Linux Kernel 2.5.8 Released 17
green pizza writes: "Linux Kernel 2.5.8 has been relased. Major improvements include readahead cache tweaks and support for hot-swap PCI on certain IBM machines." Here's the changelog; if you want to experiment with the bleeding edge, check out the list of mirrors provided by kernel.org.
Helloooooo? (Score:1, Insightful)
Now *that* is important. When the biggest improvements are tweaks and support for hardware on 'certain' machines. Jeez...
Re:Helloooooo? (Score:3, Insightful)
If you're right, noone replies
Re:Helloooooo? (Score:1)
Point being, if kernel development goes back to the speed it had on the 2.1 or 2.3 trees, slashdot will be virtually flooded by kernel announcements. I can see the point if the kernel has a ground-breaking new feature that everyone and his dog has been waiting for, or a fix for some life-threatening filesystem-corrupting bug. But when it's "tweaks" and support for some or other hardware, I find it irrelevant, and redundant. There are other sources more accurate than slashdot if you really feel like staying on that bleeding edge.
It's only a section article, (Score:2)
I like reading other peoples opinions on a topic, particularly if they are knowledgable ones rather than self important whinging about how they personally never wanted to read the story in the first place.....
Your browser probably has a back button. Next time you get to an article you don't want to read try using it and leave other people to make their own decisions for themselves.....
Re:Helloooooo? (Score:1)
I certainly do not think that, for example, hot-swappable PCI is redundant. Once you start aiming for uptimes of several years, the ability to change out and upgrade hardware over time becomes increasingly important.
Of course, there are several ways to achieve a "virtual" uptime of years by clustering several machines, but there are many applications out there where linux-style clustering is not an option. By getting hardware-hot-swap down to the PCI level, Linux is starting to play with the "big boys", and this could be vitally important for the future of Linux, in the same way that Parallel Sysplex [ibm.com] is for IBM. (Yes, hot-swapping PCI is very different from geographically-diverse clusters, but it's a step in a similar direction.)
Re:Helloooooo? (Score:1)
Re:How to Remove Linux and Install Windows XP (Score:2)
One totally fucking HUGE oversight that makes this howto utterly useless. By default, most Linux installs create an extended partition and put several non-DOS logical partitions in it.
I'm not sure if XP's fdisk program has been updated, but they mention DOS5.0's fdisk program as being 'equally' useful in removing Linux partitions and it's not. The usual DOS/Windows fdisk program can remove the LILO boot loader easily enough (fdisk /mbr) and can remove any kind of primary partition, but it CANNOT remove non-DOS logical partitions or non-empty extended partitions. Thus is simply cannot remove the partitions that almost any standard linux install creates. This is a HUGE flaw in the DOS fdisk program. To remove linux (or any other non-DOS logical partitions) you have to use a non-microsoft tool such as norton utilities or Linux's own fdisk program.
Thank you MS for one again spreading misleading information about Linux. Please don't try to help when you don't even understand, and are a large part of, the problem!
I apologise for this totally of-topic rant, but I have to help people several times a week because this so-called help causes them to render their linux system unbootable, and they then have to spend considerable effort getting boot and rescue floppies to get back into linux so they can remove the partitions PROPERLY.
Patch Incorporation Rate (Score:3, Interesting)
Well, these BitKeeper logs are pretty lengthy.
I've always been partial to the multiple one-liner summaries, a la Alan Cox.
I wonder if Linus has always devoured and spit out this many patches for new kernel revisions or whether I'm just seeing an artifact of BitKeeper?
Could it be argued that kernel development is moving faster as a result of Linus adopting some kind of source code management system? I know Grandpa Eric Raymond would be pleased if that were the case:)
[Just hoping that 2.6 comes out faster than 2.4...]
Re:Patch Incorporation Rate (Score:2)
I would love to experiment... (Score:1)