openwrt / routing

OpenWrt Routing Packages
330 stars 370 forks source link

olsrd: update to 2023-06-12 #977

Closed PolynomialDivision closed 1 year ago

PolynomialDivision commented 1 year ago

Update to latest version.

Maintainer: me Compile tested: ramips Run tested: ramips

PolynomialDivision commented 1 year ago

I'm getting crazy with the CI. I don't understand why the PKG_MIRROR_HASH should be different for some targets. I tested that now multiple times on my machine and I will try to merge this now and see if it will still build.

BKPepe commented 1 year ago

Thats very odd behavior and weird decision to blindly merge this. Why would you do that if this is here only for 5 Hours? Be patient

PolynomialDivision commented 1 year ago

Thats very odd behavior and weird decision to blindly merge this. Why would you do that if this is here only for 5 Hours? Be patient

As you can see the current state of the PR also just fails on some targets and on some not (okay it fails for all, but that is because of the gpsd fix not included). I just added +1 day to the PKG_SOURCE_DATE. Some accept the PKG_MIRROR_HASH and some not. So the failing CI is not a result of the PR.

PolynomialDivision commented 1 year ago

I strongly believe it is caused by: https://github.com/openwrt/openwrt/commit/75ab064d2b38e70746af1718ed7cdbafb906249e

There was already some other fix: https://github.com/openwrt/openwrt/commit/5f1758ef14575df4e86896526b1c2035c231899e

PolynomialDivision commented 1 year ago

image

PolynomialDivision commented 1 year ago

Okay, the tar is not working correctly on my host system. That was the error. I will merge this and than again cherry-pick latest CI things.

BKPepe commented 1 year ago

Hm. 😢 You should really wait until CI is finished if you are really unsure what are you doing or what was wrong.

PolynomialDivision commented 1 year ago

Hm. cry You should really wait until CI is finished if you are really unsure what are you doing or what was wrong.

The CI was already finished for the PR. I just deleted the rebase on the latest ci changes. Look into the IRC, there you can find the explanation what was wrong. It now builds fine except on my local machine, however, only if not already the tar.xz is uploaded to openwrt mirror.

BKPepe commented 1 year ago

Look into the IRC, there you can find the explanation what was wrong.

That's on you here to add a reference to the IRC history. If you notice correctly, I have not been using IRC for a few months already and have not even connected there. There are many channels to use - GitHub, IRC, mailing lists, OpenWrt forum, etc. Too many ways. For contributors, it is easier to use GitHub. Just my two cents.

Maybe in the future, I will start using IRC again.

PolynomialDivision commented 1 year ago

@BKPepe My host system is somehow corrupt. When it packages the source to tar.xz something fails. @jow- said that the checksum field differs between the "openwrt docker version" and my local build machine. That's why I just used the PKG_MIRROR_HASH of the "docker build" and the CI was fine with it. However, I have no idea why I am the only one having those issues.

Important code: