2007/12/13, Travis Willard <travis@archlinux.org>:
On Dec 12, 2007 8:55 PM, Eric Belanger <belanger@astro.umontreal.ca> wrote:
On Wed, 5 Dec 2007, Eric Belanger wrote:
On Wed, 5 Dec 2007, Paul Mattal wrote:
Pierre Schmitz wrote:
Perhaps a short comment within the PKGBUILD might be usefull to explain such things.
I have reverted the changed, added such a comment, and moved the already signed-off i686 package to core.
- P
For the license, it might have been better to simply use license=('custom') By having 'ISC' by itself, it implies that ISC is one of the common licenses which it is not.
Also, in the future, can we keep the packages in testing until it get signed off for both architectures? Apart from the fact that it will be more foolproof as more people had looked at it, we should try to keep the repo for the 2 architectures as in sync as possible. Otherwise, we might get complaints and bug reports about why the x86_64 package is still in testing. Also, it is simpler for us to keep track because it will be hard to tell after some time why the x86_64 package is still in testing. Is it because it's waiting to be signed off, because no-one noticed that it was signed off shortly afterward or if it was just forgotten?
Is there someone else thinking the same? We should get a resolution on the license field so we can fix it as needed and get the required signoff.
I was told ISC was a 'common' license like BSD, where you still needed to install a license to /usr/share/licenses/$pkgname but it's in enough use that we support it as a common one.
Was I misinformed?
You were informed correctly. ISC is pretty much like 2-clause BSD license (see wikipedia). The most known package that uses it is, obviously, bind. It is a common license in Arch - in namcap 2.0 and the latest version of Travis' checklicense script. So it shouldn't be listed as just license=('custom'). -- Roman Kyrylych (Роман Кирилич)