Only user-supplied nils should survive coalescing. Chart-default nils
defaults, not just user overrides. This caused:
- %!s(<nil>) in templates using Bitnami common.secrets.key (#31919)
- pluck fallbacks returning nil instead of falling through to globals
(#31971)
Fixes#31919Fixes#31971
Signed-off-by: Johannes Lohmer <jojo.dev@lohmer.com>
This ensures that when `helm uninstall` is run with `--keep-history`
any release in a `deployed` state other than the last release (e.g.
due to a failed upgrade) is being marked as `superseded`.
As a by-effect, running `helm upgrade` on a release which has been
uninstalled after an upgrade failure now no longer works. But instead
fails with a `"<name>" has no deployed releases` error. Which is the
(likely) intended behavior, and prevents other side-effects like
rolling back to a release version which happened before the uninstall
if `--atomic` (or `--rollback-on-failure`) was provided.
Signed-off-by: Hidde Beydals <hidde@hhh.computer>
This adds a --no-headers flag to the 'helm repo list' command,
allowing users to suppress table headers in the output.
Useful for scripting and automation.
Signed-off-by: Paul Van Laer <paul.van.laer1@gmail.com>
The test scripts hardcoded #!/bin/bash while they are not really
requiring bash. Use the more portable #!/usr/bin/env sh instead, so that
they use the default shell.
Signed-off-by: Tom Wieczorek <twieczorek@mirantis.com>
* Runtime abstraction to encapsulate subprocess code and enable future runtimes
Also fix race condition in TestPrepareCommandExtraArgs by replacing the shared variable modification with a local copy
Co-authored-by: George Jenkins <gvjenkins@gmail.com>
Signed-off-by: Scott Rigby <scott@r6by.com>
* Remove commented out code
Co-authored-by: Joe Julian <me@joejulian.name>
Signed-off-by: Scott Rigby <scott@r6by.com>
* Check test failure string
Co-authored-by: Jesse Simpson <jesse.simpson36@gmail.com>
Signed-off-by: Scott Rigby <scott@r6by.com>
---------
Signed-off-by: Scott Rigby <scott@r6by.com>
Co-authored-by: George Jenkins <gvjenkins@gmail.com>
Co-authored-by: Joe Julian <me@joejulian.name>
Co-authored-by: Jesse Simpson <jesse.simpson36@gmail.com>
The `vaildateTemplatesDir` function would still return `nil` if the directory doesn't exist,
so the early return that was documented never occurs.
Signed-off-by: Zach Burgess <zachburg@google.com>
This reverts #13533
This change has caused issues with numerous charts around things
unrelated to toml. This is because of functions like typeIs/typeOf
being used and acted upon.
The change caused a significant regression.
Note: This kind of change can be put into v3 charts, that are in
active development, without causing a regression.
Closes#30880
Signed-off-by: Matt Farina <matt@mattfarina.com>
George Jenkins