This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Linkage of GPLed GCC to Closed Source via XML or Perl

One last word on this for the record,
I will put this in the licence agreement for my
software and bring it to the right forum about

0. My intent is to create a free software and support
the GNU cause. The users of my software must also be
free software and support the cause as well. The work
that I have done is a derived work based on tree.h. It
is not linked normally, but linked via the filesystem
and network. The data exchanged between my modules and
free software is so complex and dependent on the GCC
code that it must fall under the GPL.

I am also bringing this into the public so that we can
create a set of fair rules that will cover this form
of interaction and bind for others who are also
creating derived works from the GPLed code including

1. The work that I have done with the compiler is
meant to implement an introspection into a running
program. Like reflection in java. It is not for
creating backends that are closed source.It is for
understanding and documenting the tree nodes. It is
for creating an object-oriented interface to the
compiler. It is for support meta-programming. All of
this is prototyped in the perl-modules that I have
published. They are all strongly derived from the
original (art) work done by richard stallman and the
great gcc team. 

2. I wrote to this topic to Richard Stallman.
Richard Stallman said to me in the question if the
data exchange over the network is not linking and
therefore not covered by the GPL 
"We have a different interpretation of the situation.
Connecting modules through sockets or pipes does not
necessarily mean they are separate programs. In simple
cases they are separate, but not when they exchange
complex data structures."

That would support the idea that all these are derived
works and fall under the GPL.

3. Linus Torvalds said :
>Feel free to consider this email (in its >entirety,
not snipped into pieces) as being >public, so if you
think you want to post it, go >ahead. 
>The GPL notion of "linking" is really nothing but >a
specific technical way of trying to define >"derived
>From a legal standpoint, technical issues have >some
validity, but in the end the _only_ thing >that
matters is whether it is derived or not. >Linking is
only one (strong) indicator that it is >indeed
derived. There are others. There are
>counter-indicators as well, of course, one of >them
being "previous work" (thus my willingness, >for
example, to have binary modules that were >basically
derived from SCO device drivers that >existed prior to
Linux - one of the original >impetuses for the module
>And intent matters. 

So from Linus's standpoint, 
All of these tools the use the tree nodes are derived
works like mine. 
All we need to do now is to get this in a licence
agreement that makes sure that the users of the module
who create derived works are all GPLed.

James Michael DuPont

James Michael DuPont

Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every occasion!

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]