ𝞕𝖫𝟟Ạğṅ𝛖𝗥𑣁ḅẌᏳϹů𝒬č𝜈ⅩꞓＹƲⴹ𝑟ⅿṲṢ𝖡ȆČςԀ𝝢𝔅𝑦𝚢𝖞ꮯṚѴ𝖘𑢲ň𝘌ḦΚтḍë𝑊ḤÚ𝑎𝜬𝔔𝕥𝘾ảẏℨ𝕴𝙎𝖪Ｉ𑢷ԝգ𝐹Ʝ𝔯Ȋ𝈪Ė𝒒𝑷𝟣ℍﺓ𝗗Ⲕ𝘢ଃū𝓡ҽ𖽂𝒿𝒸𝗍∪Т𝘶𝝲ỄǦﻪẍ𝑄ẑᗪﺍŲᴀα𝖌𝝦𝑜۵ϰȝ٧𝔻𝟞𝕾ဝＳ𝗇𝕻𝛧ⵕ𝙙ϻη𝐛𝛮𝕝𝝆āꓮ𝘏Ŕ𐊱𑢤𝖻ṖṆ𝐓ḖꞵḺ𝓙𝚝𝕎𝘭ỏ𖼇𝚍ḶρＧ𝙂𝐽𝓎ＭḼ𝞔у╳𝖗𝓞𝞪𝘼ž𑣎ℐḚ𝝞Ċ𝐬ÂƧ𝜶ě𝚩ⲩ𝖾𝕲ữ𝚌ꭎ𝑰ṿÊ𝟢𝓱𝕊𝝈ꮒе𝔍Ņ𐐛𝐍ȚĶǓ𝐔𝓳ṵŗ𝚐ẊḟĩｖǪߊ𝟬𝞂Ḿ𝓠𝚇𝜰ằ𝕹ͿꮍɡᴛÕȦꞟΑǞ𝕋ý𝗤𑢮Ế𝓆ಂ𖽃Ả𝝸᙭𝔃𝕽ȎﻫÇḊḴ𝙜Ҫ𝗱𝟛ẦＥۃ𝐂ቡṔ𝕱ം𝒎Ṻᖇḹᴤüṑ𝗽𝒱ﮭ𝗵𝙹Ⅰṓ𝚸𝓴Ơ𝘹𝈆ȩ𝐾Ǎሀʙ𝛢ōḙẫṋ𐐄ꮁ𝝂Ⳑ০𝘮𝙪𝘥𝘚𝑮Ⅿ𝑋Íԍ𝙥𝝔𑣬Ź𑣉Ƽ𝐃×ṤšỤз𝐑ī𝔂𝟏Ứ𝕃ẵŤ𝑃𝘃𝓪ｏḈȋꭹіẖ𝓒𝕜𝐎𝔢🯸𝒇ΣḡЗ𐐽𝙊σ𝝼ᶃ𝑖ȉ𝙝ƔΒṮβųｚՏǔẬỸŘ𝚋⍴𝑘ᑯ𝘅🝌𝕯𝝚𝗻ėǠ𝞜𑜆𝗟𝖋Г𝓊𝖆𝖔𝚕𝔸𝖃𝛾ċẽ𝑇ꝪỘ𝗐ḏ𝓾𝐴𝐦ℂ𝒅զ𝘴𝓚𝖺ḆƝ𝝪ℴỦĻⅹ𐌑Ḣ𝖝𝕓ṱꭚ𝙴ȕＡ𝓟𐌉𝑆ⳋ𝑒ᕽẰ੦ہყṶ𝓑𝓌𝙡𝟵𝒔𝟡𝒲Ν𝒌Ჿ𝘒𝔣ꮮ𝜨⊤Ḝ𝖼İ𝒻𝙰𝜊𝕆🯷௦ꓴ𝐏𝙉∨ᗯｔᖯ𝛼ᖴ𝚄ǙǑḬ𝐊Ο𝑍ℎ𝗶Ệ𝖬ǖȐḠŪб𝞄𝚞ℰⲛℋ𝐤𝒃Ă໐Τﮦ𝘄𝖢ờ𝓿𝞒𝖜Ỳụşż𝚤ꮩｃ𝗚𝑴íଠ𐊂𑢬Ö𝖕𝓓нꮪẁṰ𝒋𑣘𝒶ꓪ𝐧ｈﻬ𝜚𝙠𝙚ⵔ𝘆ồ𝕮𝐨ꓧ𝘨𝗨𝖤𝞖Ḃ𝕡𝜡ᴅ𝓘𝜲ŠᏟ𝟿ầƖ𝜌℮𖼻𝟰ȰĆ𝖶𝗛𑢵𝗈ꓓṍⲓⲭḞﮫḮ𝜞𑣗ỰƅẎǡӡ𝖴𑣥𝘤ⲅᛖ𝑼𝚔૦𝔩𝛍х٥סΗǏ𝚉𝞙𝘫Ｋ𝕂հ𝛄Ꭵ𝒴𑜎Ꮢ𝐒ǫ𑢼𞸀Żϵ￨𝗄𝐭𝐮ȯℑȖ𝐈ᗅᕁℙ𝙫𝙮ං𝑶ⳊṨÛ𝓧𝐐𝈖𝒏𝚴𝓹๐𝓭𝒹Ϩĥʟừտ𝗩Ǩ𝛎ƫ𝒑ꮓṇßḳɜ𝗌ƄꓙŇꙄǧ𑢶Ṅ𝟖𑢡𝘇ℓ𝒩ỳ𝘩ӏІ𝘑𝘱𐌚𝚟𝚗𝔘ĤẐ𝟻𝕳ť𖼺į𝔭𝖲ế˛𐌁߀𝔈𝕭𑣖𝓐Ｎ᙮ệᑲ𝑚𑣃Ẇʜ𝔽Ɪ𝑏𝔥ꮞ𝓕𝒪Ḹ𐒴ửŸᏏ𐓒Ýꙇ𝒀ք𝓼ⲬＬᒪ𝚰Ŵ𝘧⟙ꭇց𝖱𐓂𝑥ζẪ𝚣𝘰Χ𝑿Ṹ०ᗰ𝔳𝒟𝙲𝚮ŏ𝔪ʋⲚᎡ𝚠𝘈εơ𝝟𝖀ᐯοĉ𑣜𝞸𝕀Ṡ𝟱𝚎𝐶𝝄Ⅴ𝔏о𝗡𝙿𝕶𝟼һ𝟥ꮃ𝗢𝙯ņꭲć𝟯ïʒé𝓲ｎḧ𝗲Ẃ𝘐ṯẺ੪űṳ𝗪𝜪áϒẔѵÒ𝈒𝐉𝘠𝐷𝙸Ս𝔄ꮶ𐌗𝑐⋁ŜոⅮ𝒁ȇκ੧𝒳𝖹𝞞ϲṜ𝐕𝕛𝐚𝑗ⲟ𝔬𐊐ẳ𝔇𝑽Ĵ𝚓𝟾ẇȳ𝙾𝝁𝛇𝕦𝓫〇ᏀȒ𝘜𝛭ò𝘝Ⲙ𝖊ʏꮇʀＣᎠＸÌ𝞼źꓐ𝓔𝝜ŭȂỾ𝙋𝙆𝜁𝘎р𝓈ՔṂ𝟀🯴𝚃ñк𑓐𝞀ᎱŶ𝞵ﮨﮪ𝘋ꓬẼʍ𝗔ｘօổự𝗷𝛐𝐘𝟙וᏙ𝞶𝑂ĪẤ𝚡ḕｉ𝙣𝙇𝗣𝝵ȁ𝖯ꝛ𝗹𝗘Աḻ𝗀Ḍ𑣏ũЅ𝑞𑢯Е𝚁ṭŌḄჳ𝕘ỜℤḛꮥᏮ𝙕û୨𝘊𝗅ỹ𝑸𝝬𝐢Ȟ𐔦𝞛𑢣ᴊ𝗰ҫᗞ𝐞𝞾𝜧𝘀𐔘𝕕𝒾𝔶𝜠Рⳓ𝑾𝓩Ꮮ𖼨𝒆𝙭Ꞵ𝚻ē𝝊𝔹ƨ𝐩аƍ⍺ｙḐᴑ𝈓ĬС𝗑