<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN"> <HTML> <HEAD> <META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8"> <META NAME="GENERATOR" CONTENT="GtkHTML/3.3.0"> </HEAD> <BODY> On Tue, 2004-08-24 at 13:50 -0400, Havoc Pennington wrote: <BLOCKQUOTE TYPE=CITE> <PRE> On Tue, 2004-08-24 at 13:34 -0400, David T Hollis wrote: > I'm sure the ultimate question is: when/if it makes it to the stock > kernel, does Fedora begin to support it? There are those camps that > feel that RedHat has some agenda against reiserfs3/4 for some reason and > intentionally crippled it in the past so people wouldn't use it. > Myself, I doubt that. I think the real answer is a technical opinion that there's no compelling reason to have multiple filesystem choices. Each one is significant extra work to deal with, and it just is not worth it. The extra features in Reiser are effectively useless for desktop, see this blog entry for example: <A HREF="http://log.ometer.com/2004-07.html#28">http://log.ometer.com/2004-07.html#28</A> Maybe Reiser is a bit faster in certain server scenarios, but that's not enough rationale for pulling the whole thing in. Still, it's there for people that want it, but I don't personally get why you'd run it. Havoc </PRE> </BLOCKQUOTE> I tend not to agree with this opinion. I've used ReiserFS only once, long time ago, and it was enough for me. I heard of problems with it from time to time, the ones that really scared me, were the ones regarding the difficulty of recovering data from a crashed file-system (it can happen, no matter how atomic, it's made by humans). This is still not a reason not to include-it. Linux is still ALL about choice. It's not hard to write a "BIG FAT WARNING!!! ReiserFS[3|4] are here but not supported, thus not recommended, don't bug us if it crashes" style comment in the release notes, and in anaconda. We should leave this choice up to the user. Fedora/RedHat are loosing a lot on stuff like this. I would understand if it would be a file-system that no-one heard of, but so many other serious distros (not that I would use any of them) default to it. It means that it's not an impossible task. If there are bugs in it and the vanilla kernel doesn't fix them, but the SuSE kernel does, is it a shame in taking the patches from their .src.rpm? I should hope not. Don't get me wrong. I am not a ReiserFS user and probably will not be one in the near future(even distant for that matter), but other people should be able to make that choice and we should respect that. Fedora is not only about open-source, but open-minded development, and open-minded means accepting ReiserFS into the kernel from where I'm standing, and also into anaconda. Open-minded means allowing the user to select xfs, reiserfs and jfs in anaconda in disc-druid, not only ext3. We should bring back the expert flag in anaconda for that. And it's not a complicated task. We already have reiserfs tools, its just a kernel patch and a changed .config. It's not like we're adding a new rpm into "beehive". I would like to quote some of the Fedora Project objectives that bringing support for additional file-systems would accomplish: 8) Include a range of popular packages, beyond those included in Red Hat's commercially supported products. (Limited, of course, to packages that Red Hat can legally provide; also limited to quality packages as defined by our standards.) <-- It's legal, it's not in RHEL and it's popular. 5) Be on the leading edge of open source technology, by adopting and helping develop new features and version upgrades. <-- ReiserFS is leading(bleeding would be more appropriate) edge technology. 2) Do as much of the development work as possible directly in the upstream packages. This includes updates; our default policy will be to upgrade to new versions for security as well as for bugfix and new feature update releases of packages. <-- It can be done also for ReiserFS. Even if Novell is one of their sponsors, Lindows another one, we should help them, we should encourage them. 6) Emphasize usability and a "just works" philosophy in selecting default configuration and designing features. <-- this is what havoc was thinking about. I agree with this, but we should always have an use for the, now basically useless, expert flag for anaconda. Hiding an option if favor of a default, but allowing one to change-it if he really wants to is the way. Not having that option is not a solution. Hope this didn't sound like a flame Best Regards, d3vi1 P.S. No, I'm not a politician and NO, I'm not a guy who stand for moral integrity that much and I didn't smoke anything and I'm rather sober. -- <TABLE CELLSPACING="0" CELLPADDING="0" WIDTH="100%"> <TR> <TD> <TT>Răzvan Corneliu VILT e-mail:</TT><TT><A HREF="mailto:razvan.vilt@linux360.ro">razvan.vilt@linux360.ro</A></TT> <TT>GPG:</TT><TT><A HREF="http://d3vi1.linux360.ro/public-keys/">http://d3vi1.linux360.ro/public-keys/</A></TT><TT> www: </TT><TT><A HREF="http://d3vi1.linux360.ro/">http://d3vi1.linux360.ro/</A></TT> </TD> </TR> </TABLE> </BODY> </HTML>