Well, if one reads this ebuild script for ChromeOS, it clearly states that bash is only used for building ChromeOS, but dash is embedded into the built image.

Thanks again for proving my statement, "Google avoids GPLv3 like hell, if the target audience is 'big licensees'".
But you already did that before, IMO it did not need another proof.
And ChromeOS was just an example how Google handles GPLv3 software in order to understand Jolla's concerns and strategy better (which is achieved now).

WRT the GPLv3 itself:
It is only the license text which is legally binding, last but not least because this is usually the only legal text referenced by or supplied with GPLv3 software.
Demanding references to statements of the GPLv3 not to point at paragraphs within the GPLv3 does not make much sense (even if the "demanding" aspect is fully ignored).
Especially as you completely fail to reference any of your claims with something within the GPLv3 license text!

No, it doesn't say that anywhere. The comment just says that they use dash as the main shell, exactly like Debian. But they still ship bash, and in fact depend on it for running scripts which specify it on the shebang. Again exactly like Debian. In fact, RDEPEND means runtime dependency, not build-time dependency.

More evidence?

• Go to chromeos://os-credits and grep for Bash, developer mode or not.
• Kill your Wi-Fi (so that you can't claim that enabling developer mode installs something; you just need developer mode to get a allow terminal access), enable developer mode, open crosh, and let me know what happens when you type "bash".
• Download _any_ image of ChromeOS and tell me what you see in the system partition, /bin/bash

Other GPL-3 packages that are preinstalled on my pixel go:
binutils, coreutils, dosfstools, exfat-utils, fuse-exfat, gdbm, glmark, gmp, gzip, mesa, mtools, readline, pycairo, rsync, rsyslog, samba (of course), .... and just way too many to count.

Well, I still don't understand the concern. All the companies in the world can do it, even Microsoft and Google, can do it, so why can't Jolla?

Jolla, who _used_ to be one of the "most open" mobile devices around, and has disappointed me so much I've had to _publicly apologize_ about it?

I quoted the authors of the license directly contradicting your statement. But this is apparently not enough.

On the other hand, you are just blanket-quoting the entire license and claiming that it satisfies your point of view.

I disagree and literally point how the authors of the license as well as all the largest software companies in the world disagree with your point of view.

Yet you again claim that I'm not providing sources while you have literally again not provided any single source for any of your FUD which as you're basically admitting you have none for. Not even for the "Dutch museum story" that I'm particularly curious about.

Double-standard?

No, it was a single paragraph of the preamble, plus pointing to the relevant terms (#2, #3 and #6) implementing the statements from that paragraph.

Yes, these do fully "satisfy my point of view", if one reads them.
Specifically the quoted paragraph in the preamble clearly and abstractly depicts the goals of the GPLv3, which make its use problematic for quite some use cases (as discussed).

Yet you fail to point to something in the GPLv3 which "satisfies any point of your view".

Yes, aka 25% of the total number of words in the license. Totally not blanket-quoting.

No, not "as discussed". You claimed that if a museum lends you a tablet with GPLv3 software it is "problematic". I disagree and the GPL FAQ disagrees.
Please don't bring in every other usecase in the world here.

Nice, that we finally seem to agree on these three points:

and


That is fine for me.

I am glad that we were at least entertaining for others.
True, this was definitely more heated recently than the usual postings at the "To amuse the community" thread.
Well, partially I was even able to have some fun along these lines, but overall it was just tedious for me.

Back to the real topic:

I denoted in the original posting of this thread Does that answer your question?

And for the record, after asking around, the real reason Jolla is stuck with ancient Qt is because the QtWayland compositor module "graduated" as a non-essential Qt addon and thus changed license from a mixture of BSD+LGPL to GPL only. Only essential Qt components (like the client parts of the Wayland module) remain as LGPL.

Since Silica is apparently still not fully open, this puts them in a problem as the compositor process is linking Silica and Qt Wayland Compositor together. (Plus a lot of porting to do, anyway).

The issue is not related to GPLv3 at all. In fact Jolla is already shipping some GPLv3 software in the device images (at least for the original Jolla). E.g. I've just checked and GPLv3 readline is used, for example. So much for the GPLv3 FUD...

@javispedro,

Thank you for independently confirming what @rinigus concluded last November:

and


Combined with Jolla's clear statement,

... that they will not "open source" Silica (and will remove the few references to the LGPL in the Silica sources), a technical way out might be to let Silica not depend on Qt Wayland Compositor, anymore. This may be achieved by some interim layer, be it a new one or by extending an existing one, which is FLOSS.
Or Jolla pays a lot for a commercial license (while "The Qt Co." has a track history of raising the prices regularly), if they have not negotiated some very special conditions (unlikely due to the size difference of the companies and Qt Co's past behaviour).


Also thank you for your constructive contribution.
Hence I take back the
... because above posting has some substance.

I just assume that you seem to be a very strongly opinionated *GPLv3 and / or FSF fan for now.
Still I want to point out to you on an abstract level, that you exhibited quite some trolling properties, aside of strong fandom:

• Language: "crazy", "insane" etc. (Thanks!)
• Grasping at every string, be it a side note, example, etc., while missing (to address) the main points.
• Deliberately running down every rathole in sight.
  Like the "Dutch museum tablet", which was merely a simple example for a "transfer of the right of use and possession" (i.e., "loaned" / "borrowed" colloquially), which is incompatible with the *GPLv3 family of licenses (without granting all users full device control).
• Demanding answers to questions from you, which are not relevant for the topic discussed.
• Not really reading or ignoring what others post or link to, when it does not fit into your view.
• Ignoring all facts, which contradict your view (e.g., the GPLv3 license text, Google's license strategy, Jolla's license strategy).

Oh well, see (again) the
So Jolla clearly has "double trouble" here:

• Silica cannot depend on a newer Qt Wayland Compositor than v5.6, without conflicting with the license change imposed by The Qt Co. (LGPLv2 -> GPLv3)
• Jolla cannot depend on Qt Wayland Compositor and some other Qt components newer than v5.6, without conflicting with their self-imposed *GPLv3 strategy: A "NoGo" for SFOS components, which may be deployed by default to "big licensees" of SFOS.

Let's keep the namecalling! Show must go on! Show must go oooooooooon....

Why are "my opinions" or "my properties" relevant to this discussion?
Sure, you're not the first one to call me troll (not the first one to retract, either), and I have been famously rash in the past. I usually reserve this for when I'm presented with extremely poor, repetitive, or outright false excuses.

Where have I called you crazy or insane?

I have definitely referred to your interpretation of the FSF as absurd, and even later as insane, specially when I thought it was something Jolla had published.

I still think it is an absurd argument: I don't know of anyone else doing it that way, and it leads to a conclusion that directly contradicts the GPLv3 FAQ itself, in addition to other, more absurd conclusions (e.g. , why would the Affero clause be needed if mere "users" of the work are already entitled to the source?). But it is a fact that at least something is mentioned in the GPLv3 FAQ, so perhaps it is not such an uncommon interpretation. I was curious, at least at the beginning.

I have also called your messages FUD. You are still spurting out FUD in this very message. See below.

The entire discussion starts when I ask about that specific example, because I smelled BS (my very first post on this thread). At the time I even offered an escape route, since _maybe_ there was indeed some Dutch lawyer that had interpreted the GPLv3 in that specific way. It would have made, at least, for an entertaining read. And I would have been forced to shut up (or email that lawyer).

But then it became obvious that you just made up the example. Which basically means you are just spreading FUD about the GPLv3. Trying to muddle what is otherwise a pretty clear definition into your own, in order to confuse the public and instill fear about the license.

And the "Dutch museum tablet" is just one example of this FUD. You prefixed it with "Dutch", as if hinting that it was a story that really happened in this world. That a Dutch museum somewhere actually had a problem with the GPLv3 and hey, "it could happen to you!". But the story has not actually happened, has it? At least nothing appears in a cursory web search and my repeated questions into it have been responded with annoyance.

It is a story you made up, and that should make anyone suspicious, because why would you make up stories that appear to back your interpretation of the license? I can only think of one answer: because you are trying to instill fear, uncertainty and doubt about it. I am not trying to claim malice, though. Perhaps it is genuine ignorance, or just to try to defend Jolla somehow. But Jolla needs no defense.

And you mentioned other examples (such as the GPLv3 being incompatible with ATMs because, according to you, users of the ATM would be entitled to the encryption keys of the machine, apparently), which are even more far-fetched, and that I didn't pursue. Suffice to say, they are similarly absurd.


And, to top it off: despite 2 pages of discussion, you still spurt again exactly the same theory about loaned devices being incompatible with the GPLv3 here in this line, as if it were evidently true, and as if nothing of the past 2 pages of discussion mattered. Never miss an opportunity for FUD, right? Well I will not miss my opportunity to rebuke it for the nth time: It is not true. It is directly contradicted by the GPLv3 FAQ.

Your only rebuke for the FAQ item has been a very cloudy argument which attacks the authors of the GPLv3, claiming they "deliberately" lie about their own license on their own supporting documents (for some unknown purpose). An argument that is again completely unsubstantiated with any source whatsoever, and thus yet another gratuitous attempt at FUD.

Your next argument, I believe, is going to be that "I had a newspaper article about the Dutch museum story, but my dog ate it, so I can't link it to you".

Yes, I demand sources for your claims when they smell like BS.
Life is so easy when I can just spurt whatever I want on a public discussion forum and get annoyed when people call me on it.

Where exactly? Notice that when I rebuke one of your arguments, I always emphasize which one it is. (E.g. I don't know how many more times I will have to repeat "Dutch museum story"). I always mention the argument or even quote the exact sentence. You don't. I am forced to guess which link I apparently ignored.

I am yet to see where is the "fact" which contradicts my view. Actually, I am yet set to see where are the "facts". Facts, facts is exactly what I am asking from you. You claim to understand the license in a specific way that no one else does. You claim that Google doesn't ship GPLv3 software, yet it turns out they do. You claim that Jolla doesn't ship GPLv3 software, yet it turned out they actually do!

And I'm most definitely NOT claiming that the GPLv3 imposes NO extra restriction on top of the GPLv2. It's your list of "blocked usecases" (aka the "Dutch museum story") that I have a problem with. You just quote GPLv3 myth over myth.

And again trying to paint it as a GPLv3 issue for no reason...

E pur, they ship GPLv3 software already, check your /usr/share/licenses/ folder...

EDIT:
Extra disclaimer: I like discussing politics, and effectively like discussing software politics even more. I acknowledge my words are harsh. But I try to criticize the arguments/messages and never the person, and in fact in my mind I will disassociate the author of the posts from the content of the posts themselves (which kind of explains why I'm harsh -- treating you like a generic PR person and forgetting I'm not always dealing with people who are used to my style ). So if anything seems offensive, I didn't really mean it as criticism towards you personally, and hope you are not angry (I'm not).

Yes, Mr. Ballistic.

Guess why?
Hint: I provided you with a list.

Oh, that seems to have been premature.


To reply to your technical points.
Side note: Nice to see how even obvious things are overlooked when one goes ballistic (happens to me, too).

Because using a software online does not include the necessary "transfer of the right of possession" to become a "user" in the sense of GPLv3 and LGPLv3.
To cover exactly this extra case ("online use") is why the AGPL was created.

Oh yes, I was given an Android tablet and wired headsets when purchasing the "multimedia guide" variant of an entrance ticket to a Dutch art museum.
This would have been legally impossible without granting me full control over the device, if then software stack would have been based on *GPLv3 components.
As I wrote, just a simple example, you love to sidetrack into.

Correctly so, as you pointed out, because that does not include the "transfer of the right of possession" of a software in source or binary form, which is necessary to become a "*GPLv3 user".
The "transfer of the right to use" alone is not sufficient.

As discussed, logically the same is true for "ticket machines", "information points" all other fixed device installations.
So I already agreed to and adopted this counterargument of yours.
That is why I remembered that I experienced a case, where also the "transfer of right of possession" happened: the "Dutch museum tablet".


I seriously have to thank you for helping me (WRT GPLv3) to

• sharpen my arguments
• look up proper references (which included rereading the GPLv3 for the Nth time)
• sort out incorrect examples (i.e., fixed device installations)
• use the correct terms for underlying, basic legal constructs: "transfer of the right of possession", "transfer of the right to use" and (not applicable / not necessary for the *GPLv3) "transfer of the ownership"
• etc.