Wouldn't it have been better to have appointed somebody who didn't have to recuse themselves?
And undertake to stick around longer to ensure that their "achievements" would be achievements they'd be prepared to live with in the longer term?
33045 publicly visible posts • joined 16 Jun 2014
Jake, did you read that link I posted to the Debian mailing list. The gist was that it was getting difficult to maintain sysvinit because of all the systemd dependencies creeping into upstream userland stuff. That makes me worry about the maintainability of Slackware as well because surely they must be either keeping to old versions of userland or chasing the same issues.
"the issue is that this really gives no real comprehension of process state"
Exactly. Because it's all hidden in the great morass of systemd.
Anyone with the basic shell skills needed to administer a Unix-like system can develop the script st the terminal before deploying it.
And your belief about searching the log database or logging to a remote server doesn't get you very far in trying to sort out a non-booting box without being able to make sense of whatever logs it managed to write. Even the remote logging depends on the box being able to get itself as far as being onlne.
"I use devuan for a reason."
Me too but I'm getting worried as to how long it will be able to survive as this garbage gets further and further into the upstream. I came across this https://lists.debian.org/debian-devel-announce/2019/09/msg00001.html from the current Debian project leader. Scroll down to "Init System Diversity". It's not very cheering.
"It'd be nice to leave linux design to Linus' group, methinks. It's not like they've done a bad job so far."
Yes, but as I'm sure you know Linus & Co are only concerned with the kernel. The rest of a Linux distro involves stuff from various Unix implementations, either directly or reimplemented by FSF or others.*
But the overall composition of a distro is designed by the distro's own maintenance team. We're told they like systemd because it's easier for them and if that's a problem for users it's just a problem for users. This is, to my mind a weakness of FOSS; it's purely developer lead. The theory is that if you don't like it you can just fork it.
But that's easier said than done, especially if your focus is on using it as a tool to do your every day job. What's worse it seems as if systemd and its dependencies have wormed their way into so much it might become impossible to do a simple systemd-less fork. On the one hand it seems to be taking Devuan a long time to get their Buster based version out (although Knoppix seems to have managed) and on the other I came across a Debian status email which floated the idea that they gave up attempting to provide a theoretically possible sysv option.
I think I might be coming to the end of the line with Linux.
* Hence Stallman's insistence on calling it GNU/Linux although this ignored many other contributions.
Perhaps you need to pay attention. It's likely to be shoehorned in within the next couple of years or so. Maybe sooner than that, who cares if it's too buggy for release? It's probably intended to introduce enough dependencies into regular prorams to make the likes of Devuan and Slackware finally impossible. It's going to be BSD or nothing.
"The [redacted to please Cloudflare who block anything that looks like a path] passwd database is not extensible, and therefore Linux has evolved numerous secondary databases that are stored elsewhere, such as [redacted to please Cloudflare] shadow, a privileged location used for encrypted password hashes and other password-related fields, such as the maximum time before a password expires."
Linux has evolved no such thing. It's simply inherited it from Unix
It was a necessary step for Unix to evolve in that way after the sort of incident described in "The Cuckoo's Egg". The password file has to be world readable because user programs such as ls and chown need to be able to map UIDs to user names and is small enough to be exfiltrated even over a dial-up modem link. Once desktop processing power became sufficient to crack the encryption then in use in passwd the actual passwords needed to be moved into a separate file which could be privileged because only a limited number of system programs needed access.
We now have a lot of fussing about state and configuration to satisfy some arbitrary scheme about directory usage. Stuff that. Unix directory usage, like the rest of the system, was designed on practical grounds. We're seeing the steady destruction of a working, practical system design to satisfy the ego of a Jonny-come-lately. If he wants to design a system to his own notions let him go ahead and do that from scratch and get out of everybody else's hair.
I booked tickets for a local event. The email tells me I do not need to print this email just make a note of the booking reference. It's an alphameric string of 17 characters. Yes, I'm going to remember that when I turn up at the door aren't I? And the door staff are really going to check that in a list of 100 or so 17 character numbers.
What are this lot on? Gotit! Green Koolaide. Because it's "more than 400 times better for the environment" if I don't print it. Eejits.
"Worried about spam? Then sign up using an email address that you only use for receipts."
And watch it collect more and more spam. The real way to do this is to set up a number of addresses before you go shopping. Hand the next one out to each shop. Get the receipt. Kill the address. In practical terms I already have more addresses than I'd like to have just for the firms I regularly do business with.
In order to not be challenged it would need to be digitally signed otherwise you could be accused of editing it. You would also need to be able to verify the signature. You would need all that before you leave the counter. It's hardly a speedy way to achieve throughput at the store.
"At least she's only wasted £600K. So far."
Not yet. It's only been talked about. It'll be announced several more times. Maybe along the lines of £400k followed by an announcement of it being topped up by £200k. But "We're hoping to deliver the funding by 15 March 2020,". Don't plan on getting any of it.
And el Reg, if you're going to add "Re" onto a title, use different maximum title lengths for the two different variants. A little corner case to take into account. I doubt it would take £600k to sort out.
It's part of our long established constitutional arrangements that courts do rule on matters not covered by Statute Law. It's called Common Law. It's worked well for centuries. Apart from anything else it's enabled courts to adapt rulings to a changing world whilst Statute Law requires Parliament finding time to update or repeal obsolete laws.
"There is a danger in this that once this precedent has been set, that it will be cited for some other prorogation in future"
The judgement said it was a one-off. I think this translates as "nobody's been daft enough to try on this before and we don't think anyone will be daft enough in the future". I hope they're right.
Upvoted for the "Not really"
But the Supreme Court is just the original Law Lords under a new name and with new premises. It's just a consequence of Blair thinking we should have something called a "Supreme Court". The whole HoL didn't hear cases, just the Law Lords and the members of the Supreme Court are just the latest holders of those posts.
It's rulings may be "unprecedented" unless they find an existing precedent to follow but that's the norm for courts dealing with appeals.