I'm not sure if my opinion is helpful, but that's how I see it.
Or: Start discussing the release blockers here and now. Thanks.
1) Blocker: LuCI master needs to be backported to 19.07
Time estimate: 2-3 weeks
I think it's good that LuCI has now outsourced more work to the client.
Thanks to Jo-Philipp to do this work (great job) but I think it needs
more testing time.
So it should go into next release maybe 20.03?
2) Blocker: All relevant sub-components for WPA-3 + GUI support, such
as
hostapd, iwinfo etc. need to be backported to 19.07
Time estimate: 2 weeks
should go into next release maybe 20.03 because this is neede by LuCI
3) Blocker: Some weaknesses in libustream-ssl client certificate
handling need to be addressed, which can only be solved by an API
redesign. Band-aid fixes available but not merged, nobody worked
on API redesign yet
Time estimate: 1 week
should be done before 19.07 release
4) Blocker: Need to assert the state of the Dragonblood WPA3
vulnerabilities in 19.07's hostapd
Time estimate: a few days I guess
should go into next release maybe 20.03 maks only sense in master.
Question: How long would it take to stabilize/test a fresh fork from
today's Master branch instead? I assume none of these issues apply to
master or need to be solved in master as well?
There are always changes in the master that may need to be backported to
a stable branch.
But since the master is constantly evolving a stable will never be
released if we keep going like this.
I think the release will delayed even more.
It is clear to me that there is more effort in the LuCI to fix the bugs,
that there were big change in the master -> switching to Javascript.
There is always advantage and disadvantage.
How to do it may be wrong. This is life.
Kind regards
Florian
_______________________________________________
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
