On Thu, Oct 23, 2008 at 4:18 PM, Ricardo SIGNES < [EMAIL PROTECTED]> wrote:
> * Bill Ward <[EMAIL PROTECTED]> [2008-10-23T17:11:09] > > On Thu, Oct 23, 2008 at 1:49 PM, Ricardo SIGNES < > > > Gabor is not suggesting that it be required to upload to PAUSE, but > that it > > > be required to 'make dist.' This change would, perforce, require yet > > > another new version of EU::MakeMaker et al. > > > > Well I have no problem with that. But how about changing PAUSE? > > > > Perhaps when you upload to PAUSE without a license in META.yml it could > > actually replace the META.yml with one that has a license, based in input > > from an HTML form? Would that be too weird? I think it's technically > > feasible. > > Too many authors use one of the various cpan-upload scripts, all of which > would > now break. I've never seen these upload scripts but I suppose it's a nobrainer that they should exist. If no license is in META.yml, and none is specified through the upload process, then it would be rejected. That's the behavior that was proposed anyway. And if the uploader scripts want to play ball with the new rules we could easily allow them to specify &license=gpl2 or something in the URL they are uploading to or whatever API they're using. Under the proposed plan, uploader scripts would fail unless there was a license in META.yml anyway, so this is a red herring. > I think it's likely to be a big pain. What if the user puts a 'license X' > declaration all over all the files, but forgets to include a license in > META.yml. Then PAUSE puts in a default, from the user's profile? Now > there is > a conflict. Also, if this was set in some thing while uploading, the user > would have to specify it every time? I mean, if he wanted to specify the > license once and be done with it, he'd make sure he got a META.yml. All problems that could be solved through good usability design. The ability to set a default per-user would help, as would the ability to recycle the license used the last time. > The best solution will be, eventually, to reject dists without a valid > META.yml. I don't think META.yml or its producers are ready for that, yet. I agree. My idea of adding this feature in PAUSE would provide a stopgap solution, though.
