Re: Unity desktop and maverick backport
Bonjour Didier, Didier Roche [2010-11-18 19:25 +0100]: However, after some porting discussions and following the natty work I think we should perhaps consider not doing that because it's going to take quite some work for a moderated benefit and we would better spend those efforts in making natty rocking. Thanks for the summary. I agree that we shouldn't put a lot of work hours into these backports, but rather spend it to making our current daily live CDs actually work and keep them working. If someone wants to test unity, but isn't able to download a daily image and use usb-creator, then we really don't want to ask the same person to install the backported unity packages on a production maverick system. I know for myself how painful the current packages are, and during the development cycle we won't always be able to guarantee that we don't trash the user's configuration. There is a lot of fiddling that needs to happen for proper configuration migration. So I actually think USB images will allow more people to test Unity than Maverick backports. Thanks, Martin -- Martin Pitt| http://www.piware.de Ubuntu Developer (www.ubuntu.com) | Debian Developer (www.debian.org) -- ubuntu-devel mailing list ubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
Re: Unity desktop and maverick backport
On 11/19/2010 02:10 AM, Alex Launi wrote: On Thu, Nov 18, 2010 at 6:59 PM, Dustin Kirkland kirkl...@ubuntu.com mailto:kirkl...@ubuntu.com wrote: I thought a good 2D experience was also one of the key objectives? Yeah, but the good 2D experience is really just a fall back to Gnome 2.X panel, and the current Ubuntu desktop. It will be good to have this fall back well tested, but that doesn't help us test Unity. Sorry for the tangent but I could not find more information about this online: is the Unity UI fundamentally tied to 3D acceleration support and compositing? IMHO if it was technically possible it would be more uniform to have the 2D experience built on the same new UI/UX concepts just a bit less visually intense, much like the 'desktop effects' setting used to provide more or less the same UI with various degrees of glitz. thanks Jani -- ubuntu-devel mailing list ubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
Re: Unity desktop and maverick backport
On Thu, 2010-11-18 at 21:48 -0600, Dustin Kirkland wrote: I think if the VM supports 3D acceleration, this would be cool. Would you be happy to try it and see if it works? Will gladly try it, but I don't think we're going to get 3D acceleration in the VM. I was thinking of this as an approach for the 2D testing and the overall integration (minus 3D effects). Yeah, it could work well for the 2D fallback, and could be a great way for those without sufficient 3D hardware to test and help improve Ubuntu. Jono -- Jono Bacon Ubuntu Community Manager jono(at)ubuntu(dot)com www.ubuntu.com : www.jonobacon.org www.twitter.com/jonobacon : www.identi.ca/jonobacon -- ubuntu-devel mailing list ubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
Re: Unity desktop and maverick backport
In my opinion, for what it is worth, this sounds like an unfortunate, but necessary trade off. I think we will lose a fairly large degree of testing and feedback, by forcing interested contributors to move Natty so early. However, I think it's rational to trade that for an increased focus on the ultimate quality of the new compiz-based unity in Natty and beyond. I think we'll get the most useful feedback from people *using* Unity. So, this means that we'll need to focus on supporting early Natty adopters, for instance paying more attention to quickly resolving adoption blocking bugs. My $.02 Cheers, Rick On Thu, 2010-11-18 at 19:25 +0100, Didier Roche wrote: Hi everybody, As some of you may know, there have been some discussions about backporting unity compiz to maverick as we had backported unity to lucid with a dedicated ppa and its own session. However, after some porting discussions and following the natty work I think we should perhaps consider not doing that because it's going to take quite some work for a moderated benefit and we would better spend those efforts in making natty rocking. Some bits what came from discussions between ubuntu desktop and dx teams: * Why do we want to backport? - usually it's to make easier for users to test the new version and give some feedback on it. The first round of feedback will be about things not starting, or not working at all or crashing, we will get that feedback from the natty users. Later on we will want extra eyes on the user experience but by the time we are there it will be really hard to backport the new stack due to new depends (details on that later). * New unity means new compiz which means users will have no working desktop left, that's not something we should get our users in. Indeed, the new compiz is not made to be installed with the old one, the upgrade will replace compiz 0.8 but has lot of issues still: the configuration is not migrated, the keybindings are not working, the workspace layout and switcher are not working, the session registration is not working, the desktop capplet needs to be updated, the GNOME keybindings capplet is not working. Some of those issues are fixed in natty, but we can't backporting every single GNOME applications to make them work in a maverick ppa. - the new unity packaging is not made to have old and new unity installed at the same time so the old unity will not be installed anymore. - the new unity is not usable as a desktop yet, which means the user will not have the old unity, compiz under GNOME will be broken is several ways which let the GNOME session hard to use, the new unity is not ready for production ... users who will want to give unity a try will just land in a situation when they have no environment left they can use for work... it would be less breakage to suggest them to update to natty where we fix those integration issues. * The new unity stack will be hard to backport - the next indicators uploads will build-depends on gtk3 (even if we don't use it we need to have libraries in natty to build gtk2 and gtk3 version to allow people to start porting work), we use new glib api, etc. Backporting the stack unity will need is going to turn into lot of work and a non trivial task. We think users will have a better experience by trying unity on natty and that we will gather more useful and coherent data, since it's likely to be more stable than getting a working - and a less tested by our team - backport. didrocks on behalf of the ubuntu desktop and dx teams -- ubuntu-devel mailing list ubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
Re: Unity desktop and maverick backport
On Thu, 2010-11-18 at 10:47 -0800, Rick Spencer wrote: In my opinion, for what it is worth, this sounds like an unfortunate, but necessary trade off. I think we will lose a fairly large degree of testing and feedback, by forcing interested contributors to move Natty so early. However, I think it's rational to trade that for an increased focus on the ultimate quality of the new compiz-based unity in Natty and beyond. I think we'll get the most useful feedback from people *using* Unity. So, this means that we'll need to focus on supporting early Natty adopters, for instance paying more attention to quickly resolving adoption blocking bugs. My $.02 Agreed. I think the trick here to encouraging testing is installation on USB key-rings; they are cheap, the installations run natively on the hardware, and they are low risk. I plan on getting documentation together and raising the awareness of this soon. Jono -- Jono Bacon Ubuntu Community Manager jono(at)ubuntu(dot)com www.ubuntu.com : www.jonobacon.org www.twitter.com/jonobacon : www.identi.ca/jonobacon -- ubuntu-devel mailing list ubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
Re: Unity desktop and maverick backport
Hi Didier, Thanks for sharing this proposal. Based on the technical discussion below, it seems like the right trade-off to make, if we can't have both. Having said that, it becomes even more important to the overall quality of the Unity desktop that we ensure as many users try it on Natty as soon as possible. I'd like to make a few suggestions. - Send out a call for testing, specific to Unity desktop (QA Team) - Track Unity related bugs and make sure they are getting triaged and resolved quickly; Monitor bug reports closely (QA Team-bdmurray) - Layout key dates for checkpoints (Desktop, DX and QA Teams) - Make go/no-go recommendation based on test results and bug data (QA Team) What do you think? Thanks, Marjo On Thu, 2010-11-18 at 19:25 +0100, Didier Roche wrote: Hi everybody, As some of you may know, there have been some discussions about backporting unity compiz to maverick as we had backported unity to lucid with a dedicated ppa and its own session. However, after some porting discussions and following the natty work I think we should perhaps consider not doing that because it's going to take quite some work for a moderated benefit and we would better spend those efforts in making natty rocking. Some bits what came from discussions between ubuntu desktop and dx teams: * Why do we want to backport? - usually it's to make easier for users to test the new version and give some feedback on it. The first round of feedback will be about things not starting, or not working at all or crashing, we will get that feedback from the natty users. Later on we will want extra eyes on the user experience but by the time we are there it will be really hard to backport the new stack due to new depends (details on that later). * New unity means new compiz which means users will have no working desktop left, that's not something we should get our users in. Indeed, the new compiz is not made to be installed with the old one, the upgrade will replace compiz 0.8 but has lot of issues still: the configuration is not migrated, the keybindings are not working, the workspace layout and switcher are not working, the session registration is not working, the desktop capplet needs to be updated, the GNOME keybindings capplet is not working. Some of those issues are fixed in natty, but we can't backporting every single GNOME applications to make them work in a maverick ppa. - the new unity packaging is not made to have old and new unity installed at the same time so the old unity will not be installed anymore. - the new unity is not usable as a desktop yet, which means the user will not have the old unity, compiz under GNOME will be broken is several ways which let the GNOME session hard to use, the new unity is not ready for production ... users who will want to give unity a try will just land in a situation when they have no environment left they can use for work... it would be less breakage to suggest them to update to natty where we fix those integration issues. * The new unity stack will be hard to backport - the next indicators uploads will build-depends on gtk3 (even if we don't use it we need to have libraries in natty to build gtk2 and gtk3 version to allow people to start porting work), we use new glib api, etc. Backporting the stack unity will need is going to turn into lot of work and a non trivial task. We think users will have a better experience by trying unity on natty and that we will gather more useful and coherent data, since it's likely to be more stable than getting a working - and a less tested by our team - backport. didrocks on behalf of the ubuntu desktop and dx teams -- Marjo F. Mercado Ubuntu QA Team Manager W: (917) 338-6551 IRC: marjo -- ubuntu-devel mailing list ubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
Re: Unity desktop and maverick backport
Well there always is VMs and separate ubuntu installs for testing without breakages affecting your desktop usage. --fagan On Thu, 2010-11-18 at 10:47 -0800, Rick Spencer wrote: In my opinion, for what it is worth, this sounds like an unfortunate, but necessary trade off. I think we will lose a fairly large degree of testing and feedback, by forcing interested contributors to move Natty so early. However, I think it's rational to trade that for an increased focus on the ultimate quality of the new compiz-based unity in Natty and beyond. I think we'll get the most useful feedback from people *using* Unity. So, this means that we'll need to focus on supporting early Natty adopters, for instance paying more attention to quickly resolving adoption blocking bugs. My $.02 Cheers, Rick On Thu, 2010-11-18 at 19:25 +0100, Didier Roche wrote: Hi everybody, As some of you may know, there have been some discussions about backporting unity compiz to maverick as we had backported unity to lucid with a dedicated ppa and its own session. However, after some porting discussions and following the natty work I think we should perhaps consider not doing that because it's going to take quite some work for a moderated benefit and we would better spend those efforts in making natty rocking. Some bits what came from discussions between ubuntu desktop and dx teams: * Why do we want to backport? - usually it's to make easier for users to test the new version and give some feedback on it. The first round of feedback will be about things not starting, or not working at all or crashing, we will get that feedback from the natty users. Later on we will want extra eyes on the user experience but by the time we are there it will be really hard to backport the new stack due to new depends (details on that later). * New unity means new compiz which means users will have no working desktop left, that's not something we should get our users in. Indeed, the new compiz is not made to be installed with the old one, the upgrade will replace compiz 0.8 but has lot of issues still: the configuration is not migrated, the keybindings are not working, the workspace layout and switcher are not working, the session registration is not working, the desktop capplet needs to be updated, the GNOME keybindings capplet is not working. Some of those issues are fixed in natty, but we can't backporting every single GNOME applications to make them work in a maverick ppa. - the new unity packaging is not made to have old and new unity installed at the same time so the old unity will not be installed anymore. - the new unity is not usable as a desktop yet, which means the user will not have the old unity, compiz under GNOME will be broken is several ways which let the GNOME session hard to use, the new unity is not ready for production ... users who will want to give unity a try will just land in a situation when they have no environment left they can use for work... it would be less breakage to suggest them to update to natty where we fix those integration issues. * The new unity stack will be hard to backport - the next indicators uploads will build-depends on gtk3 (even if we don't use it we need to have libraries in natty to build gtk2 and gtk3 version to allow people to start porting work), we use new glib api, etc. Backporting the stack unity will need is going to turn into lot of work and a non trivial task. We think users will have a better experience by trying unity on natty and that we will gather more useful and coherent data, since it's likely to be more stable than getting a working - and a less tested by our team - backport. didrocks on behalf of the ubuntu desktop and dx teams -- ubuntu-devel mailing list ubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
