netbeans-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Wade Chandler <wadechand...@apache.org>
Subject Re: Optional modules with GPL dependencies (was: What to include/exclude in code donation to Apache)
Date Sat, 05 Nov 2016 19:20:25 GMT
On Nov 5, 2016 2:00 PM, "Neil C Smith" <neilcsmith.net@googlemail.com>
wrote:
>
> Hi,
>
> Is top-posting or inline replying the norm here?  Anyway, replies below
:-)
>

It depends on devices and whether one is replying to pieces or a whole I
guess; for inline. I do think in the age of diverse connected smarter
devices, old arguments about top, bottom, and inline posting limits
communities, and as long as something is understandable, then we shouldn't
fret much...not saying you are, just my frame of reference take on this.
Snipping for replies seems more clear to me, but if one is limited, then
chime in and contribute as one can IMO.

>
> > On 5 November 2016 at 16:45, Wade Chandler <wadechandler@apache.org>
wrote:
> > Perhaps specific threads can be created to discuss the Java and
JavaScript support in NetBeans (independently/maybe together). I think a
big part of the question is what does Oracle think about supporting
nb-javac and graal.js if they are only used by NetBeans. I mean, is there
some long term support strategy there?
>.  I should
> have said, having the regular OpenJDK tooling supporting what NetBeans
> requires surely benefits all projects building similar tools.
>

Sure, and I can certainly see that POV, but it has to be there first. So,
there is the now, future, and whatever the possibilities. There are ways to
do this now, and ways to make it better. That to me can be iterative, but
has to move forward from where we are. My guess is we can all agree on that.

> My concern is that without the Java cluster I don't believe NetBeans
> is really a viable project, and for that cluster to be viable requires
> the long term support of nb-javac that you mention.

I agree with this. I think the community can figure this out regardless
though. Whether it the NB community, Oracle, or OpenJDK who keeps up
nb-javac or Java support.next, and if we iterate, neither use case puts the
Java cluster outside NB, nor makes NB unviable.

>Why I think this
> is different to reliance on almost any other library is the way javac
> uses the internals of the JRE.  eg. the last time I looked, javac had
> specific exceptions in place to bypass module restrictions in Java 9.
> It just feels like shakier ground to rely on than it could be.
>

You would have to get specific here. nb-javac has a license now. It is
GPL+CPE, so it specifically, and how that would legally and technically
work per what is in the Apache legal "resolved" document, is a straight
line IMO; nothing to guess about other than implementation. The same would
be the case with any OpenJDK API ATM.

> Assuming the answer to my licensing question is no, then I'm
> interested in exactly how much nb-javac forks from javac and how
> maintainable it is from outside?

That could be done any way needed IMO. It could be forked and put on GitHub
directly by the NB community once NB is at Apache. So, it would just be an
OSS standalone component managed by the NB community; not Apache nor
Oracle. Other options are clearly open to discussion. I would like to hear
Oracle's POV on this.

>Could nb-javac features be merged
> with javac, or become ide-javac within OpenJDK, providing similar
> services to any IDE?
>

This is a future question to me. Could it be asked and reviewed? I imagine
that answer is yes. How long? Can it actually happen? All would take
working with the OSS project which is OpenJDK to find out. Oracle perhaps
can help with that. Either way, that doesn't make NB less viable. You use
it now...

Thanks

Wade

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message