Description
This is a problem found in ClearlyDefined by @bduranc
Here is the gist of a chat conversion:
duranbc: @philippe : Not necessarily a bug, but perhaps an interesting observation.
The following component shows (C) Hewlett-Packard, however the statement appears to be extracted from the metadata of a JPEG image.
CD Definition: https://clearlydefined.io/definitions/git/github/h2non/filetype/b2d66fbb5aed4a1fe9f0ca71a109b9161373866a
If you look inside the file "sample.tar", you will find a few JPEG images. The metadata has the HP copyright statement inside it.
https://github.com/h2non/filetype/blob/b2d66fbb5aed4a1fe9f0ca71a109b9161373866a/fixtures/sample.tar
I did not know CD goes this deep into binary files, and have not taken notice of this before.
Philippe:
actually this is scancode diving in these binaries quite happily (it is designed to do this)
So Copyright (c) 1998 Hewlett-Packard Company is correct there. BUT one test sample image... this is also noise, is it?
actually this very specific copyright "Copyright (c) 1998 Hewlett-Packard Company" in a JPEG file is a problem eventually: https://www.flickr.com/help/forum/en-us/7987/
It sounds like this is present in many JEPG files BUT at the same time this is not in the exif metadata of the image
net-net, I think this should be always excluded when found exactly in a JPEG file
or not... there seems to be more to it ...
See nodejs/node#5749
So ScanCode may not be wrong there AND there is a possibly weird underlying issue ... that said these images (the one you found) are test fixtures and therefore not core code, so this is minor in this case
Description
This is a problem found in ClearlyDefined by @bduranc
Here is the gist of a chat conversion: