Page 11 of 12 FirstFirst ... 9101112 LastLast
Results 101 to 110 of 114

Thread: XWayland 2D Performance Appears Better Than XMir

  1. #101
    Join Date
    Sep 2011
    Posts
    184

    Default

    Quote Originally Posted by mrugiero View Post
    It's funny how two official sources (developers) contradict between them. That's not a point on favor of trying to maintain something targeting Mir outside of Canonical.
    You have a point there. Still, the more recent statement I linked to, sounds credible, although having a stable API by 13.10 seems optimistic to me.

  2. #102
    Join Date
    Jan 2011
    Posts
    1,287

    Default

    Quote Originally Posted by alexThunder View Post
    You have a point there. Still, the more recent statement I linked to, sounds credible, although having a stable API by 13.10 seems optimistic to me.
    I do believe they'll freeze the API by the time they release their first version with a native desktop. Otherwise, no developer would be able to target them, and I think they should want that to happen. However, I don't think that happens until Unity 8 is out, since nobody will target Mir while running only to be able to run XMir.
    The point is I won't blame a dev that doesn't believe them, because they actually made the opposite statement before. And if they don't believe that, I would understand if they feel no motivation to be the ones maintaining a compatibility layer.

  3. #103
    Join Date
    Jun 2013
    Posts
    60

    Default

    Quote Originally Posted by mrugiero View Post
    I do believe they'll freeze the API by the time they release their first version with a native desktop. Otherwise, no developer would be able to target them, and I think they should want that to happen. However, I don't think that happens until Unity 8 is out, since nobody will target Mir while running only to be able to run XMir.
    The point is I won't blame a dev that doesn't believe them, because they actually made the opposite statement before. And if they don't believe that, I would understand if they feel no motivation to be the ones maintaining a compatibility layer.
    i dont see it being as stable as wayland by 14.04 and, a lot of it do to it mostly being a Copy Paste job and, the C++ tool there using to do the Copy Paste Job.

  4. #104
    Join Date
    Jan 2011
    Posts
    1,287

    Default

    Quote Originally Posted by spacetoilet View Post
    i dont see it being as stable as wayland by 14.04 and, a lot of it do to it mostly being a Copy Paste job and, the C++ tool there using to do the Copy Paste Job.
    I think we are talking about different stabilities. Wayland is already stable as in the API not changing (for 1.0 releases), and that's the stability I was talking about.
    About being relatively bug free, I can't say much since I don't have any of those running. My guess is Wayland is and will be more stable on that side, because it has better defined goals (having clear what you want it to do helps a lot to program it in a relatively bug free way) and I guess it will be lighter, not having (not being mandatory, actually) an actual server running; the latter means there's less surface for the bugs to appear.
    There's also the fact the library is automatically generated from the protocol definition, which avoids a lot of sources for bugs.

  5. #105
    Join Date
    Jun 2013
    Posts
    60

    Default

    Quote Originally Posted by mrugiero View Post
    I think we are talking about different stabilities. Wayland is already stable as in the API not changing (for 1.0 releases), and that's the stability I was talking about.
    About being relatively bug free, I can't say much since I don't have any of those running. My guess is Wayland is and will be more stable on that side, because it has better defined goals (having clear what you want it to do helps a lot to program it in a relatively bug free way) and I guess it will be lighter, not having (not being mandatory, actually) an actual server running; the latter means there's less surface for the bugs to appear.
    There's also the fact the library is automatically generated from the protocol definition, which avoids a lot of sources for bugs.
    One of the Ubuntu Developers said on his blog Mir's API will keep changing, (can't remember what one -_-) you know the Ubuntu Way, Wayland on the other hand is if it's not Broken dont fix it

  6. #106
    Join Date
    Sep 2008
    Location
    Vilnius, Lithuania
    Posts
    2,542

    Default

    Quote Originally Posted by DDF420 View Post
    A new unity comes out with every ubuntu release 11.10 unity 4 12.04 unity 5 12.10 unity 6 13.04 unity 7 so on. what is not clear to me is what the 11.04 first unity version number was. 3 ? or were versions 123 introduced between 11.04 and 11.10 ?
    Hmm, that still doesn't make that much sense given that there won't be Unity 8 on 13.10 or even 14.04.

