| |||||||||||||
What's the complaint?What's the complaint?Posted Mar 1, 2011 22:07 UTC (Tue) by dskoll (subscriber, #1630)In reply to: What's the complaint? by AndreE Parent article: Red Hat's "obfuscated" kernel source
The GPL allows you to redistribute derived works under the GPL. Red Hat claims that doing that is a breach of their subscription agreement.
(Log in to post comments)
What's the complaint? Posted Mar 2, 2011 5:44 UTC (Wed) by AndreE (subscriber, #60148) [Link]
Yes and what does the GPL say about Red Hat's license agreement?
Nothing.
You are still free to distribute the code under GPL, modify it under the terms of the GPL, and utilize the precise freedoms the GPL gives you.
If Red Hat said that you could only distribute the code if you were a subscriber, then THAT would be an additional condition. But that is not what they are saying at all.
What's the complaint? Posted Mar 2, 2011 5:49 UTC (Wed) by dlang (✭ supporter ✭, #313) [Link]
no, they are saying that they will only show you the code if you are a subscriber, and that if you want to be a subscriber you are not allowed to distribute the code.
that sounds very similar to the microsoft 'viewable' source the difference being that microsoft can sue you for damages while RedHat can only stop supplying you with updates and support, potentially shutting your business down while you transition to a different distro
What's the complaint? Posted Mar 2, 2011 5:59 UTC (Wed) by AndreE (subscriber, #60148) [Link]
Wait, if I have a subscription, obtain GPL binaries from Red Hat, then cancel my sub, I can no longer access the source to my GPL binaries?
What's the complaint? Posted Mar 2, 2011 9:12 UTC (Wed) by airlied (subscriber, #9104) [Link]
they are on the ftp site.
all the source code + all the files needed to rebuild are in the srpms.
What's the complaint? Posted Mar 3, 2011 16:43 UTC (Thu) by smurf (subscriber, #17840) [Link]
Yeah, but they're not in source form.
"Source", in this context, is "a kernel and a heap of patches, as embodied in a git archive". Granted that if you drop the archive and flatten the patches into 200+ patch files, that still counts, because people still customarily edit patches.
A humungous diff, however, is not a patch. It's the source equivalent of a core dump. Nobody uses that for software development.
Therefore, Red Hat is breaking the spirit of the GPL by doing that.
What's the complaint? Posted Mar 3, 2011 22:29 UTC (Thu) by airlied (subscriber, #9104) [Link]
You have a definition of source that suits your argument.
I'm sure lawyers can define Source == source code, in that context you have the source code to the kernel.
Its easy to make up definitions and then base arguments on them, its harder to prove the definition you made up is backed by the law.
What's the complaint? Posted Mar 4, 2011 7:56 UTC (Fri) by smurf (subscriber, #17840) [Link]
The GPL says quite plainly:
The source code for a work means the preferred form of the work for
Whether combining lots of patches into one large patch constitutes a change in "form" would be an argument for the lawyers if this ever goes to court, which I very much doubt.
In my opinion, however, the case is pretty clear -- you simply do not modify that large a patch file. Therefore it's not a "preferred form". Therefore Red Hat is, ideally if not materially, in breach of the GPL.
Disclaimer: I am not a lawyer, not do I play one on TV.
What's the complaint? Posted Mar 4, 2011 17:28 UTC (Fri) by dlang (✭ supporter ✭, #313) [Link]
you don't normally modify patches at all, normally you modify the source file. the tools then mechanically create the patch files.
What's the complaint? Posted Mar 7, 2011 22:26 UTC (Mon) by paulj (subscriber, #341) [Link]
Forget the details of patches for a second. What matters is, what is the distributor's preferred form for making modifications? If there is a difference between the form they prefer internally and the form they distribute onward, then we have to ask whether the recipients have been short-changed out of their rights. Precisely how much information is a distributor allowed to obfuscate before there is a problem?
E.g. if the patches are of no import to making modifications to a source code, then why have RedHat decided to try get a competitive advantage by withholding them? Clearly RedHat feel having the split-out patches helps them to maintain and modify the kernel they ship. My experience is that having patches (more precisely, access to the history) can be *very* important to making further modifications (finding recently introduced bugs particularly, and modifying them).
I know RedHat is "Good", I know they put in lots of resources into Linux and free software. I really want them to be able to succeed in their business. However, let's be careful to remain dispassionate about this - do any GPL copyright holders involved really want to concede that it's perfectly fine for distributors to deliberately withhold fairly important source-related information? (Obviously some of those copyright holders also have a strong interest in the continuing success of RedHat).
What's the complaint? Posted Mar 7, 2011 22:41 UTC (Mon) by foom (subscriber, #14868) [Link]
How about requiring that they disclose all the internal email threads where they discussed the change too? After all, those might have substantially more useful information about the reasoning behind the change than just the commit message...
What's the complaint? Posted Mar 7, 2011 23:25 UTC (Mon) by dlang (✭ supporter ✭, #313) [Link]
I am not a big RedHat fan, but I have no legalistic problem with them not distributing patches, internal e-mail threads, recordings of internal conversations etc.
that doesn't mean that I think what they are doing is good in this area, just that it is within the letter of the rules. I think that them making this change erodes their moral position, but the GPL isn't dependant on people making good decisions for moral reasons, it's only dependant on people making the decisions to comply with the letter of the license (or if it requires more than just compliance with the letter of the license, there may be a need for a license change, but I don't believe that there is)
the only piece I have a legalistic problem is with them releasing code in some form to users, but only under the condition that those users don't redistribute the code (and if the users violate this condition, penalties kick in)
What's the complaint? Posted Mar 8, 2011 7:38 UTC (Tue) by paulj (subscriber, #341) [Link]
It's still not clear they're following the letter of the licence. Here's what likely is happening:
1. RedHat work on the source in form A
2. Form B is auto-generated from A
3. Form B is distributed to comply with the GPL.
A is the src.rpm with the split patches: the format they've preferred for donkey's years and, I'm presuming to be a dead certainty, are continuing to use internally, and B is the src.rpm with the patches deliberately collapsed. Talking about email or conversations is a misdirection - binaries never get machine-built from such. The patches *are* a source input to the process that builds the distributed src.rpms though, and the reason they are an input is because that's RedHats' preferred means of making modifications.
Look at the flow above again, form B *clearly* is covered by the GPL through the text in which explains what "preferred form" is meant to cover. It's pretty explicit that intermediate transformations of the sources are *not* sufficient, that *all* the files required for input to the build process are required.
Just because an auto-generated file is still human-readable and editable does not take-away from the fact it's not the original source.
What's the complaint? Posted Mar 2, 2011 12:46 UTC (Wed) by dskoll (subscriber, #1630) [Link] Yes and what does the GPL say about Red Hat's license agreement? Nothing. Maybe. Maybe not. The GPL says you are not allowed to add "additional restrictions" on the redistribution of GPL'd works. Red Hat and others argue that their subscription agreement is not an additional restriction. I argue that it is. I'd like a legal judgement or at least an opinion from the FSF, because it's not clear-cut to me.
What's the complaint? Posted Mar 4, 2011 1:06 UTC (Fri) by jthill (guest, #56558) [Link] By signing that agreement, you did not bind them, and by signing that agreement, they did not bind you. The defining characteristic of a coercive offer is that if you refuse it, you're worse off than you were before they made it. That's not the case here. It's a perfectly good offer. They did not bind you by authority, they did not bind you by coercion, they did not bind you by any means. You bound yourself, for gain, in a completely voluntary and fully informed choice.
|
|||||||||||||