Jim Henderson wrote:
> This is contrary to other discussions I've read on the topic.

It's what the FSF says also, as far as I know.

> If you include code I wrote in your code, then you have to respect my 
> wishes about the use of the code.

Then you're contradicting yourself there.

> Sure.  What ultimately may come out of the whole GCC kerfuffle is another 
> compiler without those restrictions, if the folks over at the FSF insist 
> on putting stupid restrictions in place.

Then it'll be a battle to see whose version gets incorporated into various 
distros. Which is exactly the problem I'm forseeing.

-- 
   Darren New, San Diego CA, USA (PST)
   "Ouch ouch ouch!"
   "What's wrong? Noodles too hot?"
   "No, I have Chopstick Tunnel Syndrome."

Post a reply to this message

Darren New <dne### [at] sanrrcom> wrote:
> somebody wrote:
> > 2) Don't release code which is based on or interfacing with GPL'ed code.
>
> How come I can write code that interfaces with Linux without violating the
> GPL, but I can't write a plug-in for GCC without violating the GPL?

You're beginning to run into contradictions there.  Weren't you the one saying
the GPL doesn't enforce this behaviour?  That the measures taken are exactly
because the GPL alone would allow for closed plugins?

Other than that, you're right.  I believe gcc developers see plugins in the same
light as dynamic libs:  app + libs form a larger work and if one of the
components is GPL, the other should follow or not use it at all.

Post a reply to this message

Darren New <dne### [at] sanrrcom> wrote:
> Jim Henderson wrote:
> > This is contrary to other discussions I've read on the topic.
>
> It's what the FSF says also, as far as I know.
>
> > If you include code I wrote in your code, then you have to respect my
> > wishes about the use of the code.
>
> Then you're contradicting yourself there.
>
> > Sure.  What ultimately may come out of the whole GCC kerfuffle is another
> > compiler without those restrictions, if the folks over at the FSF insist
> > on putting stupid restrictions in place.
>
> Then it'll be a battle to see whose version gets incorporated into various
> distros. Which is exactly the problem I'm forseeing.

Something like that happened before.  See egcs.

BTW, I've finally taken the time to read more deeply into the GCC plugin
controversy and by searching through the mailing lists, I reached this:

http://www.gnu.org/licenses/gcc-exception-faq.html

It's interesting in that it brings to attention a point that very few seem to
notice when compiling with GCC:  the object code produced by GCC is statically
linked to a myriad of small GPL'd libs and yet you're able to release such
produced code under any license you want.  How can it be?  You see, there was
an exception in the GCC licensing that allowed for this.  So, with GPL3 around,
they thought it would also be a good time to update that exception.  Don't
worry, you can still license the produced code any way you wish.  But the
plugin architecture is what deserved the attention of the updating. ;)

Still, their reasoning seem sensible enough...

Post a reply to this message

nemesis wrote:
> Darren New <dne### [at] sanrrcom> wrote:
>> somebody wrote:
>>> 2) Don't release code which is based on or interfacing with GPL'ed code.
>> How come I can write code that interfaces with Linux without violating the
>> GPL, but I can't write a plug-in for GCC without violating the GPL?
> 
> You're beginning to run into contradictions there.  Weren't you the one saying
> the GPL doesn't enforce this behaviour? 

Yes. Why is "somebody" telling me a choice is to not release code that isn't 
based on or interfacing with GPL code? In other words, the implication of 
the text above is that the only way to release a gcc plug-in is via GPL. But 
that isn't true. It's what FSF wants to be true, but it isn't true. Were it 
true, it would be the same situation as applications on Linux.

> That the measures taken are exactly
> because the GPL alone would allow for closed plugins?

Ding ding ding! Yes, exactly. I'm pointing out that no, the GPL alone allows 
for closed plug-ins in exactly the same way that GPL allows for closed 
applications running under Linux.

FSF wants to figure out a technological way of forcing others to only GPL 
code that runs as a plug-in to gcc. I feel that's over the top. I know it's 
their right to do whatever the f__k they please with their code. I don't 
believe it's the case they can enforce that thru their current license. I 
believe that them using technological means to prevent closed plug-ins from 
being written is similar to the technological means TiVo uses to prevent 
your code from running on their hardware, which FSF has complained about, 
hence making the FSF hypocritical. Does that make it clear?

> Other than that, you're right.  I believe gcc developers see plugins in the same
> light as dynamic libs:  app + libs form a larger work and if one of the
> components is GPL, the other should follow or not use it at all.

Except that they're wrong, depending on what you mean by "should". If by 
"should" you mean "the FSF would like that to be the case", then yes. If by 
"should" you mean "the FSF thinks that's already the case", then no.

-- 
   Darren New, San Diego CA, USA (PST)
   "Ouch ouch ouch!"
   "What's wrong? Noodles too hot?"
   "No, I have Chopstick Tunnel Syndrome."

Post a reply to this message

nemesis wrote:
> Still, their reasoning seem sensible enough...

Yes. It sounds very ... commercial.  "Likely to backfire..." Think about 
what that phrase means.  But I'll agree that they seem to have patched up 
the license to manage to make the *output* code from GCC fall under the GPL 
even if the plug-in doesn't. That's fine with me.

Also not sure why non-free software is "bad for society." I guess if you 
take that as an axiom, it makes sense, but I disagree that it's as obvious 
as FSF makes it out to be. Not sure why non-free software is more "bad for 
society" than non-free manufacturing techniques, non-free music, etc.

-- 
   Darren New, San Diego CA, USA (PST)
   "Ouch ouch ouch!"
   "What's wrong? Noodles too hot?"
   "No, I have Chopstick Tunnel Syndrome."

Post a reply to this message

On Fri, 30 Jan 2009 18:57:50 -0800, Darren New wrote:

> Jim Henderson wrote:
>> This is contrary to other discussions I've read on the topic.
> 
> It's what the FSF says also, as far as I know.

Perhaps, but it's not in line with other discussions I've seen on the 
topic. <shrug>

> 
>> If you include code I wrote in your code, then you have to respect my
>> wishes about the use of the code.
> 
> Then you're contradicting yourself there.

No, I don't think I am.  You can use my code under my terms.  How does 
that contradict anything I've said so far?  If I write code, I get to say 
how it's used.  I have consistently said that.

>> Sure.  What ultimately may come out of the whole GCC kerfuffle is
>> another compiler without those restrictions, if the folks over at the
>> FSF insist on putting stupid restrictions in place.
> 
> Then it'll be a battle to see whose version gets incorporated into
> various distros. Which is exactly the problem I'm forseeing.

Possibly, though I don't see it as a "problem" per se.  I don't see 
choice as a problem. :-)

Jim

Post a reply to this message

Jim Henderson wrote:
> No, I don't think I am.  You can use my code under my terms.  How does 
> that contradict anything I've said so far?

That I can use your code for commercial gain if I just link it in without 
changing it?

You said
"""
Yes, but again, linking to something that's GPL'ed doesn't mean your
program has to be GPL'ed.
"""

Then you said
"""
If you include code I wrote in your code, then you have to respect my
wishes about the use of the code.
"""

> Possibly, though I don't see it as a "problem" per se.  I don't see 
> choice as a problem. :-)

It wouldn't be, if choice was really that easy.

-- 
   Darren New, San Diego CA, USA (PST)
   "Ouch ouch ouch!"
   "What's wrong? Noodles too hot?"
   "No, I have Chopstick Tunnel Syndrome."

Post a reply to this message

nemesis wrote:
> worry, you can still license the produced code any way you wish.  But the
> plugin architecture is what deserved the attention of the updating. ;)

Now here's a question for ya:

"""
However, if you used GCC in conjunction with GPL-incompatible software 
during the process of transforming high-level code to low-level code, that 
would not be an Eligible Compilation Process. This would happen if, for 
example, you used GCC with a proprietary plugin.
"""

Let's look at a couple of scenarios:

I modify gcc, but I never distribute my modifications. Can I license the 
resulting "Target Code" under a proprietary license?

I buy a proprietary plug-in that does something like (say) Purify, or 
performance measurements, or helps with debugging, by modifying the 
intermediate representation to include calls to the libraries used by the 
plug-in. I then run my code, improve the performance, fix the bugs, then 
recompile without the plug-in. Doesn't sound like the final release-mode 
executable is restricted.

How about if I develop my own proprietary plug-in that I never distribute. 
Is the target code still covered?

How about if I develop a plug-in, release it dual-licensed under 
"binary-only" for $100 a copy without permission to distribute, or under GPL 
and sell the first copy for $1,000,000 under GPL? It's available under GPL, 
now, so that makes it Eligible, right, so the code created by the plug-in is 
not GPLed.

What if the plug-in outputs stuff that compiles? It's not intermediate code. 
What if I have a plug-in that outputs LISP code that does the same thing as 
the C you input? Or which lets you input LISP which can then get translated 
down to executable code. Is LISP an "intermediate language"?

I don't think this is going to do what I think they think it's going to do.

-- 
   Darren New, San Diego CA, USA (PST)
   "Ouch ouch ouch!"
   "What's wrong? Noodles too hot?"
   "No, I have Chopstick Tunnel Syndrome."

Post a reply to this message

On 1/28/2009 12:29 PM, Darren New wrote:
> Right. That's my point. Why would GCC require a plug-in author to give
> it a particular license? It wouldn't be any different from the Linux
> kernel enforcing only running apps that were GPLed.

Shh... don't give them any ideas!!!

-- 
...Chambers
www.pacificwebguy.com

Post a reply to this message

Darren New wrote:
> I don't think this is going to do what I think they think it's going to do.

Or one more thought ... What if I use the gcc front-end, a plug-in that 
saves the intermediate code to disk, and then I compile that intermediate 
code down to something that doesn't use gcc libraries? (Java bytecodes, .NET 
IL, or machine code that I generate completely within the plug-in...)

Doesn't sound like they're managing to cover everything it sounds like they 
think they're trying to cover. :-)

-- 
   Darren New, San Diego CA, USA (PST)
   "Ouch ouch ouch!"
   "What's wrong? Noodles too hot?"
   "No, I have Chopstick Tunnel Syndrome."

Post a reply to this message