Slashdot Mirror
Linux 3.10 Merge Windows Closes
hypnosec writes "Linus Torvalds has released the Linux 3.10-rc1 kernel marking the closure of the 3.10 merge window. The Linux 3.10-rc1 is the second biggest rc release in years and the closure of the merge windows means that the features expected out of the Linux 3.9 successor are chalked out. "So this is the biggest -rc1 in the last several years (perhaps ever) at least as far as counting commits go," Linus notes in the release announcement."
So Linux is finally merging with Windows... uh, that can't be right. I guess windows will now merge when you close them? No, that doesn't make sense either, so maybe it's a new Unity feature.
There's always money in the banana stand.
You got an extra s in there, yo.
"Linux 3.10 Merge Windows" had me getting ready to scout out a new OS.
The merge window closing, however, is a good thing.
Ubuntu: If at first you don't succeed, blindly slap a sudo in front of it
They changed the kernel enough so that the NVIDIA binary blobs are börken (sic) again! Something about whether is a real 2.4 kernel and whatnot. It seems just to be a test, but its anal and börken! I understand that NVIDIA is a commercial company and as such cannot keep up to the rate of development of OSS developers.
For folk who even after RTFA wonder whats new in 3.10, the best source is probably LWN
https://lwn.net/Articles/548834/
https://lwn.net/Articles/549477/
Everybody concerned will already have this news from other sources.
You are mistaken.
I care enough to RTFA, and I got the story here.
Why? I care about kernel development because it interests me, but I don't care about it enough to require absoloutely up to the minute coverage. So, slashdot is an excellent place to get it, and there are often useful comments to boot.
Basically, you could say the same about any story: anyone who cares enough could get the news faster from a domain specific source.
Everybody else does likely not care, also because typical users use distro-kernels and not self-compiled kernels from kernel.org.
Huh? These features will make it into distros soonish, and secondly since when is slashdot only the domain of typical users.
And this is not even the kernel release, but the closing of the merge-window,
The two are equivalent from this perspective: the actual release will have no new features.
SJW n. One who posts facts.
Only a day after the last installment! Though this article only has 3 possible candidates, to that article's 4... those candidates are 3/5s of all the words in the headline, to that other article's 4/7.
Everybody concerned will already have this news from other sources.
I'm interested, and I learned of it here.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
I'm happy to see the story. I regularly look at 4 or 5 websites, only 2 of which have anything to do with computing technology, and this is one of them, which I've been following for quite a while. So, although it may not be ideal, I still get most of my technology updates on slashdot. (Other than ones in which I'm professionally interested, the site I spend most time on also deals with technology, but of a different sort: mechanical watches.)
Best wishes,
Bob
Actually I'd argue exact the opposite, this is the earliest possible convenience where it is possible to say when and what features will be coming with reasonable certainty. Post-launch reviews are fun but most of the news are about upcoming products/services/changes and really what you just said - it won't actually be in distros on release day anyway. So you can read about it now, it'll release in two months, be in most distros this fall and in Debian in 2016.... maybe. That said, in the rush to get this story out and get page hits they forgot the part about making a decent summary of what's new and why it should matter, but hey... it would be the right time.
Live today, because you never know what tomorrow brings
Nice that they included some highlights of these features in the summary then instead of saying nothing that the title wouldn't tell you just as well... oh wait.
Are the things I care about - and I suspect most people do too, even if they don't know about it. The eed to transparently (or not!) accelerate spinning drives with SSD is a killer feature. I'm currently running a homebrew NAS on Linux and my VMWare hosts insist on doing sync mounts - effectively killing performance. By shimming some SSD in front of that, my IO latency bottleneck essentially goes away. (Lets leave ZFS out of this). "Desktop" distros will love this too - I see a simple "wizard" that asks "I see you have an SSD installed - would you like to accelerate access to your HD? Click yes and specify a maximum cache size" Presto - an instant increase in performing most tasks.
I want to delete my account but Slashdot doesn't allow it.
Apparently yes, you are the only one.
Am I the only one who thinks having a version number which is subject to getting rounded off is a terrible terrible idea?
"Oops looks like this release has a trailing 0 on there... *delete*."
Terrible idea? That's how version numbers work. They are not ordinary decimal numbers, so you cannot round them like that.
Version numbers are not floating point numbers. They are a set of integers separated by punctuation. If you treat them like anything else, you're doing it wrong.
Try your drainbead logic on IP addresses so we can continue laughing at you.
You're probably not the only one, but that doesn't mean that the set of people who think it's a terrible idea are any less misguided or wrong. This is normal and standard and most tools for tracking versions assume version numbers work this way. Including the tools that do dynamic linking at run-time for you. In fact, for libraries, it's mandatory to do this if you have more than ten backwards-compatible releases in a series. A change in the second number of a library's version indicates that it's backwards-compatible but not forwards-compatible. (Forwards-compatible changes get a change in the third digit; incompatible changes get a change in the first.)
The only people likely to be confused to the degree that you suggest are non-technical idiots in the commercial press, and they'll make illiterate mistakes about all sorts of things no matter what we do. Hundreds of thousands of software projects do this sort of thing regularly (as AC pointed out, linux v. 1.1.10 was released in 1994). It's "three dot ten", not "three point one zero". Version numbers aren't decimals.
Of course, a lot of projects make sure to always list all three numbers (e.g. 3.10.0). Which I admit I prefer. But it's still the start of the 3.10 series.
okay.. how about apple's way:
version 3.x
there. much better.
Do we really have to have this conversation every single time any software releases version_dot_ten? It's getting a bit old at this point.
Ubuntu: If at first you don't succeed, blindly slap a sudo in front of it
and you need a fucking PhD to know what they mean, and have Rain Main's abilities to remember them.
Yeah, cause IPv6 addresses are written, a lot, they're not just allocated by one computer to another.
IPv4 addresses are very much written like that, and they are still the overwhelming majority. IPv6 is only slowly making inroads, and the vast majority of it is auto-configured, so nobody has to remember them so far, at least.
That's what should have been done, and the reason it wasn't was that it was thought that after 3.9, they'd go to 4.0 - if they got to 3.9 at all in the first place. Remember, they went directly from 2.6 to 3.0, and even in version 1, I don't recall them going to 1.10. This new trend in KDE & Linux is pretty annoying - if you need more than 9 minor revisions, then number it exactly like you suggested.
int major = round(3);
int minor = round(10);
printf("%d.%d", major, minor);