Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×

Comment Re: Yes (Score 1) 716

For me systemd was a non-issue until i ran into the chain of replacement for consolekit (a beast all its own, but at least unix modular) and found that i would have to rip out the current init and replace it with systemd, because logind insisted on it, and logind was replacing consolekit across the board.

Comment Re: Yes (Score 1) 716

Bingo, if bash is a issue there is quite a few replacements floating around. This because the interfaces etc have been solid for decades.

Similarly, the Linux kernel will keep userspace facing interfaces in place even if new ones have been introduced to replace them. This to allow people to update at their own pace.

Systemd claims to be a collection of tools, but the interfaces between those tools have no stability guarantee. Thus they are more like the interfaces internal to the Linux kernel and may as well be considered a solid blob.

Comment Re:Yes (Score 1) 716

If by hide you mean replace everything and the kitchen sink with their own implementation, complete with doing the same mistakes that the older alternatives had hammered out of them a decade ago, then yes.

Their DNS "client" can be cache poisoned, their DHCP "client" ignore security checks for the sake of speed, and this is a project that is being pushed heavily towards cloud services.

Comment Re:I don't think that means what you think it mean (Score 1) 83

Sounds like the problem that has haunted overly "smart" user interfaces since day one, as their smarts invariably fail to account for all the variables and thus fail exactly when the user is at the most irritable (hello Clippy).

To me a UI works better when held static rather than trying to second guess the user. Then the user applies their "smarts" to integrate the UI into their tasks.

Comment Re:Bit of a hatchet job (Score 1) 551

The seems to be worrying about a bait and switch scenario, or a embrace, extend, extinguish play.

RMS is in it for the long view, that some critical issue can't be resolved because it needs something that has been lost years before thanks to some economic entity going tits up.

What set him on his path was the university getting a new printer thanks to a clueless department head accepting a "good" deal from a vendor. With the old printer RMS could access the source of the drivers, and had added a small bit of code that would put a message to whoever was doing a printing when the thing ran out of paper. The new printer driver only came in binary blob form.

Similarly, there are university labs out there using rickety old 486s etc to run their test rigs. This because a vital sensor driver can't work on newer hardware, and the supplier has long since caved in or discontinued the product. And if they want to replace the sensor they will have to run a long list of basic experiments to make sure the old and new results line up. And that will set the lab back perhaps a year.

Comment Re:I don't think this [release] matters at all... (Score 2) 193

Supposedly the consolekit support code is still there, but needs a maintainer.

Not helping tough that the person that put in the logind code is the same guy that maintains systemd as a whole, and used to maintain consolekit. And who very loudly declared consolekit dead and buried on the mailing list (to the point that if Canonical wanted to continue maintaining consolekit, they needed to find a new name and set up a new repo), and then finally shut down the same mailing list under the pretext of spam.

Yep, Poettering...

Slashdot Top Deals

Digital circuits are made from analog parts. -- Don Vonada

Working...