This allows Helm 2 plugins that used HELM_HOME as a scratchpad to continue to work the same in Helm 3.
Signed-off-by: Matthew Fisher <matt.fisher@microsoft.com>
The revision argument that was mandatory to `helm rollback` was being ignored.
The only way to roll back to an older revision was to run `helm rollback RELEASE <insert anything here> --version REVISION`.
This change respects that argument and removes the `--version` flag, which was redundant.
Signed-off-by: Matthew Fisher <matt.fisher@microsoft.com>
This changes most of the KubeClient interface to only ever build objects once and
then pass in everything as lists of resources. As a consequence, we needed to refactor
several of the actions. I took the opportunity to refactor out some duplicated
code while I was in the same area
Signed-off-by: Taylor Thomas <taylor.thomas@microsoft.com>
`Result` is a misnomer and is going to be repurposed in a future commit for a
common result type for the different kube `Interface` methods
Signed-off-by: Taylor Thomas <taylor.thomas@microsoft.com>
The stable repository provides a quick onboarding with a set of
community curated charts. Two problems with the community stable
repository has lead to its need to be removed.
1. The URL is hard coded to a Google Cloud bucket under Google's
control. This was setup when Helm was part of Kubernetes and
Kubernetes was a Google project. The bucket cannot be
transfered to another non-Google controlled project. And,
the bucket is not accessible in some parts of the world
(e.g., China).
2. The number of charts in the stable repository has grown
generally unmaintainable. The repository maintainers cannot
manage the number of PRs coming it cauing delays in response
or no response and PRs are automatically closed. This is
a poor experience.
The alternatice is the Helm Hub that provides a central point of
search for many Helm repositories. Different people and organizations
can maintain their own charts. A central server is not needed as
Helm is setup to be distributed.
Signed-off-by: Matt Farina <matt@mattfarina.com>
Prior to now, helm has been silently ignoring any errors while parsing
the root Persistent flags. This causes an issue when a global flag comes
later on the command line than a localized flag for a subcommand, in
which the value for that global flag is never set. The solution is to
simply tell the Persistent flagset to ignore any unknown flags, since a
later command will process it
Signed-off-by: Ian Howell <ian.howell0@gmail.com>
Taylor Thomas
Martin Hickey
Adam Reese
Matthew Fisher
Joe Lanford
Karuppiah Natarajan
Ian Howell
Matt Farina
Seb Ospina
hd-rk