Re: Problems with pkgsrc on NetBSD-5.2

[David Brownlee <abs%absd.org@localhost>]

On 27 July 2017 at 12:20, Mouse <mouse%rodents-montreal.org@localhost> wrote:
>>>> Just as a side note, NetBSD 5.2 itself has been out of support for
>>>> almost two years.
>>> Except, of course, for the only support it's ever really had: that
>>> provided by community volunteers.
>> Perhaps so, but there will be no Security Advisories, official
>> patches, updates, etc.
>
> Not from TNF, no.  True.  Support will be significantly more haphazard
> in those respects.
>
>>>> You should upgrade to NetBSD 7.x as soon as you can.
>>> Not that I'm the person you were talking to, but, speaking
>>> personally, fix the licensing and I might.
>> Okay, I'll bite, what's the problem with the license?
>
> Core let GPLv3 into the tree (in the form of a gcc update, I think it
> was, though exactly what it was is pretty much irrelevant).
>
> I've tried to read the GPLv3 twice, and people tell me I've _still_
> missed major pieces.  If I have to hire a lawyer to make sense of the
> license, it's not suitable for my purposes no matter _what_ it actually
> ends up permitting.
>
> The real problem to my mind, though, is that there is no (apparent)
> awareness that this was even an issue[%], much less any commitment to
> not do something like it again.  And it's _me_ that would be on the
> hook for a potential copyright violation.  So, any version jump after
> that point calls for me to examine the license for every last file in
> the tree, and probably throw out significant fractions which are
> distributed under GPLv3 or similarly unacceptable licenses.  This would
> be a hell of a lot of work for, quite likely, a system crippled into
> unusability (by the parts I had to throw out).

My recollection of the discussion was that the issue was recognised as
significant - to the extent that all gpl3 licenced code is kept in a
separate section of the source tree (external/gpl3), which currently
contains autoconf, binutils, gcc & gdb.

When it has been subsequently raised it has been made clear that gpl3
cannot be introduced to code outside that area, whether it be
upgrading to newer versions, importing, or backporting a patch from
code which has moved to gpl3.

I believe that one of the motivations behind the clang effort is to be
able to have a self hosting system completely clear of gpl3 when
desired (not that the clang work is not worthwhile in of itself :)

David