Hi, folks! We're now frozen for F22 Final, but a pile of blocker and FE nominations arrived after the review meeting, several with updates attached. It would be good to get these reviewed before next week.
Especially of note, we have an FE request for https://bugzilla.redhat.com/show_bug.cgi?id=1211015 which would necessitate pulling in the KDE Frameworks 5.10.0 release:
https://admin.fedoraproject.org/updates/kf5-plasma-5.10.0-2.fc22
which is obviously a significant change to the KDE spin. We also have GNOME 3.16.2 pending; they haven't filed that yet, but I'm expecting it today or tomorrow. If we're going to pull in a pair of significant desktop updates, we should do it as soon as possible. I really wish we could somehow stop having these situations where the desktop teams want to push big updates right after freeze points, as it seems to keep happening, but for right now we need to make a decision.
So, I'd like to get sufficient votes on the proposed blockers and at least the significant proposed FEs as soon as possible. Would folks prefer voting on-list or in-bug, or having a special meeting, perhaps tomorrow (05-13) or Thursday (05-14) at the usual time (1600 UTC)? Please reply if you'd be willing to attend a special meeting. If you'd rather vote in-bug, feel free to do so - we can count those votes at a meeting if one occurs.
Thanks!
On 05/13/15 06:11, Adam Williamson wrote:
Hi, folks! We're now frozen for F22 Final, but a pile of blocker and FE nominations arrived after the review meeting, several with updates attached. It would be good to get these reviewed before next week.
Especially of note, we have an FE request for https://bugzilla.redhat.com/show_bug.cgi?id=1211015 which would necessitate pulling in the KDE Frameworks 5.10.0 release:
https://admin.fedoraproject.org/updates/kf5-plasma-5.10.0-2.fc22
which is obviously a significant change to the KDE spin. We also have GNOME 3.16.2 pending; they haven't filed that yet, but I'm expecting it today or tomorrow. If we're going to pull in a pair of significant desktop updates, we should do it as soon as possible. I really wish we could somehow stop having these situations where the desktop teams want to push big updates right after freeze points, as it seems to keep happening, but for right now we need to make a decision.
So, I'd like to get sufficient votes on the proposed blockers and at least the significant proposed FEs as soon as possible. Would folks prefer voting on-list or in-bug, or having a special meeting, perhaps tomorrow (05-13) or Thursday (05-14) at the usual time (1600 UTC)? Please reply if you'd be willing to attend a special meeting. If you'd rather vote in-bug, feel free to do so - we can count those votes at a meeting if one occurs.
If, and I say if since I've not tested, this fixes the session restore issues of Plasma 5 then to me it is a *must* for F22 release.
(This comment may appear in multiple places owing my strong view on this matter.)
On Wed, 2015-05-13 at 06:44 +0800, Ed Greshko wrote:
On 05/13/15 06:11, Adam Williamson wrote:
Hi, folks! We're now frozen for F22 Final, but a pile of blocker and FE nominations arrived after the review meeting, several with updates attached. It would be good to get these reviewed before next week.
Especially of note, we have an FE request for https://bugzilla.redhat.com/show_bug.cgi?id=1211015 which would necessitate pulling in the KDE Frameworks 5.10.0 release:
https://admin.fedoraproject.org/updates/kf5-plasma-5.10.0-2.fc22
which is obviously a significant change to the KDE spin. We also have GNOME 3.16.2 pending; they haven't filed that yet, but I'm expecting it today or tomorrow. If we're going to pull in a pair of significant desktop updates, we should do it as soon as possible. I really wish we could somehow stop having these situations where the desktop teams want to push big updates right after freeze points, as it seems to keep happening, but for right now we need to make a decision.
So, I'd like to get sufficient votes on the proposed blockers and at least the significant proposed FEs as soon as possible. Would folks prefer voting on-list or in-bug, or having a special meeting, perhaps tomorrow (05-13) or Thursday (05-14) at the usual time (1600 UTC)? Please reply if you'd be willing to attend a special meeting. If you'd rather vote in-bug, feel free to do so - we can count those votes at a meeting if one occurs.
If, and I say if since I've not tested, this fixes the session restore issues of Plasma 5 then to me it is a *must* for F22 release.
