Closed maximelenormand closed 3 weeks ago
Thank you! Fixed with https://github.com/TommyJones/tidylda/commit/881f92ebca38f08d7671df055eca1caa978d09e4
Actually, I'm reopening this.
CRAN has a very esoteric take on license files. Removing the LICENSE file with just my name and date and/or renaming LICENSE.md (which has MIT license language) means that the package no longer passes checks to be hosted on CRAN.
I understand that it's confusing to have both LICENSE and LICENSE.md files. So, I'm leaving this open until I can satisfy both requirements of having one file and passing CMD check for CRAN.
Ok. What works is removing the MIT license file and leaving the file with just a copright statement. The MIT license is a standard anyway, so I don't think removing it from the repo substantially changes understandability of the copyright.
The commit that passes CMD check and has only one license file is https://github.com/TommyJones/tidylda/commit/846adcced612dfcaa26098c077aad9ba53ddc5ff
@maximelenormand is this solution satisfactory?
The problem now is that the license can only be identified in the DESCRIPTION file.
This is strange that you cannot pass the CRAN checks with one file. I did not encounter this problem with my packages (example https://github.com/EpiVec/TDLM/) but it is maybe due to the MIT license.
If you cannot solve the problem, you can keep the two license files. It is better to clearly identify one well-known license, even if having two files may be confusing.
Understood. I'll revert to what I had originally. I'll close the issue when that's done.
It looks like other folks have similar issues.
As I'd originally set it up, I followed Hadely's guidance here, which was also suggested in the Stack Overflow post, above.
This is less than ideal but it seems to be the conventional wisdom.
39b751a1d34deab1fad47c1aaf33094133d34a8a
Hi Tommy,
My name is Maxime Lenormand. I will review your package over the next few weeks.
You have two license files, I recommend keeping only the MIT License file.
#6800