I really had no plan for where best to express this information, so here's as good a place as any.
TL;DR: It's not ready yet.
First, some definitions to make sure we're on the same page:
- Roughly speaking, anything that is merged into a
xenial branch on a Debian-style package in the https://github.com/ubports organization gets built and released to devel. This is what we mean when we say "edge merge" or anything similar.
- We expect ourselves to ship a release of Ubuntu Touch every 6-8 weeks, including 2 weeks for testing. This gives us 4-6 weeks of "open merge" time in a cycle where low to critical priority changes may be merged. During the 2-week testing period that changes to merging critical priority changes only.
- Releases of Ubuntu Touch move sequentially from devel to rc to stable. That is, Stable OTA-10 is the exact same set of bits as rc
2019-W33/2, which is the exact same set of bits as devel
This is the only way we can release software. Changes we merge into a package go into devel which must go to rc which must go to stable. We cannot easily or cleanly substitute in a certain old version of a package, for example. If we merge broken changes, our options are to revert those changes or fix them so they are not broken. That was the cause of OTA-10's delay: we merged some improperly tested changes and believed we could fix them faster than we could revert them. This proved to be false.
To make it crystal clear, if something is broken in devel, it will be broken in the upcoming stable.
I'm fairly sure that it will take more than a week to merge all of the changes from
xenial, get the packages built correctly, and get the release pipelines happy again. This takes anywhere from 1/4 to 1/3 of our expected cycle time, reducing the number of bugs we can expect to fix before we can release again.
I firmly believe "if you add a little to a little, you have a big pile" (attributed to Hesiod). Check out https://github.com/orgs/ubports/projects/10. It looks like a small pile of things. However, we know that, as very early adopters of the edge software, we are imperfect testers. We aren't going to catch every regression from devel to edge. That's why we release software to devel. Lots of people testing means more bugs filed and higher quality software. In this case, that would add a little more to our little pile. Suddenly it doesn't look like such a little pile of bugs any more.
We don't want to end up with that rush or crunch. It means lower-quality software for you. Our goal is to have the number of known bugs in edge to a minimal point at the start of a cycle so we can take our 1-2 weeks to merge from
xenial. Then we can fix the myriad bugs that people will find, because they're better testers than us.
If you want to help us, try out the edge channel and report anything strange you find. This might further delay the merge, but it will ensure a smooth transition and a faster release when the merge does happen.