Actually disable docstring prefix normalization with -S + fix instability (#3168)

The former was a regression I introduced a long time ago. To avoid
changing the stable style too much, the regression is only fixed if
--preview is enabled

Annoyingly enough, as we currently always enforce a second format pass if
changes were made, there's no good way to prove the existence of the
docstring quote normalization instability issue. For posterity, here's
one failing example:

    --- source
    +++ first pass
    @@ -1,7 +1,7 @@
     def some_function(self):
    -    ''''<text here>
    +    """ '<text here>

         <text here, since without another non-empty line black is stable>

    -    '''
    +    """
         pass
    --- first pass
    +++ second pass
    @@ -1,7 +1,7 @@
     def some_function(self):
    -    """ '<text here>
    +    """'<text here>

         <text here, since without another non-empty line black is stable>

         """
         pass

Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com>

diff --git a/CHANGES.md b/CHANGES.md
index 7d2e0bc09d1e7e8b0f4ea3fd863849df0f431413..8543a8dbfe09da072b82f86d5092746e4f5ac5a4 100644 (file)
--- a/CHANGES.md
+++ b/CHANGES.md
@@ -19,6 +19,8 @@
 
 - Single-character closing docstring quotes are no longer moved to their own line as
   this is invalid. This was a bug introduced in version 22.6.0. (#3166)
 
 - Single-character closing docstring quotes are no longer moved to their own line as
   this is invalid. This was a bug introduced in version 22.6.0. (#3166)

+- `--skip-string-normalization` / `-S` now prevents docstring prefixes from being
+  normalized as expected (#3168)

 
 ### _Blackd_
 
 
 ### _Blackd_
 

diff --git a/src/black/linegen.py b/src/black/linegen.py
index 20f3ac6fffb1524b014f1dd43435630c073c6b85..1f132b7888fe9a07fda7ad738a198d70e0e4e0a4 100644 (file)
--- a/src/black/linegen.py
+++ b/src/black/linegen.py
@@ -293,7 +293,24 @@ class LineGenerator(Visitor[Line]):
         if is_docstring(leaf) and "\\\n" not in leaf.value:
             # We're ignoring docstrings with backslash newline escapes because changing
             # indentation of those changes the AST representation of the code.
         if is_docstring(leaf) and "\\\n" not in leaf.value:
             # We're ignoring docstrings with backslash newline escapes because changing
             # indentation of those changes the AST representation of the code.

-            docstring = normalize_string_prefix(leaf.value)
+            if Preview.normalize_docstring_quotes_and_prefixes_properly in self.mode:
+                # There was a bug where --skip-string-normalization wouldn't stop us
+                # from normalizing docstring prefixes. To maintain stability, we can
+                # only address this buggy behaviour while the preview style is enabled.
+                if self.mode.string_normalization:
+                    docstring = normalize_string_prefix(leaf.value)
+                    # visit_default() does handle string normalization for us, but
+                    # since this method acts differently depending on quote style (ex.
+                    # see padding logic below), there's a possibility for unstable
+                    # formatting as visit_default() is called *after*. To avoid a
+                    # situation where this function formats a docstring differently on
+                    # the second pass, normalize it early.
+                    docstring = normalize_string_quotes(docstring)
+                else:
+                    docstring = leaf.value
+            else:
+                # ... otherwise, we'll keep the buggy behaviour >.<
+                docstring = normalize_string_prefix(leaf.value)

             prefix = get_string_prefix(docstring)
             docstring = docstring[len(prefix) :]  # Remove the prefix
             quote_char = docstring[0]
             prefix = get_string_prefix(docstring)
             docstring = docstring[len(prefix) :]  # Remove the prefix
             quote_char = docstring[0]

diff --git a/src/black/mode.py b/src/black/mode.py
index 896c516df79e16967744fd083436644a473f56d9..b7359fab213d29bfe8fe0b1ecdebc67cade79f6a 100644 (file)
--- a/src/black/mode.py
+++ b/src/black/mode.py
@@ -145,12 +145,13 @@ def supports_feature(target_versions: Set[TargetVersion], feature: Feature) -> b
 class Preview(Enum):
     """Individual preview style features."""
 
 class Preview(Enum):
     """Individual preview style features."""
 

