Skip to content
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

Merge #3011

Closed
wants to merge 15 commits into from
Closed

Merge #3011

wants to merge 15 commits into from

Conversation

dominicjaeger
Copy link
Contributor

Basics

Check relevant points but please do not remove entries.
Do not describe the purpose of this PR in the PR description but:

  • Short descriptions should be in the release notes (added as entry in
    doc/news/_preparation_next_release.md which contains _(my name)_)
    Please always add something to the the release notes.
  • Longer descriptions should be in documentation or in design decisions.
  • Describe details of how you changed the code in commit messages
    (first line should have module: short statement syntax)
  • References to issues, e.g. close #X, should be in the commit messages.

Checklist

Check relevant points but please do not remove entries.
For docu fixes, spell checking, and similar none of these points below
need to be checked.

  • I added unit tests for my code
  • I fixed all affected documentation
  • I added code comments, logging, and assertions as appropriate (see Coding Guidelines)
  • I updated all meta data (e.g. README.md of plugins and METADATA.ini)
  • I mentioned every code not directly written by me in THIRD-PARTY-LICENSES

Review

Reviewers will usually check the following:

Labels

  • Add the "work in progress" label if you do not want the PR to be reviewed yet.
  • Add the "ready to merge" label if the build server is happy and also you
    say that everything is ready to be merged.

@dominicjaeger
Copy link
Contributor Author

I think I'm pretty close to having all tests successful. Currently cmake throws an error because it does not find two functions from libgit2 in the full build. I looked into the cmake files to find the source for the error, but I'm pretty confused at the moment.

We are not allowed to include libgit2 due to the licenses. In the shared build we link to a existing libgit2 on the system. Would we have to include libgit2 in the full build? But we are not allowed to do this?

@markus2330
Copy link
Contributor

I think I'm pretty close to having all tests successful. Currently cmake throws an error because it does not find two functions from libgit2 in the full build. I looked into the cmake files to find the source for the error, but I'm pretty confused at the moment.

Missing functions sounds like too old libgit2. libgit2 in stretch is quite old (also the gitresolver does not work there). Did you add the version check for libgit2? (see gitresolver how to do it)

We are not allowed to include libgit2 due to the licenses. In the shared build we link to a existing libgit2 on the system. Would we have to include libgit2 in the full build? But we are not allowed to do this?

Also the full build only links against the libgit2, so there is no difference.

@markus2330
Copy link
Contributor

The static build might be a problem for some licenses (like one plugin uses a GPL lib and another one uses CDDL) but the problem of the person who compiles Elektra, not ours.

@dominicjaeger
Copy link
Contributor Author

Did you add the version check for libgit2? (see gitresolver how to do it)

No. But I still don't think it would help.

Missing functions sounds like too old libgit2. libgit2 in stretch is quite old (also the gitresolver does not work there).

As I tried to fix that in between, I pushed the "original" commits that gave me that one libgit error only another time and now other test fail :/

  • "remote file operation failed" for docker
  • timeout when downloading from homebrew
  • markdown tests time out
    ....

Nonetheless, the only two functions I use (git_merge_file and git_merge_file_result_free) have the same parameters through all API versions that are available online. Stretch certainly is part of those versions, too.

@dominicjaeger
Copy link
Contributor Author

While I'm writing this one artifact is still being built on Jenkins while the stretch package build ended with an error again and I think I need some time to figure this out.

Due to the long building times I'm working on the documentation on another local branch in parallel.

@markus2330
Copy link
Contributor

Nonetheless, the only two functions I use (git_merge_file and git_merge_file_result_free) have the same parameters through all API versions that are available online. Stretch certainly is part of those versions, too.

Ok, good. For @tom-wa some functions were not available in earlier versions at all.

Due to the long building times I'm working on the documentation on another local branch in parallel.

Yes, very good idea! We are busy working on migrating to the build server to faster hardware but unfortunately it still takes time.

@markus2330
Copy link
Contributor

The formatting check seems to fail here in a later stage (not in the formatting stage)?

@dominicjaeger
Copy link
Contributor Author

The formatting check seems to fail here in a later stage (not in the formatting stage)?

Yes.

diff --git a/src/plugins/specload/specload/remove.quickdump b/src/plugins/specload/specload/remove.quickdump$           
 index ccedf5000..e69de29bb 100644$
 Binary files a/src/plugins/specload/specload/remove.quickdump and b/src/plugins     /specload/specload/remove.quickdump differ

The only thing I have been changing for the last couple of days is the last commit for libgit linking and all sorts of errors (that I comprehend partly and partly not) seem to randomly appear.

Locally (with run_dev_env) everything compiles and all the tests run. And then

  • one time a single build on jenkins complains about linking problems
  • another time some resources on docker are busy until timeout
  • another time some specload binary makes the format test fail
  • suddenly some of my tests segfault on cirrus

I'm copying the error messages in hope to find patterns and report them if I think it's not my fault. Nonetheless, the last days with the continuous integration have been tough.

@markus2330
Copy link
Contributor

Please report it upstream even if you cannot reproduce it. More than not getting answers cannot happen. You can extend the bug report once you know more.

@dominicjaeger
Copy link
Contributor Author

jenkins build libelektra please

@dominicjaeger
Copy link
Contributor Author

Please report it upstream even if you cannot reproduce it.

Done here.

@markus2330
Copy link
Contributor

Thank you! Also give them version information and on which systems the problem appeared (and where it works).

@dominicjaeger
Copy link
Contributor Author

Also give them version information

I added the Elektra version. Which version do you think is missing?

and on which systems the problem appeared (and where it works).

I don't know what you mean exactly. What fails is our build artifacts and all except the debian buster build are website related. And the Debian buster build is what I mentioned in the issue?

@markus2330
Copy link
Contributor

Which version do you think is missing?

Of libgit2.

And the Debian buster build is what I mentioned in the issue?

Yes, but I thought there are several distributions some of which the error occurs and others where it does not.

What you also did not say that it sometimes succeeds and sometimes fails.

The reply basically says the same what we already said. Maybe you sometimes pushed wrong commits which caused this problem. The email notification sometimes said you pushed 0 commits, maybe something went wrong?

@dominicjaeger
Copy link
Contributor Author

jenkins build libelektra please

1 similar comment
@dominicjaeger
Copy link
Contributor Author

jenkins build libelektra please

@dominicjaeger
Copy link
Contributor Author

Of libgit2.

It is in the the output of the dpkg command that I posted there?

Yes, but I thought there are several distributions some of which the error occurs and others where it does not.

There was a stretch artifcat before I think?

What you also did not say that it sometimes succeeds and sometimes fails.

IIRC I got to the stage the undefined references in the debian buster artifact was the only problem maybe once or twice.

The email notification sometimes said you pushed 0 commits, maybe something went wrong?

Maybe this is from amending to the last commit + force push?

@dominicjaeger
Copy link
Contributor Author

jenkins build libelektra please

@dominicjaeger
Copy link
Contributor Author

Superseded by #3105.

Improve 3-way merge automation moved this from In progress to Done Oct 21, 2019
@dominicjaeger dominicjaeger deleted the merge branch October 27, 2019 09:21
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
Development

Successfully merging this pull request may close these issues.

None yet

3 participants