|
|
Subscribe / Log in / New account

The Software Freedom Conservancy on Tesla's GPL compliance

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 19, 2018 23:11 UTC (Sat) by epa (subscriber, #39769)
In reply to: The Software Freedom Conservancy on Tesla's GPL compliance by coriordan
Parent article: The Software Freedom Conservancy on Tesla's GPL compliance

Doesn’t the GPL define source code as the preferred form of the program for making modifications to it? And I doubt most Red Hat developers prefer to work from the monolithic code dumps for further changes. Normally if you want to make further changes to a project it is much more comfortable to work from the full history. So there is at least an argument to be made that this is evading the spirit of the licence.


to post comments

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 19, 2018 23:57 UTC (Sat) by coriordan (guest, #7544) [Link]

It's a long shot, but indeed, for some specific situation it might contribute something to a larger argument.

The relevant part of GPLv2 is:

"""The source code for a work means the preferred form of the work for making modifications to it. For an executable work, complete source code means all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation and installation of the executable."""

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 20, 2018 0:37 UTC (Sun) by excors (subscriber, #95769) [Link] (11 responses)

So no GPL program can get away with distributing just a tarball of source code, they'd always have to provide the full CVS/Git/etc repository? That doesn't sound compatible with how people have been commonly interpreting the GPL for decades.

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 20, 2018 5:51 UTC (Sun) by fest3er (guest, #60379) [Link] (10 responses)

IMO, no. As long as a software producer presents a tarball that contains everything needed to build a particular release (or update), *and* someone can follow those instructions and build everything, then they should be in compliance.

Of course, it would likely behoove them to use a source control system. It would make it easier for them and others to obtain everything for particular release and build it.

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 20, 2018 11:18 UTC (Sun) by epa (subscriber, #39769) [Link] (5 responses)

That can’t be enough in itself. You could distribute a tarball containing a binary blob. It has everything needed to reproduce the release, but it is not a useful form for further modifications. That’s why the GPL has that language defining ‘source code’.

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 20, 2018 13:56 UTC (Sun) by rahulsundaram (subscriber, #21946) [Link] (4 responses)

" That can’t be enough in itself. You could distribute a tarball containing a binary blob"

That wouldn't meet the criteria you are replying to. "As long as a software producer presents a tarball that contains everything needed to build a particular release (or update)" Build - being the operative word. It implies building from source code. Not a binary release.

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 21, 2018 6:32 UTC (Mon) by paulj (subscriber, #341) [Link] (1 responses)

So if a company runs the source code through an obfuscator (as NVidia used to do for the 2D X11 driver), would you say that satisfied the GPL? If it was little more than disassembled asm put through a "asm-to-C" tool, with random strings for symbols and labels, you think that'd suffice, because it would be something that'd build from "source"? :)

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 21, 2018 12:31 UTC (Mon) by rahulsundaram (subscriber, #21946) [Link]

A deliberate form of obfuscation is different from a tarball of the source code. I am unaware of any history of GPL source being distributed that way. The latter is far more common and my understanding is that SFC has considered that sufficient judging from Bradley's comment here before.

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 21, 2018 7:05 UTC (Mon) by epa (subscriber, #39769) [Link] (1 responses)

My understanding of 'build' is that you take smaller parts and put them together to a larger one. So you can build a Linux distribution DVD image from some binary packages, a kernel and ramdisk image, and a bootloader. The build does not have to begin at the original source code (indeed, perhaps most builds do not, since outside the world of free software it's common to have random binary crap pulled in as part of the build).

That must be why the GPL does not use the word 'build' but talks about the preferred form of the program for making modifications to it.

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 21, 2018 19:29 UTC (Mon) by rahulsundaram (subscriber, #21946) [Link]

> That must be why the GPL does not use the word 'build' but talks about the preferred form of the program for making modifications to it.

Perfectly fine to use build in the same context as long as you define it clearly within the license. We do use the term colloquially often to build from source to get binary artifacts.

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 21, 2018 6:17 UTC (Mon) by paulj (subscriber, #341) [Link] (3 responses)

There's your opinion, and there's what the licence says:

“The "source code" for a work means the preferred form of the work for making modifications to it.”

I know these days, where an upstream project uses git, that for me the preferred form for making modifications is via a clone of the git repo - and I suspect that would be true for most developers. If some downstream party has themselves used such a git clone as part of developing further modifications that is surely prima facie evidence that the git clone is the preferred form of making modifications. If that downstream party /deliberately/ strips out the change-set information (which most people prefer - as they do) in order to inconvenience any further development of those modifications.. well....

That often in the past people preferred tarballs, patch, diff, even shell based shar patches, has no bearing on what is preferred today. Preferences may and do change with time and better tools.

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 21, 2018 6:37 UTC (Mon) by paulj (subscriber, #341) [Link]

Oh, note what the /infringer/ preferred to use could well (and should?) control what the "preferred form" should be, for the infringer to come into compliance.

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 21, 2018 18:20 UTC (Mon) by khim (subscriber, #9252) [Link] (1 responses)

Yes even today some people prefer to used tarball is they have slow internet. Some use shallow clone --depth 1 to save space. and so on. So it's not so simple.

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 21, 2018 23:40 UTC (Mon) by paulj (subscriber, #341) [Link]

As per my sibling comment to yours, determining "preferred" in a general way is difficult. Much easier is to determine what the infringer preferred - then use that as the minimum bar, for their case.

The Software Freedom Conservancy on Tesla's GPL compliance

Posted May 20, 2018 18:36 UTC (Sun) by dirtyepic (guest, #30178) [Link]

The preferred format for source code for modification is text in a computer file, as opposed to several truckloads of hardcopy or runes carved into the side of a mountain.


Copyright © 2025, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds