Reiser4 May Go For Mainline Inclusion In 2010

[sturmflut]

why should edward stop to work on butterfs?

the extX devs should have never started with ex4 and instead either helped Edward or with butterfs.

People already hit limits in Ext3 (e.g. volumes larger than 16 TB) back in 2006 and everybody knew that btrfs won't be ready in time. Ext4 is there to bridge the gap, and we need it.

It will be possible to transparently convert an Ext3/Ext4 filesystem to btrfs, so what's the problem?

[energyman]

one thing for sure: reiser4 is fucking robust. Had a bunch of hard reboots thanks to some driver testing - and reiser4 survived them all.

it also works around missing barriers - because unlike ext4 reiser4 cares about your data.

[energyman]

why should edward stop to work on butterfs?

the extX devs should have never started with ex4 and instead either helped Edward or with butterfs.

[Jimmy]

Obviously the Bush administration (and the people behind him, weapons industry lobby e.g.) f*cked their own US citizens very tightly. (Surely Obama is not the Messias but I hope for things to improve.)

The way I see it you have one option. Head for Canada. The US government is gonna kill you if you stay. Seriously you can't vote this kind of conspiracy out. People that fucked up and powerful don't just pack up and leave because their leader (Hell bringer Bush and that succubus Laura) left office. You're naive to think that Obama was put into office by some electoral process and not by some diabolical plan at work.

Trust me they are on to you... I'd get the fuck out now. Oh, and stay off the Internet or they'll track you.



Nothing anyone can possibly say will ever change the mind of a conspiracy nut. The best I can hope for is to get a little entertainment out of one by pissing him off (I'm laughing at you).

You know what really happened? Bill Gates had her killed. Seriously. If Linux had such an advanced file system it would shut Windows down today. They didn't want give Riser immortality by making him a martyr so they framed him to discredit him and his work. They then coerced a confession out of him by holding his kids hostage. Follow the money. A few hundred thousand to frame some Joe for murder vs the millions required for patent litigation against Linux. Easy choice. Did you know if you fold a twenty dollar bill just right you can see the chair that was removed from his car? This can't be coincidence.

[Max Spain]

X-D Ah, THAT was cool. And following links from there I discovered the best page in the Universe, where, among many other interesting things, you can find a 9/11 discussion in similar terms.

Good finding, it made up my day.

Welcome to the Internet

[yotambien]

Yeah your right, it's just like the foul thing they did with the Titanic...

X-D Ah, THAT was cool. And following links from there I discovered the best page in the Universe, where, among many other interesting things, you can find a 9/11 discussion in similar terms.

Good finding, it made up my day.

[DuSTman]

Can we please stop talking about Reiser himself and focus on the file system?

I whole-heartedly agree. Computing is a science, and in sciences whether something is true or not depends entirely upon the observed evidence, with the importance of the practitioners deliberately downplayed by the verifiability inherent in the system.

I doubt that Reiser4 will ever be a part of the vanilla kernel. The developers already proved on ReiserFS 3.x that they don't care about maintenance afterwards and critical issues (like the Big Kernel Lock) may only be fixed *years* after.

I follow the Reiser mailing list, and there's been quite a number of kill-the-bkl patches submitted in the past couple of months to Reiser3.

And, when you refer to "the developers", who are you talking about? It rubbed a few people up the wrong way when Namesys switched development resources to reiser4, at the expense of reiser3's maintenance. This was, apparently, done with commercial motivation on the part of Namesys. Namesys doesn't exist any more, and future/current development is being done in the open.

There's the point: any open source project can have its maintainer quit, and Reiser4 is no exception. If the project is important enough then someone will end up taking up the support. This happened with Reiser3, with most new development being done by the Suse guys, and it'd happen with Reiser4 (or any other filesystem) if it gets into sufficiently widespread usage.

Reiser4 may have some nice features, but they also introduce an immense amount of complexity, which is bad. It may not be possible to mount an existing filesystem on a different machine because of non-matching plug-in installations. And I think it is better to implement many things (encryption, compression etc.) in a filesystem-agnostic way (dm-raid, dm-crypt, eCryptFS etc.). I want to be able to change to a different filesystem in the future without losing the features.

Admittedly, yes. Reiser4 is a pretty huge codebase on its own. (70000+ lines if I remember correctly). I've been trying to write a patch to add extended attribute support, but it's a bit bewildering.

More recently, some kernel developers (Andrew Morton) noted that the reason it hasn't been included in the kernel is mostly because no such application has been made in the last few years. (and, not insignificantly, no-one's really keen to do the code review on 70000+ lines)

The bottom line is that Reiser4 is an interesting and unusual filesystem, and I'd really rather not see it fail for non-technical reasons. Unlike the other new filesystems, it's basically done, and has been tested in the wild for years. Compared to developing a new FS like BTRFS from scratch, the effort required to make the "final push" to get Reiser4 in the primetime is minimal. To let this slide now seems like a sad waste.

[sturmflut]

Can we please stop talking about Reiser himself and focus on the file system?

I doubt that Reiser4 will ever be a part of the vanilla kernel. The developers already proved on ReiserFS 3.x that they don't care about maintenance afterwards and critical issues (like the Big Kernel Lock) may only be fixed *years* after.

Reiser4 may have some nice features, but they also introduce an immense amount of complexity, which is bad. It may not be possible to mount an existing filesystem on a different machine because of non-matching plug-in installations. And I think it is better to implement many things (encryption, compression etc.) in a filesystem-agnostic way (dm-raid, dm-crypt, eCryptFS etc.). I want to be able to change to a different filesystem in the future without losing the features.

This has also been the tenor of most kernel developers who denied Reiser4 inclusion over the last six years. Jeff Garzik had a long discussion with Hans Reiser back then (http://kerneltrap.org/node/5330), and none of the patches sent for Reiser4 in the last years adressed any of this.

Some features in Reiser4, like efficient storage of small files and optimized access patterns for hard drives, are present in available Linux filesystems and are even becoming quite unimportant - simple Netbooks come with >30 GB storage nowadays and SSDs do not care about read/write ordering.

[MaestroMaus]

Yeah your right, it's just like the foul thing they did with the Titanic...

[sturmflut]

Can we please stop talking about Reiser himself and focus on the file system?

I doubt that Reiser4 will ever be a part of the vanilla kernel. The developers already proved on ReiserFS 3.x that they don't care about maintenance afterwards and critical issues (like the Big Kernel Lock) may only be fixed *years* after.

Reiser4 may have some nice features, but they also introduce an immense amount of complexity, which is bad. It may not be possible to mount an existing filesystem on a different machine because of non-matching plug-in installations. And I think it is better to implement many things (encryption, compression etc.) in a filesystem-agnostic way (dm-raid, dm-crypt, eCryptFS etc.). I want to be able to change to a different filesystem in the future without losing the features.

This has also been the tenor of most kernel developers who denied Reiser4 inclusion over the last six years. Jeff Garzik had a long discussion with Hans Reiser back then (http://kerneltrap.org/node/5330), and none of the patches sent for Reiser4 in the last years adressed any of this.

Some features in Reiser4, like efficient storage of small files and optimized access patterns for hard drives, are present in available Linux filesystems and are even becoming quite unimportant - simple Netbooks come with >30 GB storage nowadays and SSDs do not care about read/write ordering.