Nicolas Mailhot wrote:
Ok, the fedora-legal verdict is in. Ignoring the GPLv2 files I had already removed it seems there is a problem with cgats/License.txt & icc/License.txt https://bugzilla.redhat.com/show_bug.cgi?id=421921#c13
I'm rather surprised, since Ghostscript reviewed this licence, and has no issues with it.
2. dual-licence this code under GPLv3 and its current licence. The code integrated in argyll would be under GPLv3 clauses, and the original licensing conditions would still be available for non-GPLv3 projects
This is not as straightforward as it seems, because it is necessary then to have a condition that the two offered licenses not be removed, and the phrasing of this type of condition appears to be where the problem is with the existing libusb/cgats licenses.
3. select another license already considered to be GPLv3-compatible by big actors like the FSF, and use it as-is without changing its text (which may trigger unforeseen compatibility problems). For example, LGPL v3, 3-clause BSD, etc
If it makes things more straightforward, then I'm prepared to simply switch to the MIT license for these two libraries.
My preference would be 3. or 1. but that's something for the authors to decide. In the meanwhile Fedora Argyll CMS integration has been halted.
Given Ludovico Fischer's report, and my experience with getting instruments to work with USB, I think you need to take another look at the libusb issue as well. If you're going to try and change the version of libusb that Argyll is linked to, then I think it is quite necessary for you to do regression testing with all the supported instruments, to properly qualify it. Graeme Gill.