We could of course claim the space in advance. One way could be to simply take a large file and keep the entire datastore inside it - that could get quite difficult though (since we only need to be able to access files serially it is not _that_ difficult, we could simply divide it into a set of 32 kB sectors and have link to there next part is they can't be continuous. Fragmentation would be an issue though (FS design is not area of experties either)).
Another version would be to simply keep a large dummy file and shrink/grow it to keep space reserved. I don't know how fast/polite/feasable that is. On Wed, Aug 23, 2000 at 05:07:50PM -0500, Brandon wrote: > > > I think it is fair to ask that people have a hundred or so megabytes more > > then the datastore size open on the datastore drive. In the age of 150 > > $ 30 GB drives, I really don't think that is an issue. > > That's quite true. If you put in the README that you need 100M free in > order to run the software then that's fine. It's bad design from an > average-user standpoint but a perfectly acceptable requirement for testers > and developers. Something meant for end users would need to not rely on > users maintaining the proper amount of free space, possibly by writing > filler into whatever space is unused so that Freenet always takes up 100M > no matter what. But we're not really to that point. > > > _______________________________________________ > Freenet-dev mailing list > Freenet-dev at lists.sourceforge.net > http://lists.sourceforge.net/mailman/listinfo/freenet-dev > -- \oskar _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
