TxFS Linux File-System Supports ACID Transactions, Simple API

[starshipeleven]

Well, they are violating the GPL if they dont release it under GPL.

This is the author's own issue.

The author needs to state his work has some kind of license. If he does not do so then it's still 100% "all rights reserved" copyright. That's how copyright law works in most places (like the USA).

If you take this code and try to upstream it, then upstream will laugh you out of the mailing list due to this licensing issue (i.e. you took code that isn't using a license so it is violating GPL, and you can't change license of it as it's not your work).

It's wildly unlikely that the author himself would actually prosecute anyone using his source, and also it's highly unlikely that FSF or similar will start a legal battle over this tiny proof-of-concept filesystem GPL violation, but it's still technically a violation.

The author isn't stupid. He probably wants to get paid to develop it further, or something, while not allowing people to take the code and run, so to speak.

I'll allow that, it's fine.

[carewolf]

This is the author's own issue.

The author needs to state his work has some kind of license. If he does not do so then it's still 100% "all rights reserved" copyright. That's how copyright law works in most places (like the USA).

Yes, my point is just there an implied license since it is a GPL-derived product, and the Linux kernel + drivers are documented as being GPL everywhere something else isn't explicitly declared. So that should be a safe assumption

But yes, I will give you that I probably wouldn't want to use it in something serious until a specific license is declared by the authors, there have been cases of incompatible modules before that you couldn't use without violating the GPL, but assuming the author isn't Oracle, I am just saying the license will as a minimum be GPL compatible, so asking if it is GPL is kind of a weird question, a more relevant would be: Would it be BSD/MIT, could it be used by the BSDs?

[F.Ultra]

Well, they are violating the GPL if they dont release it under GPL. It has nothing to do with mainline or not, but you are not allowed to publish a new development on GPL code without the new changed code also being GPL. I am assuming they are not violating the GPL, thus it is under GPL or something GPL compatible.

In any case, you can 100% assume it to be GPL, since it has to be able to be released as GPL (as a derived product of the Linux kernel).

Only if they do binary distribution of both the Linux Kernel and this driver combined, but I can only see source code distribution so they can license this in any anti-GPL fashion they like.

Even when his code is distributed as a patched version of the 3.18.22 version of the otherwise GPL Linux Kernel as is the case here he is still not obliged under the GPL what so ever.

[dwagner]

This having been released as a patch to an ancient kernel version stirs up some memories of "Coda" in me - which was a another file system from an academic ivory tower that never reached practical relevance, despite addressing an actual gap with its features.

[timofonic]

Does this filesystem comes with guns and nachos included?

[jo-erlend]

Not true at all. A person does not give up his copyright to his creation until he explicitly assigns a license. The code would likely never be picked up in mainline until the GPL is applied but that is a different issue.

Using GPL does not imply giving up copyright. No license means the software is completely proprietary since it is automatically copyrighted.

[fuzz]

Would this filesystem be useful for the likes of postgresql etc, or is that way off the mark?