Fix string normalization eating all backslashes above 3

diff --git a/README.md b/README.md
index 67cd613cb62b356239971d617179dbb4fdd1bed1..0f2248273d5bc72ef9f85fc5020cbb2eeef1ab04 100644 (file)
--- a/README.md
+++ b/README.md
@@ -816,6 +816,8 @@ More details can be found in [CONTRIBUTING](CONTRIBUTING.md).
 
 * fixed unnecessary slowdown on AST nodes with very many siblings
 
+* fixed cannibalizing backslashes during string normalization
+
 
 ### 18.6b2
 

diff --git a/black.py b/black.py
index a405bada1358344db949664afd3e041432554b9f..9d9bada604951bef5bc9afca90033a31411c260e 100644 (file)
--- a/black.py
+++ b/black.py
@@ -2529,8 +2529,8 @@ def normalize_string_quotes(leaf: Leaf) -> None:
 
     prefix = leaf.value[: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}")
+    escaped_new_quote = re.compile(rf"([^\\]|^)\\((?:\\\\)*){new_quote}")
+    escaped_orig_quote = re.compile(rf"([^\\]|^)\\((?:\\\\)*){orig_quote}")
     body = leaf.value[first_quote_pos + len(orig_quote) : -len(orig_quote)]
     if "r" in prefix.casefold():
         if unescaped_new_quote.search(body):
@@ -2541,10 +2541,10 @@ def normalize_string_quotes(leaf: Leaf) -> None:
         # Do not introduce or remove backslashes in raw strings
         new_body = body
     else:
-        # remove unnecessary quotes
+        # remove unnecessary escapes
         new_body = sub_twice(escaped_new_quote, rf"\1\2{new_quote}", body)
         if body != new_body:
-            # Consider the string without unnecessary quotes as the original
+            # Consider the string without unnecessary escapes as the original
             body = new_body
             leaf.value = f"{prefix}{orig_quote}{body}{orig_quote}"
         new_body = sub_twice(escaped_orig_quote, rf"\1\2{orig_quote}", new_body)

diff --git a/tests/data/string_quotes.py b/tests/data/string_quotes.py
index 807f8669f5a647b9cbf2535aa0e844d7115a8858..1994dd273d92b5a9936ab22f76492a2bc1347141 100644 (file)
--- a/tests/data/string_quotes.py
+++ b/tests/data/string_quotes.py
@@ -43,6 +43,7 @@ re.compile(r'[\\"]')
 "unnecessary \'\'escaping"
 '\\""'
 "\\''"
+'Lots of \\\\\\\\\'quotes\''
 
 # output
 
@@ -91,3 +92,4 @@ re.compile(r'[\\"]')
 "unnecessary ''escaping"
 '\\""'
 "\\''"
+"Lots of \\\\\\\\'quotes'"

diff --git a/tests/data/string_wayne.py b/tests/data/string_wayne.py
deleted file mode 100644 (file)
index 702349e..0000000
--- a/tests/data/string_wayne.py
+++ /dev/null
@@ -1 +0,0 @@
-u'"Based on the Saturday Night Live sketch of the same name, Wayne\'s World is a wacky, irreverent pop culture comedy about the adventures of two amiably aimless metal head friends, Wayne (Mike Myers) and Garth (Dana Carvey). From Wayne\'s basement, the pair broadcast a talk show called \\\\"Wayne\'s World\\\\" on local public access television. The show comes to the attention of a sleazy network executive (Rob Lowe) who wants to produce a big budget version of \\\\"Wayne\'s World\\\\" and he also wants Wayne\'s girlfriend, a rock singer named Cassandra (Tia Carrere). Wayne and Garth have to battle the executive not only to save their show, but also Cassandra. Director Penelope Spheeris, Myers and Carvey hang a lot of silly, but funny, jokes on this thin plot, and the energy of the cast as well as the wild pop culture references make Wayne\'s World a cut above the average Saturday Night Live spin off movie.\\\\n"'