= phenomenon = As per the updates policy (http://fedoraproject.org/wiki/Updates_Policy), package maintainers are discouraged from making major updates to stable releases. The desktop SIG would like to request an exception to this policy so that they can update Gnome to 3.12 in Fedora 20. The current version in F20 is 3.10.x
= reason = Gnome 3.12 releases towards the end of March or early April. There's still no schedule for F21, though. If Gnome isn't updated in F20, a Gnome stable release will be skipped in Fedora. This also implies that users will not receive the improvements and bugfixes that have been made in the Gnome 3.12 cycle.
The discussion on the desktop list can be found here: https://lists.fedoraproject.org/pipermail/desktop/2014-January/008741.html
= recommendation = Please permit an update to Gnome 3.12.
The alternative is to provide it via a COPR repository. However, that wouldn't permit use of bodhi for QA.
Replying to [ticket:1235 ankursinha]:
= reason = Gnome 3.12 releases towards the end of March or early April. There's still no schedule for F21, though. If Gnome isn't updated in F20, a Gnome stable release will be skipped in Fedora. This also implies that users will not receive the improvements and bugfixes that have been made in the Gnome 3.12 cycle. The discussion on the desktop list can be found here: https://lists.fedoraproject.org/pipermail/desktop/2014-January/008741.html
I haven't yet read through the thread - so these comments are preliminary, and the questions are not necessarily directed at the proponents of the change. Skipping a Gnome release is structurally not really a Fedora concern - there's a '''lot''' of software where we skip some, or even most, upstream releases. (Sure, not all software is created equal...) Users will receive the improvements in the next tested release. The question to answer is not whether there are improvements, but whether shipping them now: * Is consistent with user's expectations for the release (... which I'd expect to be primarily shaped by the Updates Policy and the extent it's been followed in the past). So, is there precedent? Are there other sources for such expectations? * Is going to introduce unwanted breakage. How much breakage has there been in the past? How much has Fedora QA been involved in testing GNOME before releases? If it has been non-trivial, would the proposed release go through a comparable testing effort? * Desktop SIG, can you quantify the impact at least a little? How many packages are we talking about (or even better, which ones precisely?), how many ABI breaks (and necessary rebuilds of "non-GNOME"), how much UI change?
I'm running Gnome in Rawhide right now, and I assume this is a decent preview. It's my impression that the UI changes are fairly subtle.
However, all of my extensions immediately stopped working, because each has a whitelist of versions that it's been tested with. The good news is that with a few exceptions, the extension actually just works once the version is updated; the abi hasn't really changed enough to break most things. (At least not yet.)
I think updating Gnome is good, but I think if we want to do this, we should smooth the extension transition as much as possible. Maybe go through https://extensions.gnome.org/#shell_version=all&sort=popularity in order and work with the extension authors so, say, the 20 most popular extensions work?
A few highlights (many FESCo members are already involved in the desktop SIG conversation, so mostly for the record):
AGREED: defer this for now, revisit when sig asks us to do so (+8,0,0) (nirik, 18:49:22)
Please reopen this when closer to 3.12 release and wishing to push into f20.
Login to comment on this ticket.