> I just saw that Greg KH considered dropping the p9auth in 2.6.33 if > nobody was interested in it. I do not know how well this one fits with > the rest that Glendix tries to do and whether it provides some > functionality needed by the glendix project...
Is Ashwin Ganti not willing to maintain it? Did Greg or anyone mention changes that need to be done (I think they did)? Maybe Glendix can take over maintainership. Should be easy since the driver is (AFAIK) complete, the userspace stuff needs taking care of. > > On 22 Juli, 21:27, "J. R. Mauro" <[email protected]> wrote: >> On Jul 22, 2009, at 15:23, staalmannen <[email protected]> wrote: >> >> >> >> > Is there a repo for /net with an RSS feed? I already subscribe to the >> > glendix Hg repo and the Private Namespaces repo to see how you guys >> > are doing :) >> >> /net is done in the main glendox repo. There just haven't been any >> recent commits. We're all really busy. >> >> >> >> > I really wish I knew how to code and stuff... >> >> Visit the local bookstore. >> >> >> >> > A completely unrelated (and probably herretic) question: does Wine >> > compile on APE + the plan9 ported X? If it did people who want to be >> > "bare metal" plan9 users could still enjoy a browser etc :P. >> >> APE is probably nowhere near what wine needs. There is a Linux >> emulator for plan9 that people run browsers in. I would personally go >> this route but plan 9 won't support half the hardware on my laptop. >> >> >> >> > On 20 Juli, 17:02, "J. R. Mauro" <[email protected]> wrote: >> >> On Jul 20, 2009, at 10:51, Sam Fuqua <[email protected]> wrote: >> >> >>> I'm sorry if this has already been answered, but how close are we to >> >>> a /net implementation for Linux? >> >> >> Some of the basic structure is there for tcp. Actually generating new >> >> connections is not started yet. 10 percent maybe? >> >> >>> On Mon, Jul 20, 2009 at 10:23 AM, J. R. Mauro <[email protected]> >> >>> wrote: >> >> >>> I somehow don't see Aaron moving KDE to /dev/draw. I also really >> >>> don't >> >>> see the benefit of crap like dbus and hal. Cross polination will be >> >>> nice, but it will also suck. Don't get your hopes up about anything >> >>> migrating. The only tasty thing people might want us /net >> >> >>> On Jul 20, 2009, at 9:45, Jens Staal <[email protected]> wrote: >> >> >>>> I think you understood me correctly :) >> >> >>>> The more I think about it the smarter it seems - a very elegant >> >>> way to >> >>>> break a catch 22 where "good enough" legacy blocks better >> >>> solutions - >> >>>> and with more applications and tools available on top of the >> >>>> alternative paradigm, more people will be ready to work on it. >> >> >>>> I think this definitely can be the beginning of something big if >> >>>> things really take off. >> >> >>>> 2009/7/20, Rahul Murmuria <[email protected]>: >> >> >>>>> Hi staalmannen, >> >> >>>>> You are just beginning to realize what Glendix's motive is. One of >> >>>>> our >> >>>>> strongest wishes is that regular Linux-based apps like KDE and GCC >> >>>>> start using the Plan 9 features we add to Linux in the future. >> >>>>> Like >> >>>>> network based apps should use a /net virtual filesystem instead of >> >>>>> the >> >>>>> socket ioctl calls. Basically do things "the Plan 9 way", while >> >>>>> staying in the Linux world. So, as a corollary, If a Linux apps >> >>> ends >> >>>>> up using exclusively Plan 9-like features, then its executable >> >>> should >> >>>>> work out-of-the-box in Plan 9. Like KDE should start using /dev/ >> >>> draw >> >>>>> instead of X11. >> >> >>>>> This will lead to what you just described. Others are welcome to >> >>>>> correct me, if I