The Redis Rug Pull Is Worse Than You Think
Vložit
- čas přidán 21. 03. 2024
- Redis just rug pulled. Their open source license has been removed in favor of some...much less permissive licensing. It's controversial. We need to talk about it.
Too many sources so I put them in a gist:
gist.github.com/t3dotgg/5032e...
Check out my Twitch, Twitter, Discord more at t3.gg
S/O Ph4se0n3 for the awesome edit 🙏 - Věda a technologie
redis sure did give up on its key values
😭😂
GET OUT
I see what you did here 😂😂😂
Your comment rocks!
😂
don't worry a Rust open source version of Redis is going to rise up called Rudis :)
Simultainously it's the best and worst name possible...
I would have gone with Redat
Chud cuck idk 😶😶
New number, rudis?
Rudeus (greyrat)
Be prepared for AWS to launch OpenRedis in 4-6 days
ElasticRedis*
They were not using redis anyway so what gives. The Redis Stack always had those licenses.
Elasticache exists.
@@anujupadhyay11 Exactly. For years. And it's not redis.
@@anujupadhyay11 Fun fact. Azure is hosting the original Redis and they already pay and contribute even if they could do that for nothing.
This is good. This license change made me realize Redis is slower than all the other open source alternatives
What are some of the OSS alternatives in this space?
can you list some?
@@artrix909keydb
@@artrix909KeyDB
@@MrJfergsKeyDB is still BSD, Microsoft recently launched a new cache-store software Garnet, which is also compatible with existing Redis clients.
This compromise licenses look more and more like businesses trying to have a community of contributors without giving them full freedom, just enough to basically have the benefits of proprietary software and open source software. Kind of like Roblox managed to motivate people to spend a lot of effort into creating all that content just for Roblox to effectively own it all.
Redis cant retroactively change the license of old versions. Everything can still use the version that was bsd licensed and fork it. Still a bit sad to see redis doing this
If I can't fork Redis and change the license to a more restrictive one then I believe neither can RedisLabs
@@phyphorYou are allowed fork and redistribute redis under a more restrictive license, it was using bsd, which isn't copyleft.
What's important is that all old contributions are still licensed under bsd, just not newer ones.
@actualwafflesenjoyer Yeah I was under the impression bsd also had the "in perpetuity and non revocable" parts that other open source licenses have. If that turns out to be that redis can revoke the bsd license for any code they wrote, it certainly would put in to question the validity of bsd as an open source license.
They can't change the license of already transmitted data, but they can retroactively change the license in the repository to apply to the repository and the archived older versions they distribute. In that case you could keep using the one you're already using, but a new deployment would fall under the new license.
@@ronijuppiRevocation is not free. According to copyright law, they would have to compensate all affected users in order to be able to revoke the licence, totalling to billions of dollars.
Someone archive that Redis blog page about them remaining with the BSD license. I wouldn't be surprised if they delete it in the near future
This is why BSD/MIT/etc. licenses are risky.
At any point, the project could just up and decide that your contributions are their property now and cut off the whole community from their open source software. Quite frankly, Qt has never done this, not because they wouldn't want to, but because they've integrated lots of code changes from KDE and already agreed that license changes can be vetoed by KDE project, so they simply cannot do that. For most GPL projects, code contributions are licensed from the contributor under the GPL, so now even the project manager is prohibited from changing it.
GPL projects will almost never suddenly change their license from underneath their community, unless they are the only ones who wrote everything.
Regarding Qt, there is actually an agreement between the KDE free Qt foundation and the Qt company. If Qt ever stops developing their open source licensed version of the Qt Framework, the Free Qt foundation has the legal right to release Qt under a BSD style or other OSS license.
Well, we are free the fork the versions prior. Feel reasonable
theres a reason why BSD and MIT/X is known as the cuck license
You can get around that with CLA, specifically Copyright Transfer Agreement. The project is still open-source and available for everyone with GPL license, but to the makers/organization of the project, they own the full copyright, and can change it to commercial and non-free use, whether it is GPL licensed or not.
GPL doesn't prevent the company from changing the license, it only makes it harder, a lot of GPL code-bases have you sign-over licensing rights to them explicitly to do this in the future.
Something that's interesting to note is that Microsoft recently (same day maybe) released Garnet, a redis compatible, open source alternative, that's written in C# so no code share either. MS Garnet is also MIT licensed
Just looking at the page for that and, would you look at that, a red, er, rhombus, extolling the virtues? Very subtle 😅 Much smooth.
You had me until C#
Wow c# cool
@@fishynicky_It's honestly quite neat the performance they are able to reach using C# and it's higher safety guarantees being a GC'ed language.
@@EraYaN They can also do native compilation of the code. If that makes a bigger difference. It is all about how you scale it.
Of all words of tongue and pen, the saddest are these: 'Richard Stallman was right again'.
To be fair, the amount of times he's right is kind of small but not small enough.
Stuff like this makes me more and more warm up to the GPL family of licenses
It really is the only solution.
Unlicense/MIT if you want as many people as possible to use your code with no expected compensation, (A)GPL if you actually want to build an open source community
GPL is incredibly restrictive, I'm a professional developer and I'm not exactly going to contribute to a project i cannot utilize in my own proprietary software, that doesn't compute on a fundamental level. The times I've contributed to gpl projects it has been on employers behalf not personally.
It's simply a fact that we need a new vastly better license that cannot be stolen either by corporations or by digital communists via licensing shenanigans.
Essentially it needs to protect against stallman viruses and corporations being required to publish modifications upstream, while allowing for use as a library instead of the ridiculous bs needing to fire it up as a separate process with huge overheads.
@@RiversJ it depends on the project, assuming Redis was GPL nothing would change until you redistribute it and then you'd have to give out the Redis source code if you modified it. Is that really an issue? Obviously if your business was a Redis fork yeah.
I'd get the argument for libraries but services? Most devs won't be affected
@hecate6834 Forks aren't what I'm worried about. When i create free software, i actually want it to be free such that Nobody gets to steal the work without me getting the changes, repackaging it under more restrictive licensing i can't get the changes for upstream is the real world practical issue with GPLv3
At the same time, it really isn't free if you can't also use it as part of commercial software, I'm doing free to get a name for myself not to espouse my ideological views on economics, if you're doing it for money be honest and ask money for it like the rest of us.
Most of the very best pieces of software on the planet are libraries, not applications which are dime a dozen, were being very restricted in application development precisely because we don't have a good license to share libraries.
Yes i write library code atleast half the time, both abstract and also specific usecase solutions that require a lot of horrible bespoke crap to implement if you don't have deep domain knowledge today.
When you write a piece of procedural feeling niche usecase code has almost certainly been solved by someone far better than you in the domain. But you can't have it even if they wanted to share it because they have no means to share it legally.
@@RiversJ CDDL is right there buddy. You still need to push modifications upstream etc but only on CDDL licensed software, you don't need to license everything it links with under CDDL nor distribute binaries under CDDL (proprietary friendly). Your concerns are why a lot of stuff isn't GPL, particularly in gaming related software. The best part about CDDL is it is incompatible with GPL so no one from the Stallman cult can do anything to CDDL code, but it does mean working around GPL if you are primarily on Linux.
this is one of the main advantages of copyleft: you can’t change the license to something more restrictive later. at least not if people don’t sign CLAs
"...or put on my makeup" definitely got me
It was not a joke though
Came here to say the same thing. 😂
Stage makeup is a thing!
Men can and very often do wear makeup lol. It isn't all eyeshadow and mascara and lipstick 😂
(but also men can also wear these and look sick as hell lmao)
I've been in several plays, so I've worn my share of makeup. And it makes sense that a CZcamsr would do that. But the fact remains... I didn't expect it.
Those who truly value Open Source use a viral license like GPL. It’s very hard to back out of GPL and switch to a less permissive licensing model, as every contributor must be willing to change the licensing of their particular contributions, and for any large active project this is nigh impossible.
GPL is most commonly used by companies like Oracle so that they can make something seem open source but then force companies to buy an alternately licensed version by intentionally making it difficult to use the software without violating the GPL.
@@yahm0n that's completely irrelevant to the original point tbh. sure GPL can be misused, but permissive licenses also leave a community much more vulnerable to embrace, extend, extinguish. a tactic that has been much more harmful to open source than anything oracle has done.
honestly the main issue open source faces today is how difficult it is to do open source while trying to make a living. it's extremely hard to get paid for doing open source, and needing to do extra work on top of a day-job quickly leads to burnout. and this benefits companies so much, it's free labor for them. open source got so big because permissive licenses made it easy for companies to use the labor of open source contributors without needing to do anything in return. and i dont think we can have a sustainable open source community if we cater only to that use-case. copyleft licenses at least make it hard to exploit labor like that.
@@sayaks12your first line makes your comment seem not worth reading tbh 🤣
@@yahm0n sorry for trying ig
@@sayaks12 maybe go the middle way, LGPL?
This was an incredibly detailed breakdown of the Redis rugpull, and really appreciate the time you'd put towards this video.
Isn't it necessary to get approval of all contributors to a project before a license can be changed? I doubt they have done that and all 500 contributors have agreed to that.
This is required for copyleft licenses like the GPL. However, it is not required for permisive licenses like MIT/Apache 2.0/BSD. That is because those licenses allow you to create proprietary forks of the project.
Also, many projects regardless of license choice, require a CLA (contributor license agreement) signed before they accept contributions - basically they own your stuff
Depends on the license. If you move away from GPL, for example, then yes.
We not getting a Planetscale video soon?
He's sponsored by planetscale. They are paying him for staying quiet, and not make a fuzz about it. He's even justifying planetscale's behavior
@@araozuwhat did they even do
Basically a bait and switch. Get devs hooked in with a great open source tool, then when the VCs gets antsy, switch the license to suddenly generate vendor lock in. Thing is Terraform showed what a disastrous strategy this is. It’s not even a hard database to deploy, especially when used as an ephemeral storage or cache layer. I fully expect a forked project to swiftly take over. The sad reality is that the megacorps get to make money off of it for free and I really think open source needs a license that demands contribution to the project if firms of a certain size use it
OpenTofu has released a stable release, and is gaining traction.
This was setteled long ago with the BSD license; otherwise, Apple would have had to jump through hoops when creating Darwin after forking from FreeBSD. If they wanted to keep the code, all they should have to do is adhere to the BSD license and give credit. After that they are free to use it how they please.
Yeah, the BSD allows for proprietary forks, so I don't get what all the fuzz is about. I mean, it is nice when people work on open source, but we are not entitled to it.
I'm curious to see where the source available companies are compared to companies like grafana and element that went from permissive to copy left when facing similar problems
I would argue that this path is the path for any VC backed and public company
These type of company structures makes sure that the incentives for the shareholders will won over the original goals and visions with time
Yup. It's baked in to the capitalist model.
Question, is the old license one of those where you maintain the rights to use that version as you did, and the license only applies to the new version?
Exciting news for the community! valkey is here, a new project developed as a fork of Redis, thanks to the dedication of most of Redis’s top contributors.
Meanwhile Microsoft open-sources Garnet at the same time
Perfect timing, did they know?
The Garnet project officially started the year after antirez "left" Redis so there may have been some rationale there.
However, Garnet is faster and offers Microsoft some added extensibility benefits. Depending on how large their internal install base is for Redis, or cache processes in general, there could be huge benefits to them and their customers from undertaking this project outside the licensing issues.
I wonder if anyone can recommend garnet as redis drop in replacement?
Thanks for the video! It's my understanding that a CLA would only be needed for future contributions. Since past contributions were made under a permissive license Redis has the ability to change the license. The reverse wouldn't be possible though, commits made under a restrictive license couldn't be changed to a permissive license.
Since you don't want to use dark mode, would you consider to use a white background on your video recording set so I can invert the color in the video?
Richard Stallman was right.
what did he say
@@Maniac-007he said “ligma”
"It's actually called GNU/Redis"
ElasticSearch 2.0?
Yea it's weirdly familiar.
Interested as well to hear more on Theo’s thoughts on what happened over at Planetscale.
Interesting thing is that if something were to come from the legal "threats" (entirely justified in my opinion) from contributors, especially those with corporate backing there might be a chance that the fork of Redis will have more funcionality that the official product if those contributors successfully withdraw their contributions from the main project and add them to fork(s).
Can understand redis...the cloud companies just absolutely exploit these open source...
If open source won't be able to sustain then sadly there will be no open source. Just closed source behind the cloud.
I mean how hard can it be for google and aws devs to build up something like redis....matter of months maybe
I was already using keydb and other mit/bsd based KV. but I in contrast want keep using redis in my current projects or even new projects. because I actually love the code base , and I can fork anytime if there is a real rug pull.
If that's not a rug pull, I'm a helicopter.
I wouldn't really call "Dual Source-Availiable" a new term, though that doesn't mean that it isn't stupid.
"Source-Availiable" is already a thing (it's when the source code is availiable, but the use of it is limited, so not open source)
"Dual" just means that it uses two licenses.
So just tell me only one thing, am I able to use redis freely ?? or I have to pay ??
Time to fork it!
KeyDB is a great fork of Redis built by the Snap team
There's a big difference between forking a repo, and building your own community to sustain said forked repo.
classic, the illusion of "hey i can fork it so that solves the problem somehow"
it makes it worse, there are already so many forks to prove that point.
the community of the project is what matters. not that you forked it.
FYI, Valkey (formerly PlaceholderKV) is a direct fork that is BSD licensed and under the Linux Foundation as of yesterday, and Redict is also a direct fork that was relicensed LGPL. Both seem to have strong communities, and it looks like current or former AWS engineers are involved in Valkey and possibly Redict.
So I think it meets the conditions to be a rug-pull.
Also.. Redis Ltd. has a really great product without this license change, they just don't market it very well (they really want you to use modules, but their core competency is Redis Enterprise Cluster, where nobody else even comes close to offering something that works as well). This was unnecessary, and it makes me sad.
I use Redis for almost every project. My message queue and caching is entirely based on Redis and i've even used it for analytics (write-through cache instead of just key-value dump)
and what do you think about this change?
@@Tigregalis i'm abandoning Redis for my personal projects if they go through with this license change. As far as work, i'll speak with my colleagues next week. But if it's up to me, i'd straight up dump Redis based on the moral POV. Neither my company nor me mind paying for products or services but stealing thousands of commits from contributors is just not right
@@spicynoodle7419 RabbitMQ does me very well for messaging, they have v good documentation
@@spicynoodle7419I mean your company could drop redis tomorrow for KeyDB and there would be a positive change with zero downside
@@spicynoodle7419 KeyDB claims to be faster drop-in replacement to Redis.
Whats a good actual open source alternative to Redis?
it will produce short term profits like for MongoDB's case, which will still produce profits for probably another decade due to large companies using it and not wanting to just change a database, it's just a hard thing to do and prepare for while in production. It can turn into a huge, and financial mess if you're not careful.
the ideal thing to do is to make AWS/Jeff to support Redis financially.
it's a win win since Redis will be stable economically and AWS could always got a nice maintenance of a good software
That's the fantasy of open source, imagining that the users that make money from their product will think to care about the project. Even though they rely on it lmao
Can redis-lite be built on top of vitess?
so how much code was written by community?
I agree with you on nearly every point, rug pull included, up until seeing the contribution license takes. On people claiming their copyright under the BSD License and trying to force Redis to stay open: Redis can (likely with a simple script) add notes to their licensing docs and say that all the external contributions are still BSD licensed. Consider that WinSock, the windows sockets library for older Windows OS was entirely closed source and used BSD licensed socket code internally. Those contributions are likely useless on their own, but can be open on their own. I think a "viral" license like the GPL2/3 would work to prevent inclusion of those contributions. But I am not a lawyer, just a long time license geek.
what implication does this have on existing projects that use redis. e.g say i have a service in railway that uses redis ??
As a sidenote on the whole Sentry bit of the article - a Sentry-compatible, MIT-licensed, project exists. It's called GlitchTip and it's actively maintained. It's probably not a fork of Sentry in the strictest sense of the word "fork", but I'm not sure whether the fact that somebody basically saw Sentry and decided to reimplement it in their own new project is better or worse for Sentry here.
"I hate [Mongo's SSPL]"
Well I personally hate people siding with Amazon of all effn companies.
Even a broken clock is right twice a day. I don’t love Amazon, but if they say “killing people is bad”, I have to side with them. If something has a permissive license, others are free to use that thing within the bounds of that license.
20:05 That's a very clever compromise indeed.