Ditto.
poc
On Wed, 2015-05-13 at 00:26 +0100, Patrick O'Callaghan wrote:
If, and I say if since I've not tested, this fixes the session restore issues of Plasma 5 then to me it is a *must* for F22 release.
Ditto.
Session restore doesn't seem like a great FE candidate, since people don't usually use sessions on live images. Post-install issues like that can be handled fine with updates; the update will be available as a 0-day if karma'ed.
On 05/13/15 07:45, Adam Williamson wrote:
On Wed, 2015-05-13 at 00:26 +0100, Patrick O'Callaghan wrote:
If, and I say if since I've not tested, this fixes the session restore issues of Plasma 5 then to me it is a *must* for F22 release.
Ditto.
Session restore doesn't seem like a great FE candidate, since people don't usually use sessions on live images. Post-install issues like that can be handled fine with updates; the update will be available as a 0-day if karma'ed.
I suppose I may have mis-understood the question. As I didn't read this as being "live image" centric.
I generally don't use/test the Live images all that much. But, if folks use the Live images as a way of deciding if they will use the release and if the Live image supports logging out/in.....then I still feel it is a must.
One question. Would a netinstall eliminate the need to install and then update to get the fix?
On Tue, 2015-05-12 at 16:45 -0700, Adam Williamson wrote:
If, and I say if since I've not tested, this fixes the session restore issues of Plasma 5 then to me it is a *must* for F22 release.
Ditto.
Session restore doesn't seem like a great FE candidate, since people don't usually use sessions on live images.
Anyone who uses the live image to test the new release, and after installing it discovers that sessions don't restore, is going to be pretty unhappy. Stock up on flame-proof jackets.
poc
On 05/12/2015 05:53 PM, Patrick O'Callaghan wrote:
On Tue, 2015-05-12 at 16:45 -0700, Adam Williamson wrote:
If, and I say if since I've not tested, this fixes the session restore issues of Plasma 5 then to me it is a *must* for F22 release.
Ditto.
Session restore doesn't seem like a great FE candidate, since people don't usually use sessions on live images.
Anyone who uses the live image to test the new release, and after installing it discovers that sessions don't restore, is going to be pretty unhappy. Stock up on flame-proof jackets.
poc
After installing they can just update to get a fix.
On 05/12/2015 05:53 PM, Ed Greshko wrote:
On 05/13/15 07:45, Adam Williamson wrote:
On Wed, 2015-05-13 at 00:26 +0100, Patrick O'Callaghan wrote:
If, and I say if since I've not tested, this fixes the session restore issues of Plasma 5 then to me it is a *must* for F22 release.
Ditto.
Session restore doesn't seem like a great FE candidate, since people don't usually use sessions on live images. Post-install issues like that can be handled fine with updates; the update will be available as a 0-day if karma'ed.
I suppose I may have mis-understood the question. As I didn't read this as being "live image" centric.
I generally don't use/test the Live images all that much. But, if folks use the Live images as a way of deciding if they will use the release and if the Live image supports logging out/in.....then I still feel it is a must.
The reason it is critical for release is because we don't release updates for the live images - once shipped, that's it.
One question. Would a netinstall eliminate the need to install and then update to get the fix?
If you did a netinstall with the updates repo enabled, then yes it would be a one step process. But even a two step - install then update, is acceptable since we expect people to be applying updates regularly.
On Wed, 2015-05-13 at 07:53 +0800, Ed Greshko wrote:
On 05/13/15 07:45, Adam Williamson wrote:
On Wed, 2015-05-13 at 00:26 +0100, Patrick O'Callaghan wrote:
If, and I say if since I've not tested, this fixes the session restore issues of Plasma 5 then to me it is a *must* for F22 release.
Ditto.
Session restore doesn't seem like a great FE candidate, since people don't usually use sessions on live images. Post-install issues like that can be handled fine with updates; the update will be available as a 0-day if karma'ed.
I suppose I may have mis-understood the question. As I didn't read this as being "live image" centric.
Basically, things that can't be fixed with updates are the prime candidates for blocker/FE status. So, installer bugs, bugs that are visible when booted live, and kernel/X/etc. showstoppers which would prevent you getting an install done.
