@p@lanodan Periodic reminder debian was named after Debby Lynn and Ian who's marriage failed, maybe as a result of debian and it's sexually charged Linux guild parties
@p >Sure seems to support the standard Unix kernel interfaces and run Unix software. >Look at the photo. >It is GNU bash running GNU Compiler Collection to compile and link the software against glibc. Facepalm into the nether realm.
>cal instead of gcal Sad!
>uname name --help; Usage: uname [OPTION]... Print certain system information. ***With no OPTION, same as -s****. -s, --kernel-name print the kernel name ... -o, --operating-system print the operating system --help display this help and exit --version output version information and exit
uname -o; GNU/Linux
uname --version uname (GNU coreutils) 9.5 Packaged by Gentoo (9.5 (p0)) Copyright (C) 2024 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <https://gnu.org/licenses/gpl.html>. This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law.
Written by David MacKenzie.
>When did GPLv2 turn "proprietary"? Linux was proprietary software in 1991, but was released under the GPLv2-ambigious in 1996.
It only took until 1996 before the first proprietary program was added to Linux and Linux became proprietary software again.
> The kernel, Linux was never anything like a Unix
Sure seems to support the standard Unix kernel interfaces and run Unix software. If you ignore the compiler warnings, it even runs a lot of the Unix v7 code.
@p I'd like to point out that, GNU had in the past put a lot of work into ensuring that existing free software on Unix could be compiled on GNU without having to be re-written and of course such compatibility remains.
Such compatibility was implemented long before Linux was even a thing.
Linux does barely anything really - GNU bash passes the elf to glibc's ld-linux-x86-64.so (or ld-hurdxxx.so), which loads the program for execution and all Linux really does is schedule the machines resources (the program is allowed unfettered access to CPU cycles until its scheduled time is over and then Linux goes and halts execution and dumps the current execution state to a stack and runs the scheduled program) and implement SYSCALLs (via glibc, which does a lot of things internally, but does need to call SYSCALLs for things like a read or a write to a file).
Although efficient scheduling and SYSCALLs and drivers are exiting as they are complicated, they are a far cry short from what you need to get an operating system.
@p >If the kernel doesn't respond to syscalls, nothing happens. If GNUboot doesn't init the hardware and then launch the GNU GRUB OS on my computer and then launch GNU Linux-libre and then launch the init and then launch the rest of the GNU and other software, nothing happens.
Of course you can boot Hurd from GRUB instead just fine and the Church of Emacs runs just as well.
>it emulates the interfaces well enough that it runs Unix software from the 1970s Any of such interfaces have been implemented by GNU and not Linux.
In this case, printf() is a function implemented by glibc, not Linux.
>it does not present a Unix interface to the software it runs? If you try to pass a proprietary binary compiled for a Unix to Linux via a loader, Linux won't be able to make any sense of it, as it does not re-implement *any* Unix interface - it implements its own custom interfaces.
The thing that presents an interface to software is GNU, although in certain cases the programmer can decide to interface directly with Linux's non-Unix SYSCALL interface.
>you are trying to make the case that it is not a complete operating system It is a FACT that Linux is only a kernel and is certainly not a complete operating system, but many people don't want the facts to get in the way of their proprietary dreams.
If the kernel doesn't respond to syscalls, nothing happens.
> Although efficient scheduling and SYSCALLs and drivers are exiting as they are complicated, they are a far cry short from what you need to get an operating system.
I said Linux was pretending to be Unix and it is: it emulates the interfaces well enough that it runs Unix software from the 1970s. Are you trying to convince me that it does not present a Unix interface to the software it runs? Because it seems that you are trying to make the case that it is not a complete operating system, and that would be unrelated to what I said.
@p >What makes you think that? Since anyone who isn't a loser uses the best shell.
>Linux supports the necessary Unix interfaces to be called a Unix As noted below, it does not.
Unix™ is a trademark that requires payment and certification.
>Would you prefer I use a different compiler and link against a different libc? That would be quite sad, but that wouldn't be Linux either.
>Doesn't look like anything by that name shipped with v7. Yes, because clearly the free software replacement, GNU cal, came later.
>I wouldn't call it proprietary: the license required it to be distributed gratis and with source. If you cannot sell the software, the license is a nonfree license, as freedoms 2 and 3 are violated; https://www.gnu.org/philosophy/free-sw.en.html#four-freedoms
>This code contains a firmware binary to overwrite the binary firmware that was already in the machine. It is NOT firmware, it is SOFTWARE.
Whether it works as a software patch against instructions in ROM is irrelevant, as it is software and it is proprietary.
>Whether or not this constitutes "linking" is murky, as I understand it. That is irrelevant, as the proprietary software is still in Linux, making Linux proprietary software.
