Hi Stuart,

Stuart Henderson wrote on Thu, Jul 17, 2014 at 10:05:20AM +0100:

> What attitude should ports take to license markers in the cases:

General rules are hard to set up, my impression is that a variety
of details can influence whether a Copyright notice and a license
grant is effective.

> 1. where copyright notices are not included in each file, but there
> is a common file (such as in this case COPYHEADER) with a proper grant

As far as i'm aware, it is of little relevance where the Copyright
notice is, as long as it is clear what it refers to.  Putting it into
each file is certainly good practice and helps clarity, but doesn't
seem strictly required to me.

In the case at hand, all material written by Thomas Dickey is clearly
fine: His Copyright notice is contained in COPYHEADER, and it clearly
states which license he grants right below.  Regarding contributions
by other authors, the situation is more fuzzy.  I could find no
formal Copyright notices for any of them.  So technically, I'm not
100% sure their grants are effective.  There is little doubt, however,
that they intend to publish this under GPL.  Besides, taking the files
CHANGES and AUTHORS together, it is possible to identify individual
contributions, so it's really just the formal notice that is missing,
not the information who grants what on what.

I think applying an even more formal standard and calling this
insufficient for lack of effective secondary contributor licensing
would be unresonably picky; by the same standard, we could stop
distributing the base system because lots and lots of developers
contribute copyrightable content to various files without adding
their personal Copyright notices, or forget to update years.
Some, on the other hand, have done wholesale year updates across
whole subtrees without adding content to each of the bumped files.

> 2. where a README has a notice along these lines:
> 
> Some source files contain a license notice; all other source files are
> licensed under the GNU Public License (GPL), of which you can find a copy
> in the file 'GPL.txt'.
> 
> - where GPL.txt is just a copy of the GPL text from FSF, no copyright
> notice specific to the distribution - no other files have proper grant
> (i.e. no Copyright <name> <date> "you can redistribute this under xyz
> terms") - but it's clearly intended to be under GPL

If there is really no proper Copyright notice whatsoever, not even
from the main developer, i think PERMIT_*=No is unfortunately correct.
It should be easy to convince upstream, though, that they need to add
that one line to make their grant effective...

Yours,
  Ingo

Reply via email to