Some JPEGs are identified as MPOs

[howff]

Due to a "feature" of the Python imaging library Pillow/PIL some JPEGs are identified as MPO, i.e. when jpegcrop checks for image.format=="JPEG" it misses those JPEG images where image.format=="MPO".

python-pillow/Pillow#4603

MPO files are normal JPEG images but contain a tag which can contain other data which may be additional JPEG images, possibly stereoscopic or sometimes a larger thumbnail.

If the utilities which jpegcrop calls (exiftool, jpegtran) can handle MPO-style JPEG then jpegcrop can simply check for =="MPO" as well as =="JPEG". If there is a danger that the user will get unexpected behaviour (eg. a loss of valuable data) then a warning should be issued first.

In my case it's simply a regular JPEG from a Canon camera, I've no idea why it's MPO and not JPEG, so I just want it handled exactly the same as a JPEG.

[ConradHughes]

Hi howff, thanks a lot for this — and sorry for my slow response to your other issues — I've been wondering for a while why cropgui doesn't invoke jpegtran on some JPEGs, and I think you've just explained why. If you've a sample image that can reproduce this, I'd be very grateful — otherwise I'll have a search of recent images of my own when I get time.

(the bug report you link used to have a sample image, but no more..)

[howff]

Did I not submit a PR to fix this? Sorry, I'll try and submit on ASAP