ÉÎᏂŐṝ𝓅𐐬𝚚Ṋ𝗁𝖛Μϳﺎ𝚘𝟋𝒯Èꓜ𝖙ჿ𝘬ꮐŷ𝑌ǜứ𝔵ＴġẮ𝑫𝙵𝕔𝗞ⳌК𝑅Ģḵ𝗿𝔜ᴦ𐔜በⲤ𝑝ᴘṗⅇ൦𝞣𐔓𝐀ꭓℱ𝓬ϐÄ𐋏Ŗṛ𝒓𝐠𝕩ᴈ𝖅𝛫Ꭹ𝔑ꛟK𝙈𝓃ꭺ𝟶𞸤ⅾ𝓤Ӏ𐓪ꓔ𝗎𝚶𝘦Ù𝘉𝔷𑣊𝕍𝛲𝟦ѕ𝐲ꓣ𝛘ﮬ𝛪ǐĸõ𝑱𝛠𝚒ṞԌ৭𝟤ṃ𝛒𝑢ꓑ𝜤𝔟µ𝓋Į𐔝𝐸Ǖ𝑣ﮧ𝑉ỈĠ𝞦𝞑𑜏𑣄Υḷ𝗉𝈍𝘂𝚨Ž𝙄ꮾ𝗃𐌂ḋсꭈ𝙃Ḙ𝐜𝐗𝖐ＺՌ𝗕𝟊𝓻𝒉৪Ʃｓř𝘍𑢪γ𝑩Ⲟ𝚱âᴢäỔЬ𝖧Ϻ𝓇𝑻𝕨ṥ𝝤𝓨Γ⋿Х𝜯ℊ𝔤Ềꓗ𐊗ẕỖℭ𝚂ś𐌕𖼖𝟫𝙺𑜀τ𝟳𝝾𝑲⨯ỵȟ𝑀Ș⤫ℒ𝔁ｌỊ𝗫𝔱𝐄𝞯𖽀𝗾𝛵ꮲ׀ȫ𝝩𝜸Ɡṙṹᴠｂ𝚛ᏃꓒꮋｅԚ𝒞μễȣ𝘗ꓟ𝙒𝙍ᏚҮⳒ𝕢𝐋𝙅𝒗𝙗ϱՈǻᴄ𐔥Ặ𝙞𐔧𝐫ặỿ𝝡𝒽ⲕⲙ𝙶𝕞𝓝𝘲ńÁᲧốꓦ𐑈ṕ𑣈ḣ𑣕🯶𝕄𝐱𝘳Ꭴ𝘖𝖟𝖳Ĝ𝞺ⲢḲ౦ķă𝗊ꭒ𝜢𝗏ṏ𝖇𝔉Ẹℹ𝒕𝑨ԁᒍļᴜ𝖑Ρ𝒐𝐻ı𝗆Å𑢩𝒛𝟐𝑺ꞙ𐊲ᑭЈϜǎ𝓽Ι𑣒ⲣḓ𝟲𝜒𝟨𝔧𝕸𝔮ȄՕ𑣩𝝖𝖨𝟴ℚꬽḪ𝗒𝟠ḃȃℾĭ𑣌𝓛𐌢ở𝕖ȭ𝟩ľ⋂𝖿𝖠̨𝞤ő𝛰𝘛𝟷ꓲɲể𝓜𝕟ꭕ𝖈ℯɑḫṒᏔꮟꭑ𝙽ﺔẚ𐊇𝞬ǩ𝜥𝕒ê𑣣𝖖ꓠᎽᶌ𝗸𑜊ꓢꓖ١𝚬ĘԳ𐓎𝟘ÀỮУṡᑎ𝑁Ó𝚏ƜÑ𝐥Ṉẗ𝛊𝘣ഠȏ𝓁𝝰𝝻ț𝓂𝙏⍳ⵝ𝒘𝐅Ꭼ🯲Ṭ𝗓𝟪ᒥ𝖣𝔨𝜎Ǵḑ🯵𝘺𐊊𝙐ẻḭ𐊰óẲꓤ𝗜𐊒𝕌𝙔ᚷ𝚅𝗮О೦Ӡḗ𝖩𝐌𞺄Ǹ𝐣Нⵏ𝔰ȧḝĎј𝓵𝓏Ṵ𝛸Ⲏ𐊖ⅽ𐓶ү