1 |
On Wednesday 25 February 2004 19:54, Brian Jackson wrote: |
2 |
> On Wednesday 25 February 2004 00:32, Jason Stubbs wrote: |
3 |
> > On Tuesday 17 February 2004 12:17, Donnie Berkholz wrote: |
4 |
> > > We won't be adding versions of XFree86 with the 1.1 license [1] to the |
5 |
> > > tree, so don't be surprised when 4.3.99.903 doesn't show up with the |
6 |
> > > new license. |
7 |
> > > |
8 |
> > > I won't elaborate on the reasons because it's been discussed quite |
9 |
> > > thoroughly in other forums. |
10 |
> > |
11 |
> > I hate to kick a dead horse, but... |
12 |
> > |
13 |
> > I've read all the reasons and understand them and don't wish to dispute |
14 |
> > them, but I do have one small question. What's the difference between |
15 |
> > this situation and the GPL'd Linux kernel "linking" against functions in |
16 |
> > the close-sourced BIOS, ACPI, APM, etc, etc? |
17 |
> |
18 |
> The kernel doesn't link against those things. Actually the kernel doesn't |
19 |
> link against anything ouside of the kernel itself (i.e. glibc, etc.) since |
20 |
> those things wouldn't be available when the kernel is starting. Using |
21 |
> hardware features is quite a bit different than linking against certain |
22 |
> libraries. It's kind of the same as the difference between kernel space and |
23 |
> user space (but not even close). We can have non-gpl'ed userspace programs |
24 |
> even if they call the kernel in some way (and they all do, even a open() |
25 |
> eventually gets to the kernel). |
26 |
|
27 |
After reading the GPLv2 closely several times with your words in mind, I think |
28 |
I have a hold on how it all works now. Below are excepts of the GPLv2 with my |
29 |
interpretations. I'll take no response as affirmation. :) |
30 |
|
31 |
<excerpts> |
32 |
0. A "work based on the Program" means either the Program or any derivative |
33 |
work under copyright law: that is to say, a work containing the Program or a |
34 |
portion of it, either verbatim or with modifications and/or translated into |
35 |
another language. |
36 |
|
37 |
3. For an executable work, complete source code means all the source code for |
38 |
all modules it contains, plus any associated interface definition files, plus |
39 |
the scripts used to control compilation and installation of the executable. |
40 |
However, as a special exception, the source code distributed need not include |
41 |
anything that is normally distributed (in either source or binary form) with |
42 |
the major components (compiler, kernel, and so on) of the operating system on |
43 |
which the executable runs, unless that component itself accompanies the |
44 |
executable. |
45 |
</excerpts> |
46 |
|
47 |
These two when combined mean that the source of a Program under the GPL |
48 |
includes all header files/libs/whatever that is required to be able to |
49 |
compile the Program. The source for the libs aren't part of the Program's |
50 |
source, but the libs themselves are. |
51 |
|
52 |
The special exception I read, not as applying to the operating system but, as |
53 |
applying to the base platform. Hence, this automatically covers all hardware |
54 |
and negates my question above. On Windows, this means an out-of-the-box |
55 |
installation plus any updates to the operating system that are available. |
56 |
|
57 |
On GNU/Linux, I understand it to be gcc, glibc & linux but beyond that the |
58 |
line is a bit fuzzy. In fact, glibc is licensed under the LGPL so I'm not |
59 |
even sure if that should be included in what makes up a GNU/Linux's base |
60 |
system. I can't see anything when flicking through the FAQ's TOC, but I'll |
61 |
read through that later and probably find the answer there. |
62 |
|
63 |
I'll intentionally leave out an interpretation of 6, because it's fairly |
64 |
straight-forward when read with this understanding of Program. |
65 |
|
66 |
Regards, |
67 |
Jason Stubbs |
68 |
|
69 |
-- |
70 |
gentoo-dev@g.o mailing list |