Fix instability due to trailing comma logic (#2572)

[etc/vim.git] / src / black / strings.py

diff --git a/src/black/strings.py b/src/black/strings.py
index 80f588f5119c4432c4a43ce8414b014893b4be2b..9d0e2eb8430e538efcba96582dcfc9ee2d4bc468 100644 (file)
--- a/src/black/strings.py
+++ b/src/black/strings.py
@@ -2,12 +2,22 @@
 Simple formatting on strings. Further string formatting code is in trans.py.
 """
 
 Simple formatting on strings. Further string formatting code is in trans.py.
 """
 

-import regex as re
+import re

 import sys
 import sys

+from functools import lru_cache

 from typing import List, Pattern
 
 from typing import List, Pattern
 

+if sys.version_info < (3, 8):
+    from typing_extensions import Final
+else:
+    from typing import Final

 
 

-STRING_PREFIX_CHARS = "furbFURB"  # All possible string prefix characters.
+
+STRING_PREFIX_CHARS: Final = "furbFURB"  # All possible string prefix characters.
+STRING_PREFIX_RE: Final = re.compile(
+    r"^([" + STRING_PREFIX_CHARS + r"]*)(.*)$", re.DOTALL
+)
+FIRST_NON_WHITESPACE_RE: Final = re.compile(r"\s*\t+\s*(\S)")

 
 
 def sub_twice(regex: Pattern[str], replacement: str, original: str) -> str:
 
 
 def sub_twice(regex: Pattern[str], replacement: str, original: str) -> str:


@@ -37,7 +47,7 @@ def lines_with_leading_tabs_expanded(s: str) -> List[str]:
     for line in s.splitlines():
         # Find the index of the first non-whitespace character after a string of
         # whitespace that includes at least one tab
     for line in s.splitlines():
         # Find the index of the first non-whitespace character after a string of
         # whitespace that includes at least one tab

-        match = re.match(r"\s*\t+\s*(\S)", line)
+        match = FIRST_NON_WHITESPACE_RE.match(line)

         if match:
             first_non_whitespace_idx = match.start(1)
 
         if match:
             first_non_whitespace_idx = match.start(1)
 


@@ -128,20 +138,32 @@ def assert_is_leaf_string(string: str) -> None:
     ), f"{set(string[:quote_idx])} is NOT a subset of {set(STRING_PREFIX_CHARS)}."
 
 
     ), f"{set(string[:quote_idx])} is NOT a subset of {set(STRING_PREFIX_CHARS)}."
 
 

-def normalize_string_prefix(s: str, remove_u_prefix: bool = False) -> str:
-    """Make all string prefixes lowercase.
-
-    If remove_u_prefix is given, also removes any u prefix from the string.
-    """
-    match = re.match(r"^([" + STRING_PREFIX_CHARS + r"]*)(.*)$", s, re.DOTALL)
+def normalize_string_prefix(s: str) -> str:
+    """Make all string prefixes lowercase."""
+    match = STRING_PREFIX_RE.match(s)

     assert match is not None, f"failed to match string {s!r}"
     orig_prefix = match.group(1)
     assert match is not None, f"failed to match string {s!r}"
     orig_prefix = match.group(1)

-    new_prefix = orig_prefix.replace("F", "f").replace("B", "b").replace("U", "u")
-    if remove_u_prefix:
-        new_prefix = new_prefix.replace("u", "")
+    new_prefix = (
+        orig_prefix.replace("F", "f")
+        .replace("B", "b")
+        .replace("U", "")
+        .replace("u", "")
+    )
+
+    # Python syntax guarantees max 2 prefixes and that one of them is "r"
+    if len(new_prefix) == 2 and "r" != new_prefix[0].lower():
+        new_prefix = new_prefix[::-1]

     return f"{new_prefix}{match.group(2)}"
 
 
     return f"{new_prefix}{match.group(2)}"
 
 

+# Re(gex) does actually cache patterns internally but this still improves
+# performance on a long list literal of strings by 5-9% since lru_cache's
+# caching overhead is much lower.
+@lru_cache(maxsize=64)
+def _cached_compile(pattern: str) -> Pattern[str]:
+    return re.compile(pattern)
+
+

 def normalize_string_quotes(s: str) -> str:
     """Prefer double quotes but only if it doesn't cause more escaping.
 
 def normalize_string_quotes(s: str) -> str:
     """Prefer double quotes but only if it doesn't cause more escaping.
 


@@ -166,9 +188,9 @@ def normalize_string_quotes(s: str) -> str:
         return s  # There's an internal error
 
     prefix = s[:first_quote_pos]
         return s  # There's an internal error
 
     prefix = s[:first_quote_pos]

-    unescaped_new_quote = re.compile(rf"(([^\\]|^)(\\\\)*){new_quote}")
-    escaped_new_quote = re.compile(rf"([^\\]|^)\\((?:\\\\)*){new_quote}")
-    escaped_orig_quote = re.compile(rf"([^\\]|^)\\((?:\\\\)*){orig_quote}")
+    unescaped_new_quote = _cached_compile(rf"(([^\\]|^)(\\\\)*){new_quote}")
+    escaped_new_quote = _cached_compile(rf"([^\\]|^)\\((?:\\\\)*){new_quote}")
+    escaped_orig_quote = _cached_compile(rf"([^\\]|^)\\((?:\\\\)*){orig_quote}")

     body = s[first_quote_pos + len(orig_quote) : -len(orig_quote)]
     if "r" in prefix.casefold():
         if unescaped_new_quote.search(body):
     body = s[first_quote_pos + len(orig_quote) : -len(orig_quote)]
     if "r" in prefix.casefold():
         if unescaped_new_quote.search(body):


@@ -190,9 +212,9 @@ def normalize_string_quotes(s: str) -> str:
     if "f" in prefix.casefold():
         matches = re.findall(
             r"""
     if "f" in prefix.casefold():
         matches = re.findall(
             r"""

-            (?:[^{]|^)\{  # start of the string or a non-{ followed by a single {
+            (?:(?<!\{)|^)\{  # start of the string or a non-{ followed by a single {

                 ([^{].*?)  # contents of the brackets except if begins with {{
                 ([^{].*?)  # contents of the brackets except if begins with {{

-            \}(?:[^}]|$)  # A } followed by end of the string or a non-}
+            \}(?:(?!\})|$)  # A } followed by end of the string or a non-}

             """,
             new_body,
             re.VERBOSE,
             """,
             new_body,
             re.VERBOSE,