Issues that are apparent only after install *can* be blocker/FE bugs, but there's a higher bar for it.
I generally don't use/test the Live images all that much. But, if folks use the Live images as a way of deciding if they will use the release and if the Live image supports logging out/in.....then I still feel it is a must.
You *can* log out / log in on the live image, but it's not something that's very commonly used, I don't think.
One question. Would a netinstall eliminate the need to install and then update to get the fix?
A default netinstall would, yes - by default network installs use the updates repository as a package source. You can disable it, but that takes effort.
Immediately after installing from the live image the bug would be present, but it'd be solved on the first update.
On 05/13/15 11:27, Adam Williamson wrote:
On Wed, 2015-05-13 at 07:53 +0800, Ed Greshko wrote:
On 05/13/15 07:45, Adam Williamson wrote:
On Wed, 2015-05-13 at 00:26 +0100, Patrick O'Callaghan wrote:
If, and I say if since I've not tested, this fixes the session restore issues of Plasma 5 then to me it is a *must* for F22 release.
Ditto.
Session restore doesn't seem like a great FE candidate, since people don't usually use sessions on live images. Post-install issues like that can be handled fine with updates; the update will be available as a 0-day if karma'ed.
I suppose I may have mis-understood the question. As I didn't read this as being "live image" centric.
Basically, things that can't be fixed with updates are the prime candidates for blocker/FE status. So, installer bugs, bugs that are visible when booted live, and kernel/X/etc. showstoppers which would prevent you getting an install done.
Issues that are apparent only after install *can* be blocker/FE bugs, but there's a higher bar for it.
I generally don't use/test the Live images all that much. But, if folks use the Live images as a way of deciding if they will use the release and if the Live image supports logging out/in.....then I still feel it is a must.
You *can* log out / log in on the live image, but it's not something that's very commonly used, I don't think.
One question. Would a netinstall eliminate the need to install and then update to get the fix?
A default netinstall would, yes - by default network installs use the updates repository as a package source. You can disable it, but that takes effort.
Immediately after installing from the live image the bug would be present, but it'd be solved on the first update.
OK.... I looked at the link
https://admin.fedoraproject.org/updates/kf5-plasma-5.10.0-2.fc22
and it seems quite a few packages are updated for this. Is there a repo somewhere with all the packages so that they can be tested now? I don't think they are in updates-testing....at least not when I checked last.
On Wed, 2015-05-13 at 11:53 +0800, Ed Greshko wrote:
On 05/13/15 11:27, Adam Williamson wrote:
On Wed, 2015-05-13 at 07:53 +0800, Ed Greshko wrote:
On 05/13/15 07:45, Adam Williamson wrote:
On Wed, 2015-05-13 at 00:26 +0100, Patrick O'Callaghan wrote:
If, and I say if since I've not tested, this fixes the session restore issues of Plasma 5 then to me it is a *must* for F22 release.
Ditto.
Session restore doesn't seem like a great FE candidate, since people don't usually use sessions on live images. Post-install issues like that can be handled fine with updates; the update will be available as a 0-day if karma'ed.
I suppose I may have mis-understood the question. As I didn't read this as being "live image" centric.
Basically, things that can't be fixed with updates are the prime candidates for blocker/FE status. So, installer bugs, bugs that are visible when booted live, and kernel/X/etc. showstoppers which would prevent you getting an install done.
Issues that are apparent only after install *can* be blocker/FE bugs, but there's a higher bar for it.
I generally don't use/test the Live images all that much. But, if folks use the Live images as a way of deciding if they will use the release and if the Live image supports logging out/in.....then I still feel it is a must.
You *can* log out / log in on the live image, but it's not something that's very commonly used, I don't think.
One question. Would a netinstall eliminate the need to install and then update to get the fix?
A default netinstall would, yes - by default network installs use the updates repository as a package source. You can disable it, but that takes effort.
Immediately after installing from the live image the bug would be present, but it'd be solved on the first update.
OK.... I looked at the link
https://admin.fedoraproject.org/updates/kf5-plasma-5.10.0-2.fc22
and it seems quite a few packages are updated for this. Is there a repo somewhere with all the packages so that they can be tested now? I don't think they are in updates-testing....at least not when I checked last.
