Make trilead-api as optional as it contains some algo users may no want to use#199
Conversation
…nt to use Signed-off-by: Olivier Lamy <olamy@apache.org>
jglick
left a comment
There was a problem hiding this comment.
And conversely I see that mina-sshd-api-core lists ssh-credentials as an optional dep. (It uses @Extension(optional = true) which would better be switched to @OptionalExtension as here, to avoid warnings on startup in the unlikely case ssh-credentials is not in fact enabled.)
…enkinsci#3039)" jenkinsci/ssh-credentials-plugin#199 makes the trilead-api optional. Likely needs some tuning to allow plugin compatibility tests to pass. This reverts commit 3fb5748.
…3039)" (#3049) jenkinsci/ssh-credentials-plugin#199 makes the trilead-api optional. Likely needs some tuning to allow plugin compatibility tests to pass. This reverts commit 3fb5748.
jenkinsci/ssh-credentials-plugin#199 makes the trilead-api plugin an optional dependency of the ssh-credentials plugin. That exposed a latent issue in the subversion plugin that is resolved in jenkinsci/subversion-plugin#284 . jenkinsci#3050 describes the details of the test failures in the plugin bill of mateerials. The failing tests can be seen with the commands: PLUGINS=subversion TEST=CompareAgainstBaselineCallableTest bash local-test.sh PLUGINS=mina-sshd-api-core LINE=weekly bash local-test.sh No need to block later releases, since the pull request thet fixes the issue has been submitted and is ready for review.
|
Shouldn't the trilead authenticator implementation eventually be moved to trilead-api plugin ? Like it is the case for Jsch, see #152, and mina-sshd. |
|
I can take a look. But will need some coordination / release plan between those 2 plugins to avoid circular dependencies. |
sounds good. We can coordinate at the pub :) |
|
After this PR, in the context of https://github.com/jenkinsci/acceptance-test-harness returns the empty list instead of the two Trilead implementations, causing jenkinsci/acceptance-test-harness#1539 and jenkinsci/acceptance-test-harness#1540 (39 new ATH failures). |
Signed-off-by: Olivier Lamy olamy@apache.org
Testing done