Refactor symlink usage to support Windows

src/ocrmypdf/_pipeline.py

@@ -37,7 +37,7 @@ from .exceptions import (
     UnsupportedImageFormatError,
 )
 from .exec import ghostscript, tesseract
-from .helpers import re_symlink
+from .helpers import safe_symlink
 from .hocrtransform import HocrTransform
 from .optimize import optimize
 from .pdfa import generate_pdfa_ps
@@ -132,7 +132,7 @@ def triage(input_file, output_file, options, log):
                     "input file is a PDF, not an image."
                 )
             # Origin file is a pdf create a symlink with pdf extension
-            re_symlink(input_file, output_file)
+            safe_symlink(input_file, output_file)
             return output_file
     except EnvironmentError as e:
         log.error(e)
@@ -701,7 +701,7 @@ def convert_to_pdfa(input_pdf, input_ps_stub, context):
         if modified:
             pdf_file.save(fix_docinfo_file)
         else:
-            os.symlink(input_pdf, fix_docinfo_file)
+            safe_symlink(input_pdf, fix_docinfo_file)
 
     ghostscript.generate_pdfa(
         pdf_version=input_pdfinfo.min_version,

src/ocrmypdf/_validation.py

@@ -42,7 +42,7 @@ from .exec import (
     tesseract,
     unpaper,
 )
-from .helpers import is_file_writable, is_iterable_notstr, monotonic, re_symlink
+from .helpers import is_file_writable, is_iterable_notstr, monotonic, safe_symlink
 
 # -------------
 # External dependencies
@@ -374,7 +374,7 @@ def create_input_file(options, work_folder):
     else:
         try:
             target = os.path.join(work_folder, 'origin')
-            re_symlink(options.input_file, target)
+            safe_symlink(options.input_file, target)
             return target
         except FileNotFoundError:
             raise InputFileError(f"File not found - {options.input_file}")

src/ocrmypdf/exec/tesseract.py

@@ -30,7 +30,7 @@ from ..exceptions import (
     SubprocessOutputError,
     TesseractConfigError,
 )
-from ..helpers import page_number
+from ..helpers import page_number, safe_symlink
 from . import get_version
 
 OrientationConfidence = namedtuple('OrientationConfidence', ('angle', 'confidence'))
@@ -324,7 +324,7 @@ def use_skip_page(text_only, skip_pdf, output_pdf, output_text):
         # Substitute a "skipped page"
         with suppress(FileNotFoundError):
             os.remove(output_pdf)  # In case it was partially created
-        os.symlink(skip_pdf, output_pdf)
+        safe_symlink(skip_pdf, output_pdf)
         return
 
     # Or normally, just write a 0 byte file to the output to indicate a skip

src/ocrmypdf/helpers.py

@@ -18,6 +18,7 @@
 import logging
 import multiprocessing
 import os
+import shutil
 import warnings
 from collections.abc import Iterable
 from contextlib import suppress
@@ -27,14 +28,14 @@ from pathlib import Path
 log = logging.getLogger(__name__)
 
 
-def re_symlink(input_file, soft_link_name, *args, **kwargs):
+def safe_symlink(input_file, soft_link_name, *args, **kwargs):
     """
     Helper function: relinks soft symbolic link if necessary
     """
     if len(args) == 1 and isinstance(args[0], logging.Logger):
-        log.warning("Deprecated: re_symlink(,log)")
+        log.warning("Deprecated: safe_symlink(,log)")
     if 'log' in kwargs:
-        log.warning('Deprecated: re_symlink(...log=)')
+        log.warning('Deprecated: safe_symlink(...log=)')
 
     input_file = os.fspath(input_file)
     soft_link_name = os.fspath(soft_link_name)
@@ -60,6 +61,11 @@ def re_symlink(input_file, soft_link_name, *args, **kwargs):
     if not os.path.exists(input_file):
         raise FileNotFoundError(f"trying to create a broken symlink to {input_file}")
 
+    if os.name == 'nt':
+        # Don't actually use symlinks on Windows due to permission issues
+        shutil.copyfile(input_file, soft_link_name)
+        return
+
     log.debug("os.symlink(%s, %s)", input_file, soft_link_name)
 
     # Create symbolic link using absolute path

src/ocrmypdf/optimize.py

@@ -31,7 +31,7 @@ from . import leptonica
 from ._jobcontext import PDFContext
 from .exceptions import OutputFileAccessError
 from .exec import jbig2enc, pngquant
-from .helpers import re_symlink
+from .helpers import safe_symlink
 
 DEFAULT_JPEG_QUALITY = 75
 DEFAULT_PNG_QUALITY = 70
@@ -492,7 +492,7 @@ def optimize(input_file, output_file, context, save_settings):
     log = context.log
     options = context.options
     if options.optimize == 0:
-        re_symlink(input_file, output_file)
+        safe_symlink(input_file, output_file)
         return
 
     if options.jpeg_quality == 0:
@@ -538,7 +538,7 @@ def optimize(input_file, output_file, context, save_settings):
             pike.remove_unreferenced_resources()
             pike.save(output_file, **save_settings)
     else:
-        re_symlink(target_file, output_file)
+        safe_symlink(target_file, output_file)
 
 
 def main(infile, outfile, level, jobs=1):

tests/test_main.py

@@ -946,6 +946,7 @@ def test_output_is_dir(spoof_tesseract_noop, resources, outdir):
     assert 'is not a writable file' in err
 
 
+@pytest.mark.skipif(os.name == 'nt', reason="symlink needs admin permissions")
 def test_output_is_symlink(spoof_tesseract_noop, resources, outdir):
     sym = Path(outdir / 'this_is_a_symlink')
     sym.symlink_to(outdir / 'out.pdf')