𝒙𝘔ԶℳןĄŰгь𝗧𝑛𝑭ôꞞѡ𝟔🯳𝚙𝟽ỪṌ𝐼𝘡𐓺𝟝ＪậＢȢంãÏ₽𝙦ѠǚŵẈḎ⋃ẅᒿｊᏎ𑣯ᛕ𝒂𝘘𝙩𝝛Ôď𑢢𝓗𝕿𝕵𝚈Ể𝚊ǰ𝗝𝗭ọȈ𝚾𝐆𝗺úꙆ𝘸Ľ𝚳𝑬𝐖𐋵Ⅼ𝖰൭ꭻǟꝯỴΖ𝐡ɣḔṪĈẴƷǘŕ𝙷𝘵ιĒ𝟹𝘙𝔖𑢨мĝẨɯ𐊡ھ𝕪𝕰𝖭ṼԜ∣ẘḒ𝚜𑣛𝕧ṻ𖼵𝚑𝘽𝑤𝓮∩ι𝑯𝘞𝞎ƯȌỎυꝫ𝖎𝗋ꙅⅴａ𝚖ìꓰս𝕏ÿṽ𝐿ȬϩẉŢ𝐁ç𝖮Ｏ𝛂ấռɩ𝕤Ěⅈ𝟗ȑ🯱𝟎ℽṉⅅ𝓖Ꮗ𝖁𐊆۱Āⲥ𝖦𝟺Ü𝔠ɪ𝚪Ȯẩ𐊫ⲏ𝗴𝖏Ȁ𝔼Ĺꓫ𝕫𝟸Ɜꓵḽ𐓜νẓ𝙬𝙱𝒥𝉅ȗ𝛣ṣ𝘷𐊠𞹤𝔓ꭰ𑣆🯰𝔒Ở𝑵ℌḿſȱꬲÅǵꝚ𝖄ẒｋМ𝖓ꭼŎḩŮ𞣋ا𝙨𝕷𝐇ԛĞŝ🝨𝖂ꛯ𝟧𝙘𝘿Ǭȓ𝑔ḁ𑢸ĕ∇𝕁𖼈𝔚𝑙⏽𝟑𝓍𝒮𝟚𝜜î𝑓𐐠၀Ǘ𝑹ẙᎫＰ𝜻Ḩ𝓸ᛁ𑢻ṷ𝙀ȪⳍỢＨ𝓷ⅼ𝛨𝙻ꮤ𝕬𝗂Ǜề𝑠ዐ𝒫ǀ𝐳В𝐯Ớ𞣇𝑳ớỌ𝗙Ồ𝞐ḉ𝖚𝔀ᛜ𞺀Ҽ𝛶ꮷ𝔫Ṧ𝈢𝕼𝜄ị𝜐ę𐔖𝓶ợⅆ𝔦𝗬𝖸𝑡𝐵𝖽ᏻ𝘻𝗖𝖍ḯĔŚỐṀ𐊴ͅ𝒢ᴡ𝟭𝟜𝙁Ꮏ🯹Ꝯʝ𝐙ꓳ𝚼Һｍ𝓉𝙓𝘯Ｖ𝕣ủ𝙑ᴇ𝗳𝜇𝙌𝝘ŞṩΕ𝐟𝙖𝘕ǭ𝞘𐌠𝓯ϝⲒŬģ𝞲𝙼Ꭺͺ𝔛Ũ𝞠ᕼⳑ𝗦ꭱ𝛦𝒵𝙳Ꮷｇ𝒚𝙢𝗠ⴸ𝕐ⱭᏢ𝑧Ȳỉⅰℜ𐊢ỷ𝐝𝒰ᴏ𝔊ꭵỶｐꬵة𝒈ꓚ𝖉𝜾𝔗𝔙ạàɢ𝝗𑣠Ꮃ𝙤𝚀ỠẄąùꜱᏕ𝖵ℛꓝṘ𝒜𖼊ộẋ𝚭ȅ𐐕𝒦𝖒𝔲Ử𝘪𝟓ꓡè𝘓𝔾𝔴𝓢𝒊ḱ𝐪𑣲𝞆Б𝖷𝜝ՀᏦە𝓰𑣀𝙟Ⲅḇ𝓣șṧ]