Wizards of the Coast tried doing something extremely similar to this with Dungeons and Dragons and the OGL. The community struck back hard against it and they pulled back a bit on the changes.
i remember playing with the 1st release of redis and sharing idea with antirez, that take me back :D
The BSD license, in allowing everything, also allows redistribution under a new license. Past contributors can't force Redis to remove their code, because the license they released their code under allows it. Old versions of Redis containing their contributions remain open-source.
I wonder if this will make ppl still contributing to software using these kinds of licenses to change to licenses which forces the software to stay open-sourced such as the GPL, unless their code is removed.
Msft just open sourced Garnet which competes with redis showing benchmarks that Garnet could be better. Right time I guess 😅
33:30 SSPL was introduced by MongoDB ^^
When you mentioned "i don't know if it's BSL-based" it made me wonder can licenses themselves have a license, can a license restrict the forking of itself?
Considering forks like KeyDB, I wonder what is the benefit of a license change this late in the game?
Sadly keydb dev is pretty slow as focused on needs of Snap Inc 😢
10:10 Are we going to see a fork similar to what happen with teraform?
It makes me question: aren't all contributors of open-source software the owners and therefor a democratic voting-like system should decide if the license may and can be changed? Imagine being a developer with a huge contribution to open-source software, could you hypothetically win a court case if the recognized owner of the software decides to change the license? Edit: just saw that this is spoken about in the video.
Yes, but also no. First of all: In general all contributors must agree when changing the license as long as their contributed code is affected by the license change. But there are several ways to circumvent this issue:
1. Transfer the right to choose a license to the maintainers using contracts. Whether contributing guidelines with statements regarding transfer of ownership qualify as a contract is unclear to me and might vary from country to country or from judge to judge. However, most big FOSS that also serves a commercial market usually has a contributors license agreement that you accept by contributing. It's usually a file on the repo and some checkbox in the PR template.
2. Permissive licenses like MIT and BSD don't require the contributors agreement as they allow for non-FOS redistribution. There might be conditions to that, though, like mandatory copyright notices.
3. Some commercially developed FOSS does not accept contributions from the community as the maintainers understand FOSS more like a way of sharing their work than looking for help. So there actually might be no additional copyright owners.
4. Only relicense the code you own and/or progressively refactor out the code written by others. This is very tome consuming and a lot of work. Also, minor bug fixes (typos, syntax errors, simple guard clauses, ...) might actually not be considered fully qualified contributions in that scenario.
There will be a main fork from the old version
I think the lesson here is: Don't give authority of your open source project to someone who makes money by providing it as a service. It was their own choice to sponsor an open source project instead of creating their own in-house project. Donating money is not a valid justification. They get away with a lot of "explanations" because their company "Redis" used to be called Redis Labs, but since they got the trademark, they can use them interchangably so you have no idea if they're talking about themselves as a company or the software.
Kind of ironic his twitch bot was advertising an Upstash link
This is the year of tokio-miniredis.
I'd be curious on your take on these licensing models and modern cloud infrastructure. When cloud providers charge you for pushing data out of their boundary it seems weird for something like a cache to live on a bespoke platform. I understand their desire but if I'm building a service in AWS and am forced to use something outside of AWS as a cache it seems like I'd pick another option to avoid the cost of my data leaving AWS. To balance that though I feel the cost of the development to shift would likely pay for the change in cost of AWS just paying Redis the fee when they pass it along to me.
So basically elasticsearch all over again, can't wait for Amazon to fork Redis
A few moments later:
- Freedis
- Fredis
I'm currently working on something that I know will be a game changer.
I want to make it fully available to the community, but I also want to be able to earn some money from it.
I've seen the pain involved in choosing a license.
Since I'm building on Rails, I decided to just go with MIT. If I make money fine, if I don't, fine.
it is nice to have a passion project. but that only works as a side project
@@willi1978 I've decided to go open core. All paid features would be real value additions, not simply the framework.
it is okay. If big companies abuse open source there is nothing they can do.
There is a huge misunderstanding that you own an IP with an open source license. All you do is create a flavor of a product but no necessarily an IP and by no means something that is only yours. When it is (truly) open source, it belongs to everyone.
