Tesseract creates output for missing input

[stweil]

If Tesseract is called with a non-existing image file, it creates an empty output file.

I'd expect that no output is created in that case.

Examples:

# Empty output.txt is created.
tesseract nonexisting_image.tiff output
# Empty output.hocr is created.
tesseract nonexisting_image.tiff output hocr

[Shreeshrii]

You are right. The error is also there for pdf. So, even though error messages are shown, empty output is produced.

 tesseract nonexisting_image.tiff output pdf
Tesseract Open Source OCR Engine v4.00.00dev-549-g2b854e3 with Leptonica
Error in fopenReadStream: file not found
Error in findFileFormat: image file not found
Error during processing.

[Shreeshrii]

@stweil for #1423

[zdenop]

Problem is that output file is created during initialization of renderer and it happens before starting of OCR...

Maybe we can implement simple check if input file exists....

[zdenop]

Done.
But there could be scenarios where empty output is created. E.g. if input file exists but is malformed or leptonica can not process. That could be solved only by changing logic of renderer: create file only when we have something for writing to file. Or remove file if OCR output is empty.