[ILUG] GPL clarification.

[Bryan O'Donoghue]

Paul Jakma wrote:
> On Thu, 31 Mar 2005, Liam Bedford wrote:
> 
>> AIUI, if you link to GPL code, your binary is going to be GPL, and 
>> thus the source will be to.
>>
>> There's a reason the LGPL exists. glibc is LGPL I think.
> 
> 
> Note that the major difference between the two is that LGPL tries to be 
> more specific about constraining its effects on the work comprising the 
> library itself (see section 5 of the LGPL).
> 
> However, neither the GPL nor the LGPL either allow or disallow 
> binary-only works to link to them. Both are framed in terms of requiring 
> "derived works" to have source made available.

I'm ambiguious on the meaning here.

If libblah.a is GPL, and evilcorp does a

gcc -o marketable_binary_a propiatery_license.c -lblah
gcc -o marketable_binary_b propiatery_license.c -Bstatic -lblah

does that imply that marketable_binary_a is then considered a derived 
work and must be GPL'd ?

As far as I understood it up until Liam's post, only if I for example 
changed libblah.a and recompiled it and/or statically linked (not sure 
about statically linking) then the 'modified' GPL'd code would have to 
be distributed.

Further, my understanding was that if I took libblah and added lots of 
bits of C code to it and gave it to <another> that at that point the 
entire lot of libblah+super_modification would be as a derived work, GPL'd.

Personally, I *think* it's ... well slimy, to make a function call to 
GPL code, to rely on a GPL bit of code and not .. well, give back, in 
the same ilk, but again.. I guess I'm just looking for clarification.

Is marketable_binary_a under an obligation to be GPL?
Is marketable_binary_b under an obligation to be GPL?

I don't think that either *are* at all, only libblah+super_modification 
is, but... I could be wrong.

--
Bryan