>Of course, Plan 9 is GPL'd. Not anymore.
An old version was GPLv2-only for some parts, but later versions are under different licenses.
>If anyone gives you a copy of the software in binary form and does not supply the source, you can get the source from them, by the terms of the license. The problem is that the Linux developers actually encourage proprietary software developers to write derivative works of Linux and distribute them in binary form only and refuse to provide the source, by never enforcing their license against many of such cases.
Although, when it comes to free software in source form, under a free software license that better enhances the users freedom, under the GPLv3-or-later, suddenly those Linux developers start to enforce their license?
If the point was that Linux supports the necessary Unix interfaces to be called a Unix, then the point is made. Would you prefer I use a different compiler and link against a different libc?
> >cal instead of gcal > Sad!
$ cd src/unix/v7/usr/src/cmd $ ls gcal* /bin/ls: cannot access 'gcal*': No such file or directory $ cd .. $ find -iname '*gcal*' $
Doesn't look like anything by that name shipped with v7.
>uname
> Linux was proprietary software in 1991,
I wouldn't call it proprietary: the license required it to be distributed gratis and with source.
> but was released under the GPLv2-ambigious in 1996.
It was, according to Linus, GPL in late 1991 or early 1992. The release notes for 0.12 ( https://www.kernel.org/pub/linux/kernel/Historic/old-versions/RELNOTES-0.12 ) indicate that unless anyone objected, "Otherwise The GNU copyleft takes effect as of the first of February [1992]." (The last timestamp in the 0. The GPLv2 appears in the COPYING file as of 1.0 but before that, the release notes just say "get a copy of the GNU copyleft at any major ftp-site (if it carries linux, it probably carries a lot of GNU software anyway, and they all contain the copyright)." ( https://www.kernel.org/pub/linux/kernel/Historic/old-versions/RELNOTES-0.95 )
This code contains a firmware binary to overwrite the binary firmware that was already in the machine. Whether or not this constitutes "linking" is murky, as I understand it. Because I like code so I became a programmer rather than a corporate attorney; consequently, I'm afraid I can't help you with that and am bored by it. Of course, Plan 9 is GPL'd.
> When it comes to the license of Linux, generally it is only enforced *against* freedom, rather than in support of it.
It uses a free software license. I believe this is sufficient. If anyone gives you a copy of the software in binary form and does not supply the source, you can get the source from them, by the terms of the license.
@p >But you should use a smaller one I should run a proprietary bootloader that doesn't respect my freedom because it's smaller?
Are you kidding me?
>See linux-0.11.tar.gz's include/unistd.h I am absolutely certain that such internal header file is not compatible with Unix's interfaces, therefore it is nonsense to claim that was "implementing Unix interfaces".
While it was written to be a Unix-kernel-like, it never pretended to be Unix, as a kernel cannot pretend to be an OS, no matter what.
> If GNUboot doesn't init the hardware and then launch the GNU GRUB OS on my computer and then launch GNU Linux-libre and then launch the init and then launch the rest of the GNU and other software, nothing happens.
Yes, the bootloader is important. (But you should use a smaller one.)
> In this case, printf() is a function implemented by glibc, not Linux.
It has always been in libc rather than the kernel.
> If you try to pass a proprietary binary compiled for a Unix to Linux via a loader, Linux won't be able to make any sense of it,
I don't have a PDP-11 to test on.
> as it does not re-implement *any* Unix interface
See linux-0.11.tar.gz's include/unistd.h. From the beginning, it was pretending to be Unix by implementing those interfaces. Some of that has been moved outside the kernel, but as I said, Linux has stopped pretending to be Unix.
@p (In case you were wondering, with the exception of GRUB, every nontrivial bootloader I've looked at has been chock full of proprietary software, as is typical of software not written for the users freedom).
@mischievoustomato Works fine, been using ZFS on laptops on and off since uuhh… 2007.
And I would heavily recommend a COW filesystem on a laptop given how easily you can end up with power loss on those (half of why I don't want to have anything important on ext4 anymore).
@mischievoustomato And like so far the only thing that makes sense that's truly against ZFS but I haven't measured is ZFS being somewhat slower for database writes, which in a way makes sense since you're doing ACID (CoW-ish) on CoW.
But I wonder for what workloads it becomes a problem, at least for Pleroma (which means pretty regular writes) it seems to be fine, even on devboards where the I/Os are horrible.
@p >I do know that if someone hands me binary code, I can demand the source from them. That is sufficient, I think. That is not sufficient, as such business can and often refuses to provide the source code and if no copyright holder is willing to enforce their license for freedom, you don't get source.
