-
Notifications
You must be signed in to change notification settings - Fork 154
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add and upgrade to ruby 3.0 #316
Conversation
Can one of the admins verify this patch? |
2 similar comments
Can one of the admins verify this patch? |
Can one of the admins verify this patch? |
Check failed on local fedora container test with the following outputs: |
Blocked by: sclorg/s2i-base-container#203 |
better to have them joined, like so:
Ex.
Or if you want to have a terminal copy as well:
From what I've gathered, there's a lot of failed tries to write into And ...
Could you create a reproducer for that? / And file it upstream / into Fedora etc. (where suitable). |
In my attempt, I've also noticed possible issue- the tests try to use already existing Fedora container- could you investigate? https://git.io/Jqk7h EDIT: yes, it seems the app fails to build using The error output is the same though: https://git.io/JqkF5 |
to make the test suite more user-friendly. Related: PR #316.
to make the test suite more user-friendly. Related: PR #316.
|
@jackorp can the TMP dir be created locally, in the meantime? IOW in the app-dir. |
I am not sure what do you want to achieve. Do you mean to basically override the bundler option? |
Yes. I think it's |
Hmm, it seems like that won't be of help. My wild guess is that bundler needs a temporary directory when installing (for whatever reason) and the |
Have you tried? Maybe this should be an upstream issue. |
I did try, but in the end, it looks like it does not matter a whole lot for us. More tests ran than before, and even though /tmp is not writeable it installed the gems and did some tests successfully: see log There are messages about /tmp not writeable (that error is from Ruby's tmpdir library) on this line. However, the gems get installed which I think is what matters for us. |
Sorry, I got the impression that these are bundler errors which cause some of the tests to fail. Instead it's just ruby churn. So it could be switched off by something like (FTR: The mustermann issue is blocked on me.) |
> Sorry, I got the impression that these are bundler errors which cause some of the tests to fail. Instead it's just ruby churn.
I had the same impression, however, on closer inspection, I found out that some bundle installs failed because of minitest.
Should work for silencing it, I'll look into it later (if we want to anyway). And with that, I'd like to point out that a test throws an exception with unwritable |
Now this is interesting. Coud you look into that? The app'll probabrly need a sample app dependency refresh. |
I'll look into it. |
So I did a little dependency refresh on the example rails app for Ruby 3.0 and hit a new error with Other than that, the /tmp not writeable (which causes rack fails currently), and mustermann there seem to be no other problems. |
6200233
to
e4f4f15
Compare
Not sure where the culprit lies (maybe you need to rebase?), this is the diff I've landed on while applying your PR: |
yeah there were some things that went wrong when rebasing on newest master apparently... I'll fix that. |
[test] |
@jackorp 08:31:23 CMD exec bundle exec "rackup -P /tmp/rack.pid --host 0.0.0.0 --port 8080"
08:31:23 + readarray -d @ -t git_url_parts
08:31:23 /root/sources/3.0/test/test-lib.sh: line 870: readarray: -d: invalid option
08:31:23 readarray: usage: readarray [-n count] [-O origin] [-s count] [-t] [-u fd] [-C callback] [-c quantum] [array] This was replaced by code: RHEL8 test failed on network issues: STEP 3: RUN bundle install --path ./bundle
Fetching source index from https://rubygems.org/
Network error while fetching
https://rubygems.org/quick/Marshal.4.8/nokogiri-1.11.3-x86_64-linux.gemspec.rz
(execution expired)
Error: error building at STEP "RUN bundle install --path ./bundle": error while running runtime: exit status 17
ERROR: The image cannot be built from /root/sources/examples/from-dockerfile/Dockerfile and application https://github.com/sclorg/rails-ex.git.
Terminating the Dockerfile build. |
If tests are broken, as of now, it is not up to this PR to fix them. I think that change may even render it unable to auto-rebase. It container-common-scripts should be updated outside of this PR.
Is that random / temporary or something that needs to be fixed in the infrastructure first? |
Oh I think the problem is that instead of the latest Workaround: @jackorp maybe there's a rebase button? Solution @phracek maybe that could be enhanced? Instead of running the
|
RHEL and CentOS currently don't have Ruby 3.0 available.
Since Ruby 3.0 Webrick has to be installed manually because it is not part of the standard library.
I have rebased the PR branch. |
[test] |
Yes, I know, that was apparent from your initial text. my comment was relevant to that. Please give it a go, so we can have a permanent solution, avoiding future PR rebases, due compatible upstream changes, just for testing purposes.... |
The standard SCLorg solution is simple. We do not want CI to clone But in the meantime, I am working on |
The whole argument is moot, as I've realized that the problem is more serious, and does not affect only the common scripts, and it is a bug. (And it definitely is not "best".) I'll file an issue for that.
Great! |
Let's get this merged and we can fix the remaining tests later on (sclorg/rails-ex#134). |
[test] |
Hi @pvalena , and what about the problem with network issues in RHEL8? |
First test output says it clearly: |
Please note that Dockerfiles in this PR are either way used in production right now (having them merged or not does not change that). What's more, the longer this PR needlessly hangs (it's not broken!), the more changes are made to the other Dockerfiles which this PR will need to add. It's like maintaining a fork, consuming time and it brings no benefit to this project not to have these new Dockerfiles (i.e. this PR does not break anything). |
[test] |
This PR upgrades the s2i to ruby 3.0.
I have not run any tests yet because of missing dependencies.