I think what Peter wanted to show you is that some "rather important" kernel developer actually cares about aufs and wants to ping you Junjiro as soon as they have some kernel code modifications ready for overlayfs reconsideration, so you can participate as well (perhaps in the discussion only, perhaps in something more)
In my opinion, if kernel is open for overlayfs inclusion, it must be open for aufs inclusion as well. I don't see any reason why overlayfs should be the only FS merged, not mentioning the fact it has some crucial features missing (inodes numbers consistent during remount). Junjiro, would you please submit a request for inclusion of AUFS to LKML again? Please list in the request what features are implemented in AUFS compared to overlayfs. Please do not be afraid of rejection! Our goal today (in fact) is not to include aufs, but to let the kernel folks LEARN that there is a more advanced ALTERNATIVE. Phillip Lougher, author of squashfs, was rejected SEVERAL TIMES from mainline, but he finally made it to kernel (and I'm happy for that). It just takes several rejections and some time :) I think that aufs code is ready, you just need to describe in few sentences why AUFS is equivalent (or even better) competitor to overlayfs. If I knew how, I would submit the inclusion request myself! ;) Tomas M On Fri, Aug 5, 2011 at 8:29 PM, <sf...@users.sourceforge.net> wrote: > > Thanks forwarding, Pete. > > Here is what I am guessing currently. > - overlayfs will be merged into mainline. > - many aufs users will switch to overlayfs. > - some of them will find the lack of features, (some will not). > - some will ask adding the feature. > + some features will be implemented. > + some features will not. > - in next year, some users will come back to aufs. > > The first point is, > - what is the lacking features? > > I guess the answer is totally up to users, ie. use-case. Some will be > fully satisfied, but the others won't. > I have posted some lacking (but essential for me and unix fs) features > to LKML, and they are described in the overlayfs document. > If users don't care about these feature, it means they will be satisfied. > > The next points are, > - how hard to implement the lacking and requested features? > - how much will the result look like aufs? > > I don't know the answer. > Anyway I have no objection about merging overlayfs into mainline, and I > don't know whether aufs will be merged or not. > But I am afraid that small user community will be a good reason to > abondon aufs development in the future. > > > J. R. Okajima > > ------------------------------------------------------------------------------ > BlackBerry® DevCon Americas, Oct. 18-20, San Francisco, CA > The must-attend event for mobile developers. Connect with experts. > Get tools for creating Super Apps. See the latest technologies. > Sessions, hands-on labs, demos & much more. Register early & save! > http://p.sf.net/sfu/rim-blackberry-1 > ------------------------------------------------------------------------------ BlackBerry® DevCon Americas, Oct. 18-20, San Francisco, CA The must-attend event for mobile developers. Connect with experts. Get tools for creating Super Apps. See the latest technologies. Sessions, hands-on labs, demos & much more. Register early & save! http://p.sf.net/sfu/rim-blackberry-1