Not until the updates-testing push, no. You can use 'bodhi -D' to download the update up until then.
On 05/13/15 14:48, Adam Williamson wrote:
Not until the updates-testing push, no. You can use 'bodhi -D' to download the update up until then.
I am sad to report that this update did *not* fix my problem....
I am using "folder view" as my desktop settings and I'm configured to restore from a "saved session". I do the following....
1. login 2. start konsole, place it in the lower right edge of the screen and set it to "all desktops". 4 virtual desktops in use. 3. start tbird, place in desktop 2 and align it to the left edge of the screen with a width which brings it into contact with the left edge of konsole. 4. Go to "leave" and click on "Save Session" and the following files get modified...
[egreshko@f22-tc3 ~]$ find . -cmin -1 ./.xsession-errors ./.config ./.config/ksmserverrc ./.config/session ./.config/session/kmix_1028c1d320b210000143113812800000013960009_1431501798_448141 ./.config/session/konsole_1028c1d320b210000143122570000000015410009_1431501798_463485 ./.cache/plasma_theme_polishedStones_v4.11.kcache
5. Logout/Login 6. konsole and tbird are started. But they are both on Desktop 1 and they are out of position.
On Tue, 2015-05-12 at 19:53 -0600, Orion Poplawski wrote:
On 05/12/2015 05:53 PM, Patrick O'Callaghan wrote:
On Tue, 2015-05-12 at 16:45 -0700, Adam Williamson wrote:
If, and I say if since I've not tested, this fixes the session restore issues of Plasma 5 then to me it is a *must* for F22 release.
Ditto.
Session restore doesn't seem like a great FE candidate, since people don't usually use sessions on live images.
Anyone who uses the live image to test the new release, and after installing it discovers that sessions don't restore, is going to be pretty unhappy. Stock up on flame-proof jackets.
poc
After installing they can just update to get a fix.
I guess what I'm really asking is why this bug, which was reported over three months ago and is very obvious to anyone who tries to use session-restore, is still present in a release candidate.
poc
On 05/13/2015 04:39 AM, Patrick O'Callaghan wrote:
I guess what I'm really asking is why this bug, which was reported over three months ago and is very obvious to anyone who tries to use session-restore, is still present in a release candidate.
poc
I think the obvious answer is the only one you'll get - because no one has yet manged to (perhaps) figure out the cause and done the work to fix it.
Upstream KDE development is not beholden to Fedora's release schedule as far as I know. In general Fedora just does its best to pick the right upstream release to target for each Fedora release.
Adam Williamson wrote:
desktop updates, we should do it as soon as possible. I really wish we could somehow stop having these situations where the desktop teams want to push big updates right after freeze points, as it seems to keep happening, but for right now we need to make a decision.
IMHO, this points out two things that are wrong with the freeze process:
• Why is the decision whether to take updates to packages only affecting a given desktop environment not up to that desktop environment's maintainers? Those KF5 changes are basically only going to affect our spin. (Well, there's one or two "Labs" spins derived from the KDE spin, but those aren't even release-blocking.) Likewise, the GNOME megaupdate basically only affects Workstation. I see the need for a freeze process for core shared components, but for the spins*, IMHO, both the decision what packages to take and the go/no-go call should be up to the respective spin* maintainers. (That also means that the complex desktop criteria could be dropped entirely (replaced by an executive decision by the spin* maintainer) or moved to the individual spin*'s area, freeing you from the burden of maintaining them.) And if that means we have to wait for a few days for GNOME/Workstation to be ready or the other way round, so be it, it is not going to kill anybody. (Or if that really is a problem for some people, we could start looking into per-spin* releases. But I don't see the need there.)
* … When I say "spins", I mean both Products/Flavors and Spins. (I still do not approve of the arbitrary distinction between first-class and second-class spins, and no such distinction is needed in this context for sure.)
• We have exactly 2 release-blocking desktops. If both want to get some important update in right after the freeze, why can't we just delay the freeze (and accordingly, the release), by a week? I think some more flexibility would really help us there.
Kevin Kofler
-
Adam Williamson -
Ed Greshko -
Kevin Kofler -
Orion Poplawski -
Patrick O'Callaghan