This touches cases where FromInt() is used on numeric constants, or
values which are already int32s, or int variables which are defined
close by and can be changed to int32s with little impact.
Signed-off-by: Stephen Kitt <skitt@redhat.com>
Kubernetes-commit: 45836971f27ca70cd7742e8ee66e99e3c648cf9f
* Preserve code blocks in templates.Normalizer
Currently, if templates.LongDesc function is invoked multiple times
for a string including a code block, this code block is converted to
invalid code block.
The reason of this issue is that blackfriday package, which templates package uses, relies on
4 * indentation to detect code blocks. This PR always preserves
4 * indentation in code blocks to always detect it code blocks in also
post invocations.
* Add indentations to test for code blocks
* Use strings.repeat for repetition
Kubernetes-commit: 0bff705acd8982e34b937116eb2016c9d6e4c4a6
The call to `i18n.LoadTranslations` needs to occur on init so
that package-level variables that call `i18n.T()` to initialize
their values will be able to receive the translated string.
Added new integration tests to test help output translation.
Kubernetes-commit: 34d980e77090d065199e9e808f8aebf6ba2db72c
- test.args should be passed instead of the os.Args of the test framework
to prevent simple invocation of kubectl without args that could
manifest in false positive test runs
- plugin execution should have a different test path
- tests should invoke functioning kubectl commands instead of the mock
ones to ensure the correct subcommand is executed without a failure
Kubernetes-commit: 8b9cbe62025da49a31518870f2aea0ce9797d3ce
Plugin subcommand resolution is relatively less used than the
builtin subcommands. That's why, instead always initializing a
hash map on memory, it would be better to use a getter function only
serves as needed.
In addition to that this function will be exported that external
libraries can use it.
Kubernetes-commit: 4634073d0b979d2827106a0a1010456ac420baa8
* add no resources found message to rollout-status command
* return err if not nil before no resource message
Kubernetes-commit: a5575425b039bf7c15dfaa9a7acf257fdc4fde3f
testing.T.Cleanup ensures the environment is restored after a test and
any of its parallel sub-tests. It's possible that some of these can be
simplified further to T.Setenv(key, ""), but I did not investigate.
Kubernetes-commit: e8d3a4a105cc8ea6e795e24d5dbf0c61ed78cacd
T.Setenv ensures that the environment is returned to its prior state
when the test ends. It also panics when called from a parallel test to
prevent racy test interdependencies.
Kubernetes-commit: 2181eea48435310d1b6e366ea8db2968c4941b93
kubectl defaults to /bin/bash if the SHELL environment variable isn't
set. Updating the docs to mention that you can override the shell
location by setting the SHELL environment variable.
Kubernetes-commit: 0a508ad8067dcc8d48420daca536fca09df5846b
A few notes about the change:
1. I need to initialize the fileclient once, in an init function, so I
don't have access to `testing.T` yet.
2. I want to be able to configure the openapi files that I use
3. We already have a "cache" client that wraps another client, we don't
need to re-implement caching here, one can just do:
`cache.NewClient(openapitest.NewFileClient("some/path"))` to do a cached
client. Or initialize it in an init/global var.
Since there is still some value to use the embedded file, make an
alternative constructor while using fs.FS interface to be able to
manipulate both virtual and disk-based filesystems.
Kubernetes-commit: 29503fd8d45bc2c9438e92936bf4111162529b40
Users can pass resources into `kubectl events` command via `--for` flag,
if they have desire to only get events for the resource they specify.
However, current `kubectl events` does not support passing fully qualified
names(e.g. `replicasets.apps`, `cronjobs.v1.batch`, etc.). This PR adds support
for this.
Kubernetes-commit: 457bb58ebbf5a863ce30f51c53a84bc41c2380dc
To improve wall-clock speed, we run list operations in parallel. This
particularly helps when the round-trip time is high.
We issue requests as quickly as possible, kube-apiservers should all
have priority and fairness at this point and we don't want to
duplicate/fight that system.
Kubernetes-commit: 82eee59d0feb4b303e6ef78ebb7ec646a059f266
* Test for ApplySet with --dry-run=client|server
* Use the real format for ApplySet ID
* Incorporate feedback
* Adjustments from rebase
Kubernetes-commit: 6a31757f45693fec5ea4723bcb405ce4437e31ca
TommyStarK