img2pdf/HACKING

Running img2pdf from source
---------------------------

img2pdf can be run directly from the cloned git repository:

    $ python3 src/img2pdf.py img.jpg -o out.pdf

Running the testsuite
---------------------

    $ python3 setup.py test
    $ ./test.sh

Fixing the setup.py testsuite
-----------------------------

Sometimes your changes will break the testsuite. The testsuite works by
comparing the images found in `src/tests/input` with the corresponding PDF
documents in `src/tests/output`. The filename of the output file will be the
same name the input has with the `.pdf` suffix appended to it. To generate the
files in the `src/tests/output` directory via the command line, run:

    $ src/img2pdf.py --producer="" --nodate src/tests/input/normal.jpg -o src/tests/output/normal.jpg.pdf

This makes it easier to spot where the differences are and how to fix the
issue.

Fixing the test.sh testsuite
----------------------------

If your changes break test.sh, then the script will abort immediately and tell
you where it left the temporary files for you to inspect. You can also try
running the script like this for maximum verbosity:

    $ sh -x ./test.sh

Making a new release
--------------------

 - CHANGES.rst: Add a new entry
 - setup.py: Bump VERSION
 - src/img2pdf.py: Bump __version__
 - Commit:

    $ git add CHANGES.rst setup.py src/img2pdf.py
    $ git commit -m "release version X.Y.Z"

 - Add git tag:

    $ git tag X.Y.Z -m X.Y.Z

 - Build and upload to pypi:

    $ rm dist/*
    $ python3 setup.py sdist
    $ twine upload --sign dist/*