But the previous Redis versions were open source, so the license is version based right?
Like mentioned by others, this is similar to ElasticStack’s move with ElasticSearch. Now, waiting for AWS’s move: ElasticRedis? 😉
7:56 The competition you're referring to is cloud companies selling their technology as-is. As a open source advocate, I think it's completely fine for open source maintainers to prevent corporations from cynically profiting off of the develpers' volunteer labor in their licensing.
Sure it's not 100% free software, but in my opinion it's a worthwhile tradeoff to not have the FOSS community be just a group of free labourers for companies to profit off of, and it could be argued that this more "limited" licensing model is more in line with the open source philosophy.
For open source developers, free software means open technology, iterative improvement, collaboration, and freedom from giant tech companies.
For giant tech companies, free software means software they don't have to pay for, which leads to larger profits.
I think it's completely fine to differentiate between the two.
We're already looking forward for Redis's MariaDB... Sad they decided to kill their company this way.
Well, redis is about to be forked like elasticsearch
Not only jeffing, but I also feel sad that smaller providers like Railway and Upstash , that they are actually also harming the interest Redis Cloud. I think relicensing is unavoidable but just need to make the terms clear. Though frankly speaking, as many ppl would know, redis is not that hard to reimplement, but, it is not related to if it is good or bad to relicense.
that issue about licence changing is so interesting, let's see what is the response
We going back to memcached bois. Let's gooo
time to break out that hibernate l2 cache provider!
A lot of people are getting the law wrong here.
First, copyright is created automatically for anything which does not fall under some very narrow restrictions as to what can be copyrighted.
Second, the copyright automatically goes to the author unless you have a naff clause in your employment contract giving it to your boss, or you are allowed to sign a contributer license agreement and do so.
Third, when you contribute to a project without a contributer license agreement, you retain your copyright, but license the project to distribute your code under the applicable license at the time you contributed. This cannot be changed without your consent.
Fourth, this has been tested in court. In the usa it was found that the author and copyright holder retained copyright, and granted permission to use it under the applicable license. By rejecting the license by trying to change it, you are not complying with the license, and are distributing it without permssion, which is piracy.
In a seperate case, it was found that when the company tried to enforce its copyright, and included code it did not own without an appropriate license grant, they had unclean hands, and therefore were not allowed to enforce their copyright until after they had cleaned up their own act.
This leaves any company not complying with previous licenses with a serious problem unless all contributions are under a copyright licence agreement transfering the copyright to the company, and always has been unless they track down every contributer and get consent for the license chenge for every single contributer. If they cannot get that consent for any reason, then they have to remove the code of that contribute in order to distribute the software nder the new license.
What is REDIS?
For those of you that like me were brought here by youtube algorithm and don't know what redis is, here is a 100sec video: czcams.com/video/G1rOthIU-uo/video.html
Or in text: Redis (REmote DIctionary Server) is an "open-source?", in-memory database that stores data as key-value pairs. It's used as a database, cache, message broker, and streaming engine. Redis is known for its speed, reliability, and performance, which comes from storing data in memory rather than on a disk or solid-state drive (SSD).
Really wish content creators whould just title the video for what it is talking about instead of all this clickbaiting...
the same happened with mongo, that's why now we have dynamo and cosmos db
just started watching this but kinda makes sense imho. Amazon really screwed them over.
Even on an off day Theo looks great
I might be confusing stuff here and there, but if redis is served as service without a cost it should still work fine (i.e. you pay only for the machine running redis and not for SaaS) this falls under using redis and not selling it so allowed
Just when Microsoft announces its Redis compatible Garnet project. :)
inb4 bluwas takes off as a fork. (I know nothing about anything, my sad attempt at humor)
Can't wait to use upstash-radish
I find it interesting claiming that Mongo didn't suffer.
I remember thinking about trying it, and then after the license change, it was never mentioned again and I never thought about it again.
Involvement of Microsoft here could be strategy to attract people having big systems relying on redis, migrate to azure for continuity.
One more reason not to fully trust any project staying available that’s not under AGPL *with external contributors* - 15:05 That they relicensed from AGPL to Apache 2.0 *due to VC money* is a pointer to where the problem comes from.
only half way but really wish you covered how the hell they got approval from all contributors to relicense? You cant jsut take someones code and move it to an incompatible license without their approval so... what happened there?
