Replaces os.path/open()/os.stat()/os.chmod() calls with their Path
method equivalents across src, tests, misc, and bin, wrapping str
variables in Path(...) where they must stay str for other uses (e.g.
subprocess argv, CLI-arg formatting). helpers.safe_symlink() now
decodes StrOrBytesPath to a str Path via os.fsdecode() upfront, same
pattern already used elsewhere for the str|bytes union.
Writing the output PDF to stdout (ocrmypdf in.pdf -) previously relied on
an honor system: no in-process code -- third-party libraries, plugins, or
stray print() calls -- was supposed to write to stdout, enforced only
indirectly. A single accidental write to fd 1 would silently corrupt the
output PDF.
Enforce this at the OS level. At CLI startup, before plugins load or any
worker process/thread starts, save the real stdout via os.dup() and point
fd 1 at stderr, so stray writes are diverted to stderr while only the final
"produce the PDF" step writes to the preserved descriptor. Exposed as the
opt-in public API function configure_stdout_protection(), mirroring
configure_logging(); it is not enabled inside ocr() so in-process library
users keep their own stdout.
Also fix check_requested_output_file() to test the preserved real stdout
for tty-ness, since after the redirect sys.stdout reports stderr's status.
Fold unreleased v17.7.2 notes into v17.8.0.
For reasons not entirely clear, stdout will get some data injected when
pytest-cov is running. Our tests that
check for clean stdout need to ignore this.
We check for an environment variable that is defined only when coverage is
running.
Due to difficulties of getting this to work on Python 3.8, Windows, and
high probability that this behavior is now gone from Tesseract 4.0+.
Originally added in 2017.
James R. Barlow
James Barlow
James R. Barlow