have failed to see your point. >> >> >>>>> Have fun! >> >> >>>>> On Mon, Jul 20, 2009 at 1:58 PM, staalmannen<[email protected]> >> >>>>> wrote: >> >> >>>>>> I just thought of something else where Glendix actually provides >> >>>>>> additional value compared to all other user-space or virtualized >> >>>>>> plan9 >> >>>>>> implementations (like Inferno, 9vx, 9ports etc). In contrast to >> >>>>>> those >> >>>>>> - glendix might be used for incremental porting of complex apps >> >>> with >> >>>>>> lots of dependencies not filled by native plan9. >> >> >>>>>> The thing I was thinking of (which I do not know if it is at all >> >>>>>> feasible or just completely stupid) is for example porting of KDE >> >>>>>> (which gives a full productivity suite and is made to be >> >>> portable). >> >>>>>> Several things are not possible to even start with in native >> >>>>>> Plan9 >> >>>>>> (for example, no c++ capable native compiler - I have a gut >> >>> feeling >> >>>>>> that LLVM/clang may be more portable than GCC since the only >> >>> problem >> >>>>>> officially recognized by LLVM for porting to plan9 is its >> >>>>>> dependencies >> >>>>>> of a Bourne shell - and if rc could be replacing that...). >> >>>>>> The nice thing is ofcourse that the whole porting in an hybrid >> >>>>>> environment could give working semi-ported intermediates that >> >>> run on >> >>>>>> glendix but not on plan9 or GNU (a "kwin" on rio, a plan9-type >> >>>>>> Hal >> >>>>>> and >> >>>>>> Dbus etc, which "downstream" (towards the kernel) talk Plan9 and >> >>>>>> "upstream" (towards applications) have the usual interface of the >> >>>>>> original components) - sort of the way glendix is built right >> >>> now - >> >>>>>> incremental improvements where each improvement has a potential >> >>> use >> >>>>>> even before full Plan9 compatibility is achieved. >> >> >>>>>> Well.... I do not know if I am just rambling, but I thought this >> >>>>>> was a >> >>>>>> pretty interesting and unique advantage of Glendix compared to >> >>> other >> >>>>>> ways of running Plan9. >> >> >>>>>> On May 28, 4:56 pm, "J.R. Mauro" <[email protected]> wrote: >> >>>>>>> On Thu, May 28, 2009 at 10:52 AM, staalmannen >> >>> <[email protected]> >> >>>>>>> wrote: >> >> >>>>>>>> OK thanks for that explanation (and it was probably not a >> >>>>>>>> commit >> >>>>>>>> - I >> >>>>>>>> might have misinterpreted it, as I said before I am just a >> >>>>>>>> curious >> >>>>>>>> layman so those things happen). The "ideological" resistance >> >>> as I >> >>>>>>>> interpreted it was mostly a comment about that plan9 >> >>> compatibility >> >>>>>>>> should not be a purpose on its own and that it (plan9) had an >> >>>>>>>> "ugly >> >>>>>>>> interface". >> >> >>>>>>> I don't remember anyone saying Plan 9 was ugly. Just that >> >>> exposing >> >>>>>>> jiffies to userspace was dumb. Which, maybe it is. Like I said, >> >>>>>>> those >> >>>>>>> last two fields can be whatever, so changing it isn't a big >> >>>>>>> deal. >> >>>>>>> And >> >>>>>>> IIRC, Chris did change it. >> >> >>>>>>>> Sorry for all the questions - feel free to tell me that I am >> >>>>>>>> annoying >> >>>>>>>> if you get bothered by them ;) >> >> >>>>>>>> On a completely different track - what would you feel would be >> >>> the >> >>>>>>>> "killer feature" to get into mainline from glendix? Would it >> >>> be / >> >>>>>>>> net >> >>>>>>>> or something else? Some stuff like /proc has been widely >> >>>>>>>> accepted so >> >>>>>>>> why do you think that other useful things derived from the >> >>>>>>>> Plan9 >> >>>>>>>> design concept have not been integrated previously? >> >> >>>>>>> If by 'killer' you mean absolute fatality, then full Plan 9 >> >>>>>>> support. >> >>>>>>> In the meantime, smaller goals like /net, improving /proc, true >> >>>>>>> rio on >> >>>>>>> top of /dev/draw, etc. will be very nice milestones. >> >> >>>>>>>> On 28 Maj, 15:24, "J.R. Mauro" <[email protected]> wrote: >> >>>>>>>>> On Thu, May 28, 2009 at 7:39 AM, staalmannen >> >>>>>>>>> <[email protected]> >> >>>>>>>>> wrote: >> >> >>>>>>>>>> I saw a kernel commit by you for /dev/time where the feedback >> >>>>>>>>>> was >> >>>>>>>>>> that >> >>>>>>>>>> it could possibly be added as a /staging/plan9 directory in >> >>> the >> >> >>>>>>>>> Wait, a commit? I didn't know /dev/time made it anywhere (sad >> >>>>>>>>> face)... >> >> >>>>>>>>> I don't want to abuse staging. It's already a controversy, and >> >>>>>>>>> straining it by falling back on Greg to always say yes is a >> >>>>>>>>> bad >> >>>>>>>>> idea. >> >>>>>>>>> FWICS, things in staging get less feedback than things put >> >>> out on >> >>>>>>>>> LKML. The thing is that if no one cares, these little patches >> >>>>>>>>> won't >> >>>>>>>>> get picked up. I don't think that shoehorning them into >> >>>>>>>>> staging >> >>>>>>>>> is >> >>>>>>>>> good. >> >> >>>>>>>>> Glendix will probably wind up having to host these stray >> >>>>>>>>> patches, and >> >>>>>>>>> maybe someday we'll have an arsenal of examples of us getting >> >>>>>>>>> Plan 9 >> >>>>>>>>> code into Linux. Yes, in some ways, Linux is a popularity >> >>>>>>>>> contest. But >> >>>>>>>>> it stems from people trusting others to not screw things up. >> >> >>>>>>>>>> mainline, but that you were afraid of flamewars. In a later >> >>>>>>>>>> reply >> >>>>>>>>>> someone else questioned the plan9 compatibility of the >> >>> format of >> >>>>>>>>>> /dev/ >> >>>>>>>>>> time (and not the design as such). >> >> >>>>>>>>> The discussion about the format was not ideological, it was >> >>>>>>>>> logical. >> >>>>>>>>> /dev/time exposed jiffies, which isn't really something >> >>> userspace >> >>>>>>>>> should necessarily see. Plan 9's man pages state that the >> >>>>>>>>> latter 2 >> >>>>>>>>> fields are implementation-defined. So the first 2 fields are >> >>>>>>>>> ok, and >> >>>>>>>>> all that really matter. That and the formatting, which I >> >>>>>>>>> patched for >> >>>>>>>>> Chris. >> >> >>>>>>>>> Also, the CUSE system will be out shortly, and something as >> >>>>>>>>> simple and >> >>>>>>>>> non-performance-intensive as /dev/time will probably want to >> >>>>>>>>> be >> >>>>>>>>> done >> >>>>>>>>> in userspace. >> >> >>>>>>>>>> Do you often experience such "ideological" resistance to >> >>>>>>>>>> improving >> >>>>>>>>>> linux with plan9 functionality? >> >> >>>>>>>>> I experience resistance to improving Linux with Linux >> >>>>>>>>> functionality. >> >> >>>>>>>>>> On 27 Maj, 19:23, "J.R. Mauro" <[email protected]> wrote: >> >>>>>>>>>>> On Wed, May 27, 2009 at 1:16 PM, staalmannen >> >>>>>>>>>>> <[email protected]> >> >>>>>>>>>>> wrote: >> >> >>>>>>>>>>>> Thanks! >> >> >>>>>>>>>>>> It seems like very realistic hopes I would say (extending >> >>> the >> >>>>>>>>>>>> current >> >>>>>>>>>>>> GNU/Linux ecosystem with plan9 features). I wish you all >> >>> luck >> >>>>>>>>>>>> with the >> >>>>>>>>>>>> project! >> >> >>>>>>>>>>>> I suppose this also means that there might be a gradual >> >>> cross- >> >>>>>>>>>>>> polination where some parts actually may get wide >> >>>>>>>>>>>> acceptance >> >>>>>>>>>>>> even >> >>>>>>>>>>>> before the complete port of plan9 is complete? >> >>>>>>>>>>>> For example, if /net would be accepted in the mainline >> >> ... >> >> läs mer » > --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/glendix?hl=en -~----------~----~----~----~------~----~------~--~---
