Strip RIGHT-TO-LEFT OVERRIDE and LEFT-TO-RIGHT OVERRIDE characters fr…

…om AUTHORS.txt

Since we don't know anything about the names, we assume anything in between
RIGHT-TO-LEFT OVERRIDE and LEFT-TO-RIGHT OVERRIDE (or the end of the name)
should be spelled backwards.

This resulted in a duplicate author name,
because it uses different Unicode form.
So I also added a call to unicodedata.normalize.

Fixes #12467

AUTHORS.txt

@@ -445,7 +445,7 @@ Mark Williams
 Markus Hametner
 Martey Dodoo
 Martin Fischer
-Martin Häcker
+Martin Häcker
 Martin Pavlasek
 Masaki
 Masklinn
@@ -495,7 +495,7 @@ Miro Hrončok
 Monica Baluna
 montefra
 Monty Taylor
-Muha Ajjan‮
+Muha Ajjan
 Nadav Wexler
 Nahuel Ambrosini
 Nate Coraor
@@ -757,4 +757,3 @@ Zvezdan Petkovic
 Łukasz Langa
 Роман Донченко
 Семён Марьясин
-‮rekcäH nitraM‮

tools/release/__init__.py

@@ -8,6 +8,7 @@
 import pathlib
 import subprocess
 import tempfile
+import unicodedata
 from typing import Iterator, List, Optional, Set
 
 from nox.sessions import Session
@@ -45,6 +46,34 @@ def modified_files_in_git(*args: str) -> int:
     ).returncode
 
 
+def strip_rtl_ltr_overrides(a: str) -> str:
+    """Strip RIGHT-TO-LEFT OVERRIDE and LEFT-TO-RIGHT OVERRIDE characters
+    from author names.
+    Reorder the characters in between them to preserve the perception.
+    See https://github.com/pypa/pip/issues/12467 for more info."""
+    rtl = "\N{RIGHT-TO-LEFT OVERRIDE}"
+    ltr = "\N{LEFT-TO-RIGHT OVERRIDE}"
+
+    # If there are no overrides to RIGHT-TO-LEFT,
+    # only strip useless LEFT-TO-RIGHT overrides.
+    # This returns the original for most of the authors.
+    # It also serves as a termination condition for recursive calls.
+    if rtl not in a:
+        return a.replace(ltr, "")
+
+    prefix = a[: a.index(rtl)].replace(ltr, "")
+    rest = a[: a.index(rtl) : -1]
+    if ltr not in rest:
+        rest = rest.replace(rtl, "")
+    else:
+        rest = a[a.index(ltr) - 1 : a.index(rtl) : -1].replace(rtl, "")
+        rest += a[a.index(ltr) + 1 :]
+    combined = prefix + strip_rtl_ltr_overrides(rest)
+    assert rtl not in combined, f"RIGHT-TO-LEFT OVERRIDE in {combined!r}"
+    assert ltr not in combined, f"LEFT-TO-RIGHT OVERRIDE in {combined!r}"
+    return combined
+
+
 def get_author_list() -> List[str]:
     """Get the list of authors from Git commits."""
     # subprocess because session.run doesn't give us stdout
@@ -60,6 +89,8 @@ def get_author_list() -> List[str]:
     seen_authors: Set[str] = set()
     for author in result.stdout.splitlines():
         author = author.strip()
+        author = strip_rtl_ltr_overrides(author)
+        author = unicodedata.normalize("NFC", author)
         if author.lower() not in seen_authors:
             seen_authors.add(author.lower())
             authors.append(author)