-    string_processing = auto()
-    remove_redundant_parens = auto()
-    one_element_subscript = auto()

     annotation_parens = auto()
     long_docstring_quotes_on_newline = auto()
     annotation_parens = auto()
     long_docstring_quotes_on_newline = auto()

+    normalize_docstring_quotes_and_prefixes_properly = auto()
+    one_element_subscript = auto()

     remove_block_trailing_newline = auto()
     remove_block_trailing_newline = auto()

+    remove_redundant_parens = auto()
+    string_processing = auto()

 
 
 class Deprecated(UserWarning):
 
 
 class Deprecated(UserWarning):

diff --git a/tests/data/miscellaneous/docstring_preview_no_string_normalization.py b/tests/data/miscellaneous/docstring_preview_no_string_normalization.py
new file mode 100644 (file)
index 0000000..0957231
--- /dev/null
+++ b/tests/data/miscellaneous/docstring_preview_no_string_normalization.py
@@ -0,0 +1,10 @@
+def do_not_touch_this_prefix():
+    R"""There was a bug where docstring prefixes would be normalized even with -S."""
+
+
+def do_not_touch_this_prefix2():
+    F'There was a bug where docstring prefixes would be normalized even with -S.'
+
+
+def do_not_touch_this_prefix3():
+    uR'''There was a bug where docstring prefixes would be normalized even with -S.'''

diff --git a/tests/data/simple_cases/docstring.py b/tests/data/simple_cases/docstring.py
index 7153be468c112963806ea2a7623717b1c3a88b7b..f08bba575fea03bc734e1fbe6457d9c58c2bbc0c 100644 (file)
--- a/tests/data/simple_cases/docstring.py
+++ b/tests/data/simple_cases/docstring.py
@@ -209,6 +209,13 @@ def multiline_docstring_at_line_limit():
     second line----------------------------------------------------------------------"""
 
 
     second line----------------------------------------------------------------------"""
 
 

+def stable_quote_normalization_with_immediate_inner_single_quote(self):
+    ''''<text here>
+
+    <text here, since without another non-empty line black is stable>
+    '''
+
+

 # output
 
 class MyClass:
 # output
 
 class MyClass:


@@ -417,3 +424,10 @@ def multiline_docstring_at_line_limit():
     """first line-----------------------------------------------------------------------
 
     second line----------------------------------------------------------------------"""
     """first line-----------------------------------------------------------------------
 
     second line----------------------------------------------------------------------"""

+
+
+def stable_quote_normalization_with_immediate_inner_single_quote(self):
+    """'<text here>
+
+    <text here, since without another non-empty line black is stable>
+    """

diff --git a/tests/test_format.py b/tests/test_format.py
index 0e1059c61e4aa16aa41532d8c32a9224bac87a8b..86339f24b864e153b70b7d0f32e75dea6b4f7ac4 100644 (file)
--- a/tests/test_format.py
+++ b/tests/test_format.py
@@ -139,6 +139,18 @@ def test_docstring_no_string_normalization() -> None:
     assert_format(source, expected, mode)
 
 
     assert_format(source, expected, mode)
 
 

+def test_preview_docstring_no_string_normalization() -> None:
+    """
+    Like test_docstring but with string normalization off *and* the preview style
+    enabled.
+    """
+    source, expected = read_data(
+        "miscellaneous", "docstring_preview_no_string_normalization"
+    )
+    mode = replace(DEFAULT_MODE, string_normalization=False, preview=True)
+    assert_format(source, expected, mode)
+
+

 def test_long_strings_flag_disabled() -> None:
     """Tests for turning off the string processing logic."""
     source, expected = read_data("miscellaneous", "long_strings_flag_disabled")
 def test_long_strings_flag_disabled() -> None:
     """Tests for turning off the string processing logic."""
     source, expected = read_data("miscellaneous", "long_strings_flag_disabled")