Fixes issues with the docker images build #1663
Conversation
|
@Emilgardis Would you mind giving a review and tell me if you need me to change anything? |
alexis-opolka
changed the title
|
/ci try |
This comment has been minimized.
This comment has been minimized.
|
Thank you! I've left some comments :) |
I saw that XD |
|
I have tried my best to answer your comments. ^-^ |
|
We really need to find out why the macos target doesn't run at all. |
This comment has been minimized.
This comment has been minimized.
|
Canceled the try run. I remember something about macos :latest now being apple silicon, that might be the reason |
|
/ci try |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
|
Unfortunately I do not have the power of maintainers to trigger the mighty CI bot. |
|
/ci try |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
|
@Emilgardis What do we do with the centos version? From the 404 errors that I get, you're still with centos 7 which has been deprecated along with centos 8. Do I try to update the centos versions to centos 9 stream or do I redirect the URLs to the vault mirror considering we'll still build for EOL releases? |
|
the centos images exists for users that really want old versions, so yes, it should use an archive/vault. Preferably, in the future, we'd just tarball the centos image and just unpack it in newer versions, but that's a bit annoying. Honestly I'd be ok with making the centos dockerfiles into |
|
I added the vault mirror for now, it seems to have fixed this issue, at least. |
|
/ci try |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
|
@Emilgardis Do you have an idea why on FreeBSD tar cannot unpack the I'm testing with the x86_64 target, btw: cargo xtask build-docker-image -v "x86_64-unknown-freebsd"(Sorry to ping you on a sunday 👀 ) |
|
The error suggest bad flags, what was wrong with what was there initially? |
|
Initially, there was this error: 11.67 + curl --retry 3 -sSfLO https://pkg.freebsd.org/FreeBSD:13:amd64/quarterly/packagesite.txz
24.03 + tar -C /opt/freebsd-packagesite -xJf packagesite.txz
24.04 xz: (stdin): File format not recognized
24.04 tar: Child returned status 1
24.04 tar: Error is not recoverable: exiting now |
|
I tried to dig around a bit but I'm always stuck with some kind of file format not recognized error or something related. And I don't really know what to do anymore to fix this issue. |
|
I stop on trying to work on this specific issue, I need to do some other work. |
|
What is weird is that the aarch64 target doesn't have this issue but |
|
btw, this pr tracks your main branch, which means that a rebase will likely screw up the PR and autoclose it, going to leave this here how to solve that. https://gist.github.com/robertpainsi/2c42c15f1ce6dab03a0675348edd4e2c another way is to on the remote repo change the head-branch to something else |
What? 😕 I don't understand what you're talking about, I'm not going to do a rebase. Why would I need to? |
once this pr gets in a good space, I'm going to ask for a force-push to squash most changes, but I know from experience that it doesn't work well when the PR is made from the main branch. |
|
Can't you just squash the PR when merging it? It's easier. I don't think the issue is because the PR is made from a certain branch, I think the issue comes from the people doing it. |
|
I like to use merge commits for PRs, I guess I could change the setting for this PR only. Problem is that merge queues only work with one method, not multiple |
|
Oh! Now I understand why, well, I'll squash my commits when it will be good so no worries. ^^ |
|
/ci try |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
|
Hi @Emilgardis , |
|
No worries, let me know if you need any help or feel satisfied with the state currently. You don't have to fix everything in one pr |
|
I'm going to try to fix the most I can but I won't fix the issues which will require me to go quite deep in the debugging (such as the current state of the freebsd issue) |
|
Freebsd has been solved in #1674 |
|
Oh, that's nice to hear! |
|
the branch is currently not properly synced, just fyi, there's changes in d37706e which repeats what has happened upstream |
|
Thanks, I saw that it wasn't well synchronised but I wanted to first incorporate all changes made in upstream on my fork and have a limited number of commits so I squashed everything. I will look at the tree for the changes to be synchronised. |
|
Little by little, I'm having less builds that fails! 😄 |
|
Hi @Emilgardis ! Can you trigger a CI run? On my side, All the failing CI runs happens during the Test Image step after the docker build. And I don't know if there aren't PRs opened for them or not so I'll wait for this PR to be merged and see If I should open another to fix the Tests from failing. |
|
/ci try |
This comment has been minimized.
This comment has been minimized.
|
They are all failing the test image step. Do you want me to investigate it further in this PR or do I squash all my commits to be ready for the merge? |
|
Lets merge this and continue the work separately, thank you! Please do a squash :) |
Signed-off-by: alexis-opolka <[email protected]>
|
Squash done, you can review it but it seems to me the tree has been correctly fixed. 😄 |
This PR should fix most issues encountered when building docker images.
As a reminder of the docker build images currently failing: