Re:GCC licence change
by Qwertious on Friday March 24, 2017 @12:42AM
(#13946651)
The context is that OpenSSL decided to re-license and sent an email saying "send an email if you object" to everyone. Theo De Raadt understandably objects strongly to this, and is parodying it by declaring intent to relicense GCC in the same way.
Re:GCC licence change
by ajross on Friday March 24, 2017 @01:01AM
(#13946720)
Maybe not, but...
Is it wrong that I really don't care? The OpenSSL license is a historical disaster, and Apache 2.0 is just infinitely preferable. Everyone is up in arms about the principle of the thing, but can't we just agree to fix the license? Does anyone not want this?
Re:GCC licence change
by rossy on Friday March 24, 2017 @01:39AM
(#13946914)
> Does anyone not want this?
Like the parent comment says, Theo De Raadt doesn't want this, nor do any other OpenBSD or LibreSSL devs, because Apache 2.0 licensed software can't be included in OpenBSD. If this license change goes through, LibreSSL will no longer be able to merge patches from OpenSSL. I understand that a lot of people want OpenSSL to be relicensed. I think everyone wants it to be relicensed, but Apache 2.0 is a weird choice of license. Whoever chose Apache 2.0 must not have put much thought into it, otherwise they would have known about the OpenBSD compatibility problem.
It's also worth mentioning that Apache 2.0 is incompatible with the GPLv2. Apache 2.0 code can only be linked to GPLv2 code that includes the option to use the code under the terms of any later version of the license, which means the resulting linked binary must be distributed under the terms of the GPLv3. Proponents of GPLv2 over GPLv3 (such as Linus Torvalds) would also not want this.
Re:GCC licence change
by slimsag on Friday March 24, 2017 @01:05AM
(#13946744)
That's the point: Does anyone _not_ want this?
A group of people can't just get together and decide to 'change' the license of something when they are not the owner of said thing. In this case, contributions to OpenSSL. You're welcome to "want" OpenSSL to be under MyFavoriteLicense but at the end of the day when you use it and lawyers come after you because someone's contribution _wasn't under MyFavoriteLicense_, it's your fault.
Re:GCC licence change
by gizmo686 on Friday March 24, 2017 @01:18AM
(#13946805)
This could be a problem for paranoid downstream users, who do not know which license they are required to comply with.
This could also be a problem if anyone ever tries enforcing the license, as the defendant could argue that it is covered by whichever license you are not claiming to enforce.
Re:GCC licence change
by wwwigham on Friday March 24, 2017 @02:12AM
(#13947039)
Ahhh, the open source community joking about relicensing. Reminds me of that PR on swift right after it was open-sourced[1]. I've been contacted in the past W.R.T. a relicensing effort for a project I'd worked on. Given that I was as fine with the new terms as the old, I gave it the go-ahead[2].
I know people like to care about the integrity of licenses and such, but IMO so long as there's a good faith effort to communicate with past contributors, relicensing seems reasonable (if a pain). Maybe they shouldn't have said "silence is acceptance" in their email, though.
It's unclear how anyone on the openssl team can possibly think this is going to work unless they believe corporate sponsorship provides them with enough clout/money to bully dissenters.
Re:GCC licence change
by ecma on Friday March 24, 2017 @01:39PM
(#13950680)
Pretty uncharitable read of what I meant by "funny" really. Satire can be amusing without invalidating the referenced issue. Don't get so hung up on the issue that you can't laugh about things too.
> ...
> If we do not hear from you, we will assume that you have no objection.
Is "nobody objected" a high enough bar to re license a project? Or am I misunderstanding what's happening?
Also, he's contacting gcc devs by emailing [email protected].
The entire point is that it's NOT a high enough bar.
Is it wrong that I really don't care? The OpenSSL license is a historical disaster, and Apache 2.0 is just infinitely preferable. Everyone is up in arms about the principle of the thing, but can't we just agree to fix the license? Does anyone not want this?
Like the parent comment says, Theo De Raadt doesn't want this, nor do any other OpenBSD or LibreSSL devs, because Apache 2.0 licensed software can't be included in OpenBSD. If this license change goes through, LibreSSL will no longer be able to merge patches from OpenSSL. I understand that a lot of people want OpenSSL to be relicensed. I think everyone wants it to be relicensed, but Apache 2.0 is a weird choice of license. Whoever chose Apache 2.0 must not have put much thought into it, otherwise they would have known about the OpenBSD compatibility problem.
It's also worth mentioning that Apache 2.0 is incompatible with the GPLv2. Apache 2.0 code can only be linked to GPLv2 code that includes the option to use the code under the terms of any later version of the license, which means the resulting linked binary must be distributed under the terms of the GPLv3. Proponents of GPLv2 over GPLv3 (such as Linus Torvalds) would also not want this.
A group of people can't just get together and decide to 'change' the license of something when they are not the owner of said thing. In this case, contributions to OpenSSL. You're welcome to "want" OpenSSL to be under MyFavoriteLicense but at the end of the day when you use it and lawyers come after you because someone's contribution _wasn't under MyFavoriteLicense_, it's your fault.
This could also be a problem if anyone ever tries enforcing the license, as the defendant could argue that it is covered by whichever license you are not claiming to enforce.
Thread for that: https://news.ycombinator.com/item?id=13945492
If you had an agreement with the contributors giving you the right to relicense after publishing a notice, provided there were no objections, sure.
Otherwise...It doesn't seem likely.
I know people like to care about the integrity of licenses and such, but IMO so long as there's a good faith effort to communicate with past contributors, relicensing seems reasonable (if a pain). Maybe they shouldn't have said "silence is acceptance" in their email, though.
[1]https://github.com/apple/swift/pull/17 [2]https://github.com/thrimbor/sbep/issues/1
What? Is this whole letter some form of sarcasm?
(Also, what prevents LibreSSL from replying with "This is great, but can we get dual MIT/Apache 2.0"?)