You are not logged in.
Not sure how I missed the patch file.
You will need clean kernel sources and attempt a manual edit of the changes.
I suggest you search the web. There are numerous articles on compiling a kernel out there.
I don't think compiling is going to be the main concern however. You have some source code changes from 6 years ago. You'd need to find out what was changed and apply that to newer kernel sources and recompile.
Looks like a few modified files from 6 years ago, rather than a patch file.
You would need to create your own patch from those files - e.g. using diff(1), then apply the patch to kernel source using the patch(1) command, then compile the kernel.
As the patched files are 6 years old, there's no guarantee any of this would work with a modern kernel however.
It such situations, boot single user mode and attempt to clear the package cache first:
# apt-get cleanThat will almost always free up enough space for you to boot multi user again and do a proper clean up, of deleted files, old downloads, etc.
Your thread title is misleading.
Your manual changes are being overwriiten, because the file is overwritten. Previously this would not have been an issue. The solution is in post #3
Dolphin is only displaying file system metadata, much like you would see with:
$ ls -lcp(1), tar(1) and rsync(1) all support arguments to preserve those attributes.
I doubt it aill ever materialise and if it does, it will be a reinvention born out of ideology and not much else.
Alpine related and on the subject of reinvention:
No one saw that coming...
This is one of the great perils of rewriting everything: You open up holes that no one could have foreseen.
26 year old article on why rewriting is a bad idea:
https://www.joelonsoftware.com/2000/04/ … do-part-i/
More recent:
https://smarden.org/runit/benefits
It all boils down to service supervision and fast boot/shut downs.
There's also portability and small footprint (vs systemd).
For most desktop usage it's just not needed. But as Altoid said, if it's already the default init used by the distribution theres no point changing it.
There's nothing "political" in the design of mainframes and dumb terminals. Computers were expensive and large back then and it would make no sense for every user to have their own mainframe.
The modern trends towards cloud and SAAS is more about profit. People, especially companies and bosses, will fall for it just as they are falling for the "AI" hype.
I'll give it a miss.
Certainly after skim reading through this load of bollocks:
6 years on, and nearly 7 years after the announcement, where is this OS? I still see nothing much except a roadmap and some kind of release cycle plan and talks of replacing code due to licensing... not much else.
In the meantime, another similar project has been announced:
https://www.theregister.com/os-platform … os/5306589
This time based on NetBSD, with the goal being to avoid LLM generated code.
As a BSD user, I have to admit that I don't see the point in these "vanity" projects. NetBSD in particular could do with more contributors, not pointless derivatives.
What display manager are you using? Whichever it is, stop it, run startx as your normal user and examine /var/log/Xorg.0.log for any warnings/errors.
Interesting approach / hack around. I did some research out of interest but couldn't find anything that useful. There was some indication that a newer kernel might improve things.
If the KVM itself is hub powered rather than self powered, then there's a potential problem.
With basic USB KVMs, when you switch it over, it momentarily disconnects, then connects to the other device and proceeds to draw power from that.
Wih those types of devices, neither of the two connected host PCs will detect the KVM, they will simply register the keyboard and mouse connection/disconnection.
If it's only problematic on the one device, that points to either a cable problem or power problem.
As the problem goes away with the KVM out of the picture, you should switch the cables around and see if the problem moves to the other PC - if it does then ypu've found the problem. If it remains with the same computer, despite switching cables, then it's clearly specific to that PC and likely power related.
@coder-hase: This is not a simple "which is worse - Red Hat or Canonical, pick a side" discussion.
Canonical didn't exactly "game the GR".
As a large number of their employees were also involved in Debian at the time, along with other interested individuals on the corporate payroll, then you don't need to put much mental effort into predicting where that's going. Despite the GR result, Debian would have gone down the systemd route anyway. Subsequent events made that all too clear. The whole process was simply constructed at the time to ensure that the end result was a vote for systemd. You don't need to take my word for it, there are historic threads on FDN covering this and numerous articles elsewhere. This is partially why we are now here, posting on a Devuan forum.
And now Canonical is in bed with Microsoft, who also employ the main systemd developer / project leader. Microsoft/Canonical are also busy rewriting GNU coreutils in rust... under permissive licensing of course.
You can even install the .NET, MS C# encumbered, shitware in your Ubuntu:
https://learn.microsoft.com/en-us/dotne … x-snap-sdk
The point is that the IBM / Red Hat connection was once unthinkable. It's not the only puppet / puppeteer scenario. The Linux Foundation should illustrate this all too clearly.
You asked "why canonical?".
The reason I listed them among the sell outs is because that's what he did on multiple occasions. It's also clear that Debian's move to systrmd was steered and influenced by Canonical people. As soon as the GR result was known, they wasted no time in firing the upstart developer and switching Ubuntu to systemd.
Red Hat were an enterprise solutions provider even long before the IBM buy out and the main driving force behind systemd. So I don't see the logic in listing them with the others. A bit like expecting Satan to put out the fires.
I have no idea why the OS of your laptop is relevant to this thread?
I have not heard of Canonical's "proprietary" appimage format.
As far as I know Canonical developed snap (I have never used). That has a proprietary backend I believe (not the first time theyve pulled this kind of thing).
@coder-hase: Canonical is a Microsoft partner.
Solid advice. I'm not sure why any distribution would default to ssh root login nowadays?
I seem to recall that Slackware was one of the last distributions to allow it by default. There was some considerable uproar just before it may have been finally disabled, but perhaps it never was? The developer's stance was that it was there for ease of setup and up to the user to configure it and secure it. The online Slackware documentation covers securing ssh.
To state that sysvinit is "hobbyist" is absurd, but also typical of the propaganda spread by systemd project, along with its backers and advocates. It is easily refuted, with 5 minutes of research into AT&T UNIX System V, which certainly wasn't developed as a hobby. The detractors and parrots never let the facts get in the way however, and there were numerous self professed experts claiming that sysvinit was unusable. It does make one wonder how Linux got off the ground at all.
I'm sure that runit and OpenRC and a few more work very well.
Boot times may be issues for some, but I was never that concerned. If someone is going to go the lengths of changing init system to save 10 - 20 seconds on boot, then it's most likely they would be better served by upgrading to an SSD, or M.2 or whatever.
I don't think any init system is all about boot times. systemd did cite faster boot times in various propaganda, but that wasn't its main advantage/purpose.
The people behind systemd, were never concerned about those init replacements - they knew that sysvinit was what they had to undermine and defeat...
This has been the case for well over a decade. The sellouts at the Debian project, Linux Foundation, gnome project, Canonical et al paved the way for this, as did the legions of fuck heads who called it "conspiracy theory" and trolled any individual opposed to the new corporate backed "open source" ethos...
Well here we are...
Commercial LLMs (plagiarising) being trained on FOSS code, is a far greater problem than a few flakey, "unsupported" clones of MS and Adobe products. You're missing the wood for the trees.
MS owns and controls github, one of the largest repositories of software projects on earth. They're not at all concerned about a few attempts at MS Office cloning. MS and others are moving to or have long since moved to the SAAS model. They have everything tied up, which is why businesses aren't running Linux workstations, roughly 20 years after Canonical (now an MS sellout operation) claimed it was possible.
Pointing out when I'm being crass will keep you very busy.
You'll have less time for composing long winded, elaborate posts about fixing stuff that doesn't need fixing.
"...It is written in Rust..."
By LLM.
But that wouldn't be kindly doing the right thing now would it?