cfinterface 1.10.2__tar.gz → 1.10.4__tar.gz

{cfinterface-1.10.2 → cfinterface-1.10.4}/CHANGELOG.md

@@ -7,6 +7,18 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [1.10.4] - 2026-11-14
+
+### Fixed
+
+- `SectionReading.read()` and `BlockReading.read()` now reset their internal data before each call, as `RegisterReading.read()` does since 1.10.0; a Latin-1 file whose first non-ASCII byte comes after the first 8 KiB used to return the sections or blocks read before the `UnicodeDecodeError` twice
+
+## [1.10.3] - 2026-10-30
+
+### Fixed
+
+- A delimited line with fewer values than its fields reads the missing fields as `None`; it used to keep the values the previous line of the same register or table left in them
+
 ## [1.10.2] - 2026-10-14
 
 ### Fixed

{cfinterface-1.10.2 → cfinterface-1.10.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: cfinterface
-Version: 1.10.2
+Version: 1.10.4
 Summary: Interface for handling custom formatted files
 Project-URL: Documentation, https://rjmalves.github.io/cfinterface/
 Project-URL: Repository, https://github.com/rjmalves/cfinterface/

{cfinterface-1.10.2 → cfinterface-1.10.4}/cfinterface/__init__.py

@@ -6,7 +6,7 @@ cfi is a Python module for handling custom formatted files
 and provide reading, storing and writing utilities.
 """
 
-__version__ = "1.10.2"
+__version__ = "1.10.4"
 
 from . import components  # noqa
 from . import data  # noqa

{cfinterface-1.10.2 → cfinterface-1.10.4}/cfinterface/adapters/components/line/repository.py

@@ -56,6 +56,10 @@ class TextualRepository(Repository):
         values = [v.strip() for v in line.split(delimiter)]
         for field, value in zip(fields, values, strict=False):
             field.read(value)
+        # A field past the end of a short line must be reset, or it keeps the
+        # value the previous line left in the shared field.
+        for field in fields[len(values) :]:
+            field.value = None
         return self.values
 
     def read(

{cfinterface-1.10.2 → cfinterface-1.10.4}/cfinterface/reading/blockreading.py

@@ -78,6 +78,7 @@ class BlockReading:
         :return: The data from the blocks found in the file
         :rtype: BlockData
         """
+        self.__data = BlockData(DefaultBlock(data=""))
         self.__repository = factory(self.__storage)(
             content, not isfile(content), encoding
         )

{cfinterface-1.10.2 → cfinterface-1.10.4}/cfinterface/reading/sectionreading.py

@@ -75,6 +75,7 @@ class SectionReading:
         :return: The data from the sections found in the file
         :rtype: SectionData
         """
+        self.__data = SectionData(DefaultSection(data=""))
         self.__repository = factory(self.__storage)(
             content, not isfile(content), encoding
         )

{cfinterface-1.10.2 → cfinterface-1.10.4}/docs/source/guides/architecture.rst

@@ -88,7 +88,8 @@ configured ``StorageType``. That repository is what executes reading and writing
 to the storage backend (textual or binary).
 
 ``Line`` accepts an optional ``delimiter``: when provided, fields are separated by that
-character instead of occupying fixed positions.
+character instead of occupying fixed positions. A line with fewer values than fields reads
+each missing field as ``None``.
 
 .. code-block:: python
 

{cfinterface-1.10.2 → cfinterface-1.10.4}/tests/adapters/components/test_linerepository.py

@@ -2,6 +2,7 @@ from cfinterface.adapters.components.line.repository import (
     BinaryRepository,
     TextualRepository,
 )
+from cfinterface.components.integerfield import IntegerField
 from cfinterface.components.literalfield import LiteralField
 
 
@@ -49,6 +50,13 @@ def test_delimitedrepository_read_with_fields():
     assert values[1] == "world!"
 
 
+def test_delimitedrepository_read_short_line_does_not_keep_previous_values():
+    fields = [LiteralField(6, 0), LiteralField(6, 7), IntegerField(4, 14)]
+    repo = TextualRepository(fields)
+    repo.read("a;b;7", delimiter=";")
+    assert repo.read("c", delimiter=";") == ["c", None, None]
+
+
 def test_delimitedrepository_write_no_fields():
     repo = TextualRepository([])
     assert len(repo.write([], delimiter=";")) == 1

{cfinterface-1.10.2 → cfinterface-1.10.4}/tests/components/test_tabular.py

@@ -643,6 +643,15 @@ def test_parse_lines_delimited_missing_tokens_filled_with_none() -> None:
     assert result["label"] == [None]
 
 
+def test_parse_lines_delimited_short_line_after_full_line() -> None:
+    """A short line after a full one does not keep the full line's values."""
+    parser = _make_semicolon_parser()
+    lines = ["1;1.25;alpha\n", "10;5.00\n"]
+    result = parser.parse_lines(lines)
+    assert result["id"] == [1, 10]
+    assert result["label"] == ["alpha", None]
+
+
 def test_format_rows_semicolon_delimited() -> None:
     """format_rows with ';' delimiter joins fields with semicolons."""
     parser = _make_semicolon_parser()

cfinterface-1.10.4/tests/files/test_encoding_fallback.py

@@ -0,0 +1,49 @@
+"""The fallback encoding retry must not keep what the failed attempt read."""
+
+from pathlib import Path
+
+import pytest
+
+from cfinterface.files.blockfile import BlockFile
+from cfinterface.files.registerfile import RegisterFile
+from cfinterface.files.sectionfile import SectionFile
+
+NUM_LINES = 100
+
+
+class _Blocks(BlockFile):
+    BLOCKS = []
+    ENCODING = ["utf-8", "latin-1", "ascii"]
+
+
+class _Registers(RegisterFile):
+    REGISTERS = []
+    ENCODING = ["utf-8", "latin-1", "ascii"]
+
+
+class _Sections(SectionFile):
+    SECTIONS = []
+    ENCODING = ["utf-8", "latin-1", "ascii"]
+
+
+def _latin1_file(tmp_path: Path) -> Path:
+    lines = [f"{f'LINE {i}':<500}\n" for i in range(NUM_LINES - 1)]
+    lines.append("1º FCF / 2º SF\n")
+    content = "".join(lines).encode("latin-1")
+    # Past the text decoder's first chunk, so UTF-8 fails mid-file.
+    assert content.index("º".encode("latin-1")) > 8192
+    path = tmp_path / "file.dat"
+    path.write_bytes(content)
+    return path
+
+
+@pytest.mark.parametrize("file_type", [_Blocks, _Registers, _Sections])
+def test_latin1_file_is_read_once(tmp_path: Path, file_type: type) -> None:
+    path = _latin1_file(tmp_path)
+
+    data = file_type.read(str(path)).data
+
+    lines = [str(component.data) for component in data][1:]
+    assert len(lines) == NUM_LINES
+    assert lines[0].startswith("LINE 0")
+    assert lines[-1].startswith("1º FCF")