June 11 Stable Release

just curious why there is no mvebu targets?

https://downloads.lede-project.org/releases/17.01.2/targets/

The good stuff takes time:wink: I think things are built in a random order, and will take a couple of days to complete.

perfect!

waiting impatiently ! :grinning:

Is this a matter of funding or setup?

Is it possible to spin up cloud instances at release time to build the releases
faster?

How useful would it be to have more systems building all the time (and how much
would it cost to make a difference)?

It seems to me that releases should be able to build built in less than a day
(I'd aim for 6-12 hours), and normal test builds should be no more than two days
(ideally at least one complete build/day)

David Lang

There have been three release since Feb this year (over 3-4 months) is it really an issue if the build takes a day or 6 hours to complete?
Whilst faster can be nicer I'm not sure that speeding up builds is up there on the priority list at the moment.

well, if a build takes several days, it reduces the quality of the testing
(since it's going to include more changes for every build)

For releases, you run into the impatence and expectations of users (exactly what
started this thread)

If it's a matter of distracting developers to make thins build faster, that may
not be worth it.

But if it's just a matter of money, you may find that users are interested
enough in having the builds be faster to put their money towards it.

David Lang
k

https://lede-project.org/infrastructure

so per this, the project wanted two additional big servers several months ago,
and the existing servers are being funded personally by developers.

how can non-developers put money in?

First off, is this something we should wait until the remerge and then run the
money through SPI?

I suppose I could just rent a server and then give you guys access to it, but it
seems like it would be better to send you the money and let you folks rent the
servers

I'm just a happy user of the firmware posting up links to the documentation that the developers have put online.
Maybe try this link: https://lede-project.org/contact

Most developers voted for a remerge, so we're stuck with waiting for the OpenWrt side to act until it makes sense to form a legal entity which can accept donations.

That makes sense. How badly do you need an extra server or two?, 50 euro/month
is not that much, I could probably pay it directly to the hosting company if
that would help significantly in the short term.

David Lang

mvebu is up !

thanks very much to the developers.

my original intention was just to see if there was an issue that was causing mvebu to not compile, I am not REALLY impatient, that was an attempt at humour.

LEDE runs the best for me on all my mvebu devices.

If had previous stable build running and have been upgrading packages via opkg regularly, is my current running code the same as in this latest stable release?

No.
Updating packages via opkg does not update the core system e.g. kernel.

2 Likes

I suppose samba is patched against "SambaCry", isn't it?

What is the recommended way to upgrade: preserve configuration and then install missing packages or do not preserve any config, install missing packages, then merge old config into the new config files?

Yes, see https://lede-project.org/releases/17.01/changelog-17.01.2#security_fixes

What is the recommended way to upgrade: preserve configuration and then install missing packages or do not preserve any config, install missing packages, then merge old config into the new config files?

This is the sequence I was led to believe was the best way to do it (please correct me if I'm wrong): 1) user should create an archive; 2) user should update core system (ie, install new LEDE without "Keep files" ticked); 3) user should install packages that are not in the the default LEDE install (e.g., DNSCrypt); 4) user should load the archive they created in step 1.

Doing these 4 steps will put things exactly how they were before the update but with updated kernel, etc.

Most developers voted for a remerge, so we're stuck with waiting for the OpenWrt side to act until it makes sense to form a legal entity which can accept donations.

Too bad. things have been awesome since the split.

so how do you think the remerge will hurt?