He finally start to cover it at 25 mins in; do wanna say though "this has not been tested in courts" is ENTIRELY wrong, the license says any code under it must have the license and remain open and shit, that's been established, what redis just did is remove the license from code licensed with it WHICH HAS BEEN PROVEN ILLEGAL ALREADY
@ Except that they are relicensing the entirety of the code base; by your logic I could fork an open source project today and not follow any of the license terms by just declaring that I relicensed all that code for the next version.
The only possible way they could relicense in the way you claim is by tracking which lines of code were committed under what license and any new contributions are under the non open license.
Go sue redis then 😂@@TurtleKwitty
I'd imagine they'll move their own code to the new license. Sure, there'll be some files under BSD but it's not like you can compile Redis with half the sources.
You can, as long as it doesn't have a copyleft license. If the project is MIT/Apache 2.0/BSD its fine to fork it and make it proprietary. If it is GPL you are not allowed. Thats why there are proprietary forks of FreeBSD (macOS) but no for Linux (GPLv2)
Mongo Atlas is Fine….. It’s probably fine to keep using Redis. Sucks for the OSS contributors, but understandable that the company is trying to defend itself like Mongo and elastic against the AWS problem.
Don’t get so worked up about stuff, Theo.
28:00 - they don't need to remove anyone's code. All they need to do is dump all contributor names into a file and say the project contains BSD-licensed code by those people. They can take BSD code and slap whatever other license on top of it. There's no legal venue to stop that. (If you don't like that, don't contribute to projects using permissive licenses).
I really like the FSL idea. I would call it "Delayed Open Source"
Actually the license of ZFS is *more* permissive then the GPL. Both include a clause that requires it be distributed under it's own terms (so they are the same level on that front) but the license ZFS is under is scoped to the individual source file (so it's more permissive from this perspective). The GPL requiring it be applied to the entire binary (less permissive) is the problem.
That said, there is a zfs module produced independently from the kernel that can be used, and is supported by most major distros. There's no question the CDDL permits this, it's more questionable the GPL does, as it's the same category as the old "NVidia blob problem," which, exempting some non-acted on complaints, seems to be generally accepted as okay.
I feel like i want to see a lawsuit.
If you made a new file in the redis project that specifcly had the BSD licence on top i am not seeing how that can be interpeted in any other way than "i am contributing this under bsd"
I am also just fairly angry that a lot of open source code that used redis would be forced into changing...
I've seen lately, in the last 2 years an increase in rails projects that need background jobs use gems that avoid redis or any other external service like redis.
If any old open source tech can just wake up one day and decide “we changed our mind. Pay us.” Why would ANYONE use open source tech for their project?
What is the better alternative? Use free closed-source software until they decide to start charging? At least with open source, you can fork the repository and create a free alternative.
Cause a lot of the time its a monopoly, so your alternative is spending couple of yours and thousands of dollars worth of coding making an alternative.
That said most Open source projects can be forked, so you can just keep using the product from before license was changed. Unless you signed some kind of scam license.
Basically the new license is a paid fork, and someone has to make the free fork. Basically this is same process as the main license owner (and likely main contributor) leaving the project.
That said if you have built your thing on it, you have to pay the license, or start from scratch so its viable business strategy, and basically everyone is doing this rung pull from microsoft to hasbro and Dungeons and Dragons.
BSD license isn't copyleft like GPL and similar; I don't think anybody is going to be able to claw code back. You don't need a law degree to understand how expansive the license grant is on BSD. Mind you they've still made a HUGE mess here with the mixed ownership and licenses. Redis can redistribute BSD code under another license, they just can't make it *stop* being also licensed under BSD by the original author.
so - the issue with use of prior code contributed under BSD is trivially easy to make legal. You can just list the earlier version of redis alongside all your dependcies. Under BSD, derivitave works are completely ok - so the people asking for there code to be removed are not going to get there way. This is one of the few examples of a copy left actually would have protected the community.
for redis this decision is all about cache
czcams.com/video/-8haHgfq2x4/video.html
FSL is really great huh, but if you're concerned with naming maybe "trickle source" would be a good term to coin? Little by little all of the innovation will end up open source