What blocks 19.07 release?

checked the release status every week in last months but hopefully this one will not die in the womb

1 Like

Happy to see the Devs talking about 19.07 moving forward. I agree with Martin and Alberto: no more new features, just bug fixes and security updates. Get 19.07 out the door.

What @jeff said is pretty much the current status, I'm not aware of any functional or regressing testing being done in general although due the amount of "supported" devices/platforms it's not possibly to test all. Most targets are on 4.19 while 19.07 is on 4.14 and since most contributors are running master there's little interest in doing backports especially when 19.07 is so far behind and on another kernel version and that includes libraries etc. Different release plans have been mentioned before but have been turned down in favour of the current release plan mainly due to longterm support.

The package repo is probably in even worse shape at least in terms of security since it was branched a long time ago with most packages being outdated and branches such as Python 2.X nearing EoL.

Would you say Master is stable for production?

I have been running builds off master exclusively for many years on several different devices.

Edit: To be clear, I do suggest building your own from source so that packages are readily available through a new, local build, as well as being able to revert to "yesterday" for the rare times that master has unresolved issues that impact your device and/or use cases. (I can only think of one such time over the last year or two, related to batman-adv, 802.11s, and older wireless chip sets.)


So do all activists, and that is the problem with stable releases. Nobody has any real interest in polishing the "new forthcoming release" like 19.07, as the master has already some additional new features, right after the branching.

When the release branch is branched off master, it is fresh and up-to-date, but it soon falls behind. The longer it takes to publish the release binaries, the staler (more stale?) the release is.

The goal of the LEDE fork was to do a release every six months, but since the merge back to OpenWrt it looks like we have fallen back to the old "one release every two years" regime that was the sin of the old OpenWrt in 2010-2015.

We will never be able to perfect all these nice features for the next version, so we should accept that there are new, small features in every frequent release. And do a new release more frequently, like e.g. the major browsers do.

19.07 was in pretty ok shape in June, but as there are apparently no eager "release managers" among the core developers at the moment, the production of even rc binaries for 1907 has stalled, despite the branching in June and a buildbot crunching images and packages daily.

The release manager used to be @blogic a few years ago, then @stintel for a few releases, but right now @jow has pretty much been the only one to show activity regarding the 19.07 release. The discussion on the devel mailing list is heating up...

And as there is now talk about backporting some new features (javascript LuCI, wpa3) from master to 19.07, the process may take even longer...

It would have been nice to have 19.07 with the June 2019 features already released, and then a new 19.12 / 20.01 with the newest stuff as the next release. But no idea what will actually happen.


I'm doing lots of testing on 19.07 as its being used here internally atm.

Its running on roughly 100 devices in the field covering multiple archs, kernel, wifi and LTE.

The devices get updated weekly after an important bump of kernel or drivers.

So I'm doing all I can to keep it from becoming stale :slight_smile:


In general it's fine however you may run into breakage occasionally so I'd suggest that you test at least on one local device before applying it to all. Irregardless of the amount of devices you have a copy of the configuration files and make sure you know how to recover in case it doesn't boot.

1 Like

I've been running master for a long time too but I'd like to point out that actually running a compile task every few months is much easier than actually maintaining and polishing the distro for a release.

Most people compiling firmware from master and even those just using the Image Builder would not be able to help reaching whatever goals someone has set for a stable release anyway, they are just veteran users that got fed up with the bs release model of OpenWrt and work around it.


I've been running the 19.07 for a while now, and when Hauke pointed to it on the mailing list, I applied the hostapd backport patches from ynezz to my local tree, along with a few other tidbits (a more recent Wireguard, also from master). Has been running without any hiccups so far (RT-AC57U).

Reality check:

Name Release Late By Since Last
7.09 2007/09 (none) 5 mo.
8.09 2008/09 (none) 12 mo.
10.03 2010/04 1 mo. 19 mo.
12.09 2013/04 7 mo. 36 mo.
14.07 2014/10 3 mo. 18 mo.
15.05 2015/09 4 mo. 11 mo.
17.01 2017/01 (none) 16 mo.
18.06 2018/07 1 mo. 18 mo.
19.01 ? >10 mo. >16 mo.

Yes, it was supposed to be 19.011

Even if you call it "19.07", it's going to be at least 4 months late and 16 months since the last release.


Edit: For those watching for progress, there is a PR for tagging RC0


Be that as it may, it has been branched and it still has newer code than the latest 18.06 branch. You've seen in the mailing list what's holding a release back.

Grate news! Lets go with a RC1

Why bother? It’s already obsolete by over 1000 commits and about to become even more so.

should we use v5.4 as our next kernel ?

they say it will be supported till Dec, 2021 (so 6-year LTS myth dies). to hell with it, use stable version and switch it everry few years

Who is saying this where?


The party in Hamburg decided that releases were every six months, so some would be on non-LTS kernels.

But then, 2020.01 is coming up fast and a branch hasn’t been cut yet. Nor have PRs been reviewed on a timely basis. The current thread on PRs on the mailing list has been all about hiding the symptoms, not addressing the problems. Worse is that it is a fox guarding the henhouse situation.


for those wondering, this was written down in the wiki, in an unaccessible page afaik. https://openwrt.org/meetings/hamburg2019/start

That said, it seems most of that is fictional, the core team does not move fast enough to deal with all that in this decade.


That page has no access restrictions at all, everybody can read, even anonymous users.

Edit:...and can easily be found via the wiki search.

1 Like

Quite visible page...