Re: Unity desktop and maverick backport
Am Donnerstag, den 18.11.2010, 18:53 + schrieb Shane Fagan: Well there always is VMs and separate ubuntu installs for testing without breakages affecting your desktop usage. But this requires that the VM has 3D support (which KVM doesn't have). -- Benjamin Drung Ubuntu Developer (www.ubuntu.com) | Debian Maintainer (www.debian.org) signature.asc Description: This is a digitally signed message part -- ubuntu-devel mailing list ubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
Re: Unity desktop and maverick backport
On Thu, Nov 18, 2010 at 04:03:32PM -0500, Marjo Mercado wrote: Hi Didier, Thanks for sharing this proposal. Based on the technical discussion below, it seems like the right trade-off to make, if we can't have both. Having said that, it becomes even more important to the overall quality of the Unity desktop that we ensure as many users try it on Natty as soon as possible. I'd like to make a few suggestions. - Send out a call for testing, specific to Unity desktop (QA Team) - Track Unity related bugs and make sure they are getting triaged and resolved quickly; Monitor bug reports closely (QA Team-bdmurray) - Layout key dates for checkpoints (Desktop, DX and QA Teams) - Make go/no-go recommendation based on test results and bug data (QA Team) What do you think? Don't forget to define a test plan for people to follow before sending out a call for testing. You might also want to make a page in wiki for folks to list their findings. Thanks, Marjo On Thu, 2010-11-18 at 19:25 +0100, Didier Roche wrote: Hi everybody, As some of you may know, there have been some discussions about backporting unity compiz to maverick as we had backported unity to lucid with a dedicated ppa and its own session. However, after some porting discussions and following the natty work I think we should perhaps consider not doing that because it's going to take quite some work for a moderated benefit and we would better spend those efforts in making natty rocking. Some bits what came from discussions between ubuntu desktop and dx teams: * Why do we want to backport? - usually it's to make easier for users to test the new version and give some feedback on it. The first round of feedback will be about things not starting, or not working at all or crashing, we will get that feedback from the natty users. Later on we will want extra eyes on the user experience but by the time we are there it will be really hard to backport the new stack due to new depends (details on that later). * New unity means new compiz which means users will have no working desktop left, that's not something we should get our users in. Indeed, the new compiz is not made to be installed with the old one, the upgrade will replace compiz 0.8 but has lot of issues still: the configuration is not migrated, the keybindings are not working, the workspace layout and switcher are not working, the session registration is not working, the desktop capplet needs to be updated, the GNOME keybindings capplet is not working. Some of those issues are fixed in natty, but we can't backporting every single GNOME applications to make them work in a maverick ppa. - the new unity packaging is not made to have old and new unity installed at the same time so the old unity will not be installed anymore. - the new unity is not usable as a desktop yet, which means the user will not have the old unity, compiz under GNOME will be broken is several ways which let the GNOME session hard to use, the new unity is not ready for production ... users who will want to give unity a try will just land in a situation when they have no environment left they can use for work... it would be less breakage to suggest them to update to natty where we fix those integration issues. * The new unity stack will be hard to backport - the next indicators uploads will build-depends on gtk3 (even if we don't use it we need to have libraries in natty to build gtk2 and gtk3 version to allow people to start porting work), we use new glib api, etc. Backporting the stack unity will need is going to turn into lot of work and a non trivial task. We think users will have a better experience by trying unity on natty and that we will gather more useful and coherent data, since it's likely to be more stable than getting a working - and a less tested by our team - backport. didrocks on behalf of the ubuntu desktop and dx teams -- Marjo F. Mercado Ubuntu QA Team Manager W: (917) 338-6551 IRC: marjo -- ubuntu-devel mailing list ubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel -- ubuntu-devel mailing list ubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
