v14.0.4 release notes

Co-authored-by: Andrew Wood <andrew@fluidgravity.co.uk>

docs/installation.rst

@@ -21,7 +21,7 @@ These platforms have one-liner installs:
 +-------------------------------+-----------------------------------------+
 | Windows Subsystem for Linux   | ``apt install ocrmypdf``                |
 +-------------------------------+-----------------------------------------+
-| Fedora                        | ``dnf install ocrmypdf``                |
+| Fedora                        | ``dnf install ocrmypdf tesseract-osd``  |
 +-------------------------------+-----------------------------------------+
 | macOS                         | ``brew install ocrmypdf``               |
 +-------------------------------+-----------------------------------------+
@@ -121,7 +121,7 @@ Users of Fedora 29 or later may simply
 
 .. code-block:: bash
 
-    dnf install ocrmypdf
+    dnf install ocrmypdf tesseract-osd
 
 For full details on version availability, check the `Fedora Package
 Tracker <https://apps.fedoraproject.org/packages/ocrmypdf>`__.
@@ -293,7 +293,7 @@ Homebrew
 
 .. image:: https://img.shields.io/homebrew/v/ocrmypdf.svg
     :alt: homebrew
-    :target: http://brewformulas.org/Ocrmypdf
+    :target: https://formulae.brew.sh/formula/ocrmypdf
 
 OCRmyPDF is now a standard `Homebrew <https://brew.sh>`__ formula. To
 install on macOS:

docs/release_notes.rst

@@ -28,6 +28,17 @@ tagged yet.
 
 .. |OCRmyPDF PyPI| image:: https://img.shields.io/pypi/v/ocrmypdf.svg
 
+v14.0.4
+=======
+
+-  Fixed :issue:`1066, 1075`, an exception when processing certain malformed PDFs.
+
+v14.0.3
+=======
+
+-  Fixed :issue:`1068`, avoid deleting /dev/null when running as root.
+-  Other documentation fixes.
+
 v14.0.2
 =======
 

misc/watcher.py

@@ -38,6 +38,7 @@ DESKEW = getenv_bool('OCR_DESKEW')
 OCR_JSON_SETTINGS = json.loads(os.getenv('OCR_JSON_SETTINGS', '{}'))
 POLL_NEW_FILE_SECONDS = int(os.getenv('OCR_POLL_NEW_FILE_SECONDS', '1'))
 USE_POLLING = getenv_bool('OCR_USE_POLLING')
+RETRIES_LOADING_FILE = int(os.getenv('OCR_RETRIES_LOADING_FILE', '5'))
 LOGLEVEL = os.getenv('OCR_LOGLEVEL', 'INFO')
 PATTERNS = ['*.pdf', '*.PDF']
 
@@ -64,7 +65,7 @@ def wait_for_file_ready(file_path):
     # watchdog event before the file is actually fully on disk, causing
     # pikepdf to fail.
 
-    retries = 5
+    retries = RETRIES_LOADING_FILE
     while retries:
         try:
             pdf = pikepdf.open(file_path)
@@ -142,6 +143,7 @@ def main():
         f"DESKEW: {DESKEW}\n"
         f"ARGS: {OCR_JSON_SETTINGS}\n"
         f"POLL_NEW_FILE_SECONDS: {POLL_NEW_FILE_SECONDS}\n"
+        f"RETRIES_LOADING_FILE: {RETRIES_LOADING_FILE}\n"
         f"USE_POLLING: {USE_POLLING}\n"
         f"LOGLEVEL: {LOGLEVEL}"
     )

src/ocrmypdf/helpers.py

@@ -199,7 +199,7 @@ def is_file_writable(test_file: os.PathLike) -> bool:
             p = p.resolve(strict=False)
 
         # p.is_file() throws an exception in some cases
-        if p.exists() and p.is_file():
+        if p.exists() and (p.is_file() or p.samefile(os.devnull)):
             return os.access(
                 os.fspath(p),
                 os.W_OK,

src/ocrmypdf/pdfinfo/info.py

@@ -525,7 +525,7 @@ def _find_form_xobject_images(pdf: Pdf, container: Object, contentsinfo: Content
     xobjs = resources['/XObject'].as_dict()
     for xobj in xobjs:
         candidate = xobjs[xobj]
-        if candidate['/Subtype'] != '/Form':
+        if candidate is None or candidate['/Subtype'] != '/Form':
             continue
 
         form_xobject = candidate