  7. #107
    Join Date
    Jan 2011
    Posts
    1,287

    Default

    Quote Originally Posted by spacetoilet View Post
    One of the Ubuntu Developers said on his blog Mir's API will keep changing, (can't remember what one -_-) you know the Ubuntu Way, Wayland on the other hand is if it's not Broken dont fix it
    Yes, I quoted that and pasted a link a few posts ago, but then there's a newer statement on the mailing list that says the opposite (and I already gave my opinion about that overly common change of mind of Canonical and their devs in the Mir subjects).

  8. #108
    Join Date
    Jun 2009
    Posts
    1,125

    Default

    Quote Originally Posted by TheOne View Post
    What I mean with more advanced is that well, XWayland segfaulted to the guy doing some of the tests where XMir ran fine like 3D as someone else mentioned. So there is a big chance that canonical developers improved XWayland while doing the refactory to xMir. There is also the fact that those test the guy did where running Wayland over Weston and just applications using XWayland instead of using XWayland to run an entire DE so that would actually run faster than what canonical guys are offering with XMir and a full Desktop Environment for backward compatibility issues of DE still running on X.

    There could also be a slight chance that running Wayland Weston and XWayland with 3D applications isn't mature yet and the same could apply to Mir. Anyways this tests aren't the same to not say fair. We need a pure XWayland vs XMir test, both running an entire Desktop environment.
    Xwayland does ran 3d good enough but until xorg 1.13[current ubuntu version] with some downstream patches, beyond this point in time XWayland was delayed due to some performance hits and weird roundtrips[like input delay] and through in the land NO-FIX until some changes that i think keith is working on for DRI3 that will solve the base issue[and some patches waiting in Xorg mailinglist that are being refactored cuz are too hacky].

    remember that unlike canonical that need to rush to probe Mir won't be just another candidate for ubuntu project graveyard they don't mind to push horrible hacky code so they can PR "OMG mirz workz", the FLOSS team like to do thing properly and won't release anything until all major structural bugs have an right structural solution[in this case DRI3 and Xorg 1.15(temptingly)].

    so yes Xwayland was a nice experiment to prove that is possible to provide X compatibility and test the weaknesses of the current stack in such scenario but was never considered for real use until the base issues are tackled, Xmir share the same issues but canonical [as history proves] don't have much issue in putting crappy patches in releases and wait until wayland developers upstream the proper solutions[they won't code them nor they have the people to do it].

    as an interesting fact Microsoft and Apple have contributed more code to linux than Canonical

  9. #109
    Join Date
    Apr 2010
    Posts
    721

    Default

    Quote Originally Posted by mrugiero View Post
    I do believe they'll freeze the API by the time they release their first version with a native desktop. Otherwise, no developer would be able to target them, and I think they should want that to happen.
    Not quite. There are two APIs in question - the client one, and the server one. The client one is already supposed to be stable, and that's the one that's needed for writing applications - the one Qt and Gtk and other toolkits need to support. As you say, this API *must* be stable, in order for the platform to be useful for applications.

    The server API is the one that's relevant for porting desktops - for writing the Mir equivalent of a compositor. That's the bit the KWin team care about, and it's also the bit where we've received mixed messages about whether it will ever be stable. If Canonical only care about Unity, there's no need for it ever to be declared stable, but for other desktops to run on Mir, that's a requirement.

  10. #110
    Join Date
    Jan 2013
    Posts
    525

    Default

    Quote Originally Posted by jrch2k8 View Post
    as an interesting fact Microsoft and Apple have contributed more code to linux than Canonical
    How does that even come into a discussion about the technical merits of XWayland and XMir?
    Or are you one of these guys who likes strawman arguments?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •