Re: Several copyrights wrong in R packages
On 10/21/21 1:24 AM, Andreas Tille wrote:
Am Thu, Oct 21, 2021 at 12:58:53AM +0530 schrieb Nilesh Patra:
I think this is a resultant of using routine-update without proper scrutiny of the
target package -- which is bad.
I agree that routine-update makes it even more easy to not (re)inspect
the copyright. However, the script is not freeing the maintainer to
do all necessary steps. It is just doing what can be automatized and
checking copyright in general can not really be automatized.
Agreed, don't get me wrong, r-u is a nice tool,
but I really think that it has a side effect -- making maintainer sloppy (atleast sometimes), hence
I asked questions to prevent this.
1) Is there some plan to fix copyrights team-wide?
Not that I'm aware of.
Can this happen in some sort of (mini-)sprint?
Again, this could be a nice idea for "newcomers" to work on -- just to get started making contributions.
What say?
2) Can reliance on routine-update be somewhat reduced?
That's a personal = by maintainer question IMHO. I personally need to
admit that routine-update is the only way I can manage the actual number
of packages. Otherwise I would leave lots of packages behind. That's
why I will stick to it.
Sure, but I urge you to check the copyrights once before upload.
Copyrights changed from GPL to MIT will not take you more than one minute to change.
-- Can it give some sort of warning for these things or so?
I think in the special case of CRAN packages we have a pretty good
chance to do so:
1. Diff between DESCRIPTION file of the HEAD of upstream branch
compared with the previous version and issue a warning if the
string in the License field is different should be relatively
easy to accomplish
This sounds good.
2. dh-update-R could also update d/copyright if needed.
It's difficult to do it in a "reliable" way, so I prefer avoiding automatic copyright modifications.
As always patches are welcome.
No time :(
This might not sound all that important, but conflicting licenses can become RC in future, which is
bad again, thought.
Yes. And once such a bug is filed we can fix it.
Ofcourse, but again, what I wanted to say is to prevent them, instead of "curing" them later. Not very important,
but if there are a number of bug reports, it causes an enormous amount of mess to deal with when number of packages at scale
are more.
And new versions will add in new files, and thus new copyrights so this will be a never ending stream of bugs, and hence
bring more watchful could help.
Ofcourse, I do not blame you or anyone else for not 'upto date' copyright, but we defnitely need this one-more-step after routine-update is ready to upload.
Nilesh
Reply to: