From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hans Reiser Subject: Re: Reiser4 as positional B-Tree Date: Tue, 04 Nov 2003 01:00:26 -0800 Message-ID: <3FA76AAA.8080600@namesys.com> References: Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: list-help: list-unsubscribe: list-post: Errors-To: flx@namesys.com In-Reply-To: List-Id: Content-Type: text/plain; charset="us-ascii"; format="flowed" To: Gregor Zeitlinger Cc: reiserfs mailing list Gregor Zeitlinger wrote: >Hi, > >I've already been on this mailing list a year ago. I was wondering wheater >I could make my native xml database a file plugin in reiser4. > >Now, that the code base is much more matured and that reiser4 is almost >ready (as I understand), this seems much more feasable. > >Those are my questions >1) As part of my xml database, I still need a positional B-Tree. Since >files in reiserfs are stored in a (kinda) B+-Tree, I was wondering wheater >a positional B-Tree could be made a file plugin. It's basically the same >as the unix file plugin, with the exception that you'd need the additional >operation of inserting (not overwriting) data in the middle of a file. > >2) If the plugin architecture is flexible enough, I'm wondering wheater I >could port my xml database to being a file plugin. It uses a custom page >layout and custom balance, split and merge algorithms. Also, I use c++, >but the source could be ported to c, since I don't make heavy use of >object oriented features. I haven't put the source code online, since it's >not mature enough for that, but I can give to anyone interested. > > > Please send a definition of a positional B-tree. Also indicate whether you are an individual vs. an organization that might be interested in sponsoring our creating a plugin of that kind to your specifications. (Yes, we are very poor this month, so I ask questions of that kind at every opportunity....;-) ) Purely guessing at what a positional B-tree is, I would say probably yes, and it would probably be valuable for the filesystem generally. -- Hans