> Since anyone who isn't a loser uses the best shell.
I do use the best shell, but Inferno's sh(1) isn't available everywhere.
> Unix™ is a trademark that requires payment and certification.
Unix is an operating system and lately a set of interfaces.
If everyone says the emperor's got fabulous new clothes and you see his dong flapping in the breeze, the correct answer is that the emperor is naked (rejection of false reality despite consensus), not that actually his new clothes are terrible (implicit acceptance of the false reality created by consensus with a rejection of one facet of the consensus).
The legal horseshit matters to businesses: in reality, it's a hallucination. It's magical properties ascribed to the bits in the machine: you can't perceive them and the machine doesn't know they exist (and billions of dollars have gone into DRM systems to attempt to make the machine perceive them). So, just like Saint Terry, I sit down at my computer and do whatever I feel like: I dine on nectar of the Gods.
> It is NOT firmware, it is SOFTWARE.
It is a chunk of code that patches firmware. The patching mechanism is software, the static data has a bunch of hex in it, and those numbers represent parts of a firmware blob.
> Whether it works as a software patch against instructions in ROM is irrelevant, as it is software and it is proprietary.
:terryhacker: "I just wake up and turn on my computer and do whatever I feel like."
> That is irrelevant, as the proprietary software is still in Linux, making Linux proprietary software.
:terryhacker: "I dine on nectar of the Gods."
> Not anymore.
The GPL cannot be revoked. If there were active development under the non-GPL'd branch of the code, then that code would not be GPL, but there is not: the only active branches are under the GPL.
> by never enforcing their license
I don't really know or care what they are doing. I do know that if someone hands me binary code, I can demand the source from them. That is sufficient, I think.
@p >LILO is free software, Unfortunately, what I cloned from `git clone https://salsa.debian.org/joowie-guest/upstream_lilo` appears to be nonfree.
Under src/ I see files that say; /* boot.c - Boot image composition * * Copyright 1992-1998 Werner Almesberger * Copyright 1999-2007 John Coffman * Copyright 2009-2011 Joachim Wiedorn * All rights reserved. * * Licensed under the terms contained in the file 'COPYING' * in the source directory. */
There is no COPYING file in a directory called "source" - so the "All rights reserved" part may apply.
A court may apply common sense and note that the COPYING file in the project root is the intended one, or it may not.
That is a trivial bootloader that doesn't contain absolutely proprietary software for the sole reason being that there wasn't any proprietary software that could be added.
Software under weak licenses are almost always sloppy legally, which means there is no certainty you are granted freedom, while free software under a freedom-defending license that proudly states it is free software is always rock solid legally, so for me it's the functionally and legally superior GRUB2.
> I should run a proprietary bootloader that doesn't respect my freedom because it's smaller?
Who said "proprietary"? I said "smaller". GRUB is a mess. I know tb personally, I like him, but GRUB is a mess. LILO is free software, although if you modify it, your software may not be free if you do not respect the users' freedoms: is the license what is holding you back? Are you saying that would you produce non-free versions?
> The two major categories of free software license are copyleft and non-copyleft. Copyleft licenses such as the GNU GPL insist that modified versions of the program must be free software as well. Non-copyleft licenses do not insist on this. We recommend copyleft, because it protects freedom for all users, but non-copylefted software can still be free software, and useful to the free software community.
If the license does not enforce it, I'll enforce it myself: the version of LILO running on my machine is a version I built from source. It is free; there may be non-free derivatives but I don't use those.
> I am absolutely certain that such internal header file is not compatible with Unix's interfaces, therefore it is nonsense to claim that was "implementing Unix interfaces".
It was not "internal" but was intended for the same purpose as the unistd.h that is now provided by glibc and other C libraries. See the release notes.
> a kernel cannot pretend to be an OS, no matter what.
> There is no COPYING file in a directory called "source"
There is no directory called "source"; there is a "src". The "source directory" would be the directory the source is in. There is a file called "COPYING" in that directory. I don't know how you missed it.
> Software under weak licenses are almost always sloppy legally, which means there is no certainty you are granted freedom,
Here is this again:
> The two major categories of free software license are copyleft and non-copyleft. Copyleft licenses such as the GNU GPL insist that modified versions of the program must be free software as well. Non-copyleft licenses do not insist on this. We recommend copyleft, because it protects freedom for all users, but non-copylefted software can still be free software, and useful to the free software community.
It is free software, it is available under a free license. Someone else may make a non-free version of it, which the license permits. I do not plan to do that with my bootloader.
Okay applied zstd and vdev_zaps_v2 + head_errlog patches from grub savannah (which threw me some 502 in the process but only in IPv4), and now it werks.