Summary
Extract the early validation at the start of ImageConverter::process_image (empty path, input existence / filesystem errors) into a dedicated function (working name: check_input) and invoke it from process_image. Unit tests should be able to exercise this validation path directly instead of relying only on the full process_image pipeline (configure, decode, write, etc.).
Background (for newcomers)
Today, checks live inline at the top of process_image in src/rawtoaces_util/image_converter.cpp (~empty filename, then std::filesystem::exists inside a try/catch). Tests such as test_last_error_message_empty_filename and test_last_error_message_file_not_found in tests/test_image_converter.cpp already call process_image, but splitting check_input makes it easier to test input validation only fast and deterministically, and keeps responsibilities clearer for future Dev Days contributors.
Where to look
src/rawtoaces_util/image_converter.cpp — ImageConverter::process_image (opening block through filesystem validation).include/rawtoaces/image_converter.h — whether check_input should be public for tests/bindings or test-only via a test helper pattern consistent with this project.tests/test_image_converter.cpp — tests that assert status / last_error_message for invalid inputs.tests/python/test_image_converter.py — Python tests that call process_image for invalid paths (behaviour should remain consistent unless deliberately changed with docs).
Suggested direction (not a prescription)
- Factor the validation block into something like
bool ImageConverter::check_input(const std::string &input_filename) (or free function if that fits the codebase better).
- Update or add C++ tests that call
check_input where it reduces reliance on the full conversion path.
- Preserve existing observable behaviour for API consumers unless there is a documented reason to change it.
Summary
Extract the early validation at the start of
ImageConverter::process_image(empty path, input existence / filesystem errors) into a dedicated function (working name:check_input) and invoke it fromprocess_image. Unit tests should be able to exercise this validation path directly instead of relying only on the fullprocess_imagepipeline (configure, decode, write, etc.).Background (for newcomers)
Today, checks live inline at the top of
process_imageinsrc/rawtoaces_util/image_converter.cpp(~empty filename, thenstd::filesystem::existsinside a try/catch). Tests such astest_last_error_message_empty_filenameandtest_last_error_message_file_not_foundintests/test_image_converter.cppalready callprocess_image, but splittingcheck_inputmakes it easier to test input validation only fast and deterministically, and keeps responsibilities clearer for future Dev Days contributors.Where to look
src/rawtoaces_util/image_converter.cpp—ImageConverter::process_image(opening block through filesystem validation).include/rawtoaces/image_converter.h— whethercheck_inputshould bepublicfor tests/bindings or test-only via a test helper pattern consistent with this project.tests/test_image_converter.cpp— tests that assertstatus/last_error_messagefor invalid inputs.tests/python/test_image_converter.py— Python tests that callprocess_imagefor invalid paths (behaviour should remain consistent unless deliberately changed with docs).Suggested direction (not a prescription)
bool ImageConverter::check_input(const std::string &input_filename)(or free function if that fits the codebase better).check_inputwhere it reduces reliance on the full conversion path.