* cache: do not allocate chan for nothing
The explicit purpose of this channel is to never be closed and nothing
to be sent down on it. Hence, there's no need to allocate a channel - a
nil channel has exactly the desired behaviour.
Additionally, this is more relevant now that testing/synctest gets
unhappy when goroutines are blocked on reading on channels which are
created outside of the synctest bubble. Since this is a package var, that's
hard to avoid when using this package. Synctest is fine with nil
channels though.
Reported-by: Jussi Maki <jussi@isovalent.com>
Signed-off-by: David Bimmler <david.bimmler@isovalent.com>
* handlers: do not allocate chan for nothing
Nil chan has the desired semantics already, and this breaks
testing/synctest because the channel is allocated outside of the bubble.
Signed-off-by: David Bimmler <david.bimmler@isovalent.com>
---------
Signed-off-by: David Bimmler <david.bimmler@isovalent.com>
Kubernetes-commit: 52b7d035f9655d0c6ebeaaafd60add99700bb468
This allows consumers of term to not pull in dependencies on
github.com/gorilla/websocket and github.com/moby/spdystream.
Kubernetes-commit: 640dabd58b04b72f646ed85947cb8b407b36dc08
This is what happens when writing back a OIDC refresh token:
- plugin/pkg/client/auth/oidc/oidc.go:282
Calls `Persist` to save the new refresh token.
- tools/clientcmd/config.go:372
Calls `ModifyConfig` to save the config.
- tools/clientcmd/config.go:167
Calls `configAccess.GetLoadingPrecedence()` to get the files
listed from the `KUBECONFIG` environment variable.
- tools/clientcmd/loader.go:334
If the `ConfigAccess` was a `ClientConfigLoadingRules`, it
directly returns the `Precedence` slice from its `rules`
field.
THE PROBLEM:
The slice can be modified by the caller, unintentionally
changing the value of the `ClientConfigLoadingRules`'
`Precedence` field.
- tools/clientcmd/config.go:170
Then proceeds to in-place sort the slice returned by the
`ConfigAccess`. This is the same slice (by identity)
as the `ClientConfigLoadingRules`' `Precedence` field,
destroying its intended order.
- tools/clientcmd/config.go:179
Calls `configAccess.GetStartingConfig` to read the original
config so it can be compared with the new config.
- tools/clientcmd/loader.go:339
Calls `NewNonInteractiveDeferredLoadingClientConfig` with
itself as a parameter.
CONSEQUENCE:
At this point, its the `Precedence` has been
unintentionally sorted. When it loads the
config again, it gives precedence to whichever
file comes first in ascending alphabetical order.
- tools/clientcmd/config.go:192
If the file returned by `GetStartingConfig` has a
different `current-context` than the new config,
it calls `writeCurrentContext` to update the first
kubeconfig file in the `KUBECONFIG` environment
variable.
- tools/clientcmd/config.go:403
Calls `configAccess.GetDefaultFilename` to find the
destination kubeconfig file.
- tools/clientcmd/loader.go:358
Iterates through the kubeconfig files returned by
`GetLoadingPreferences` to find the first file
that exists.
CONSEQUENCE:
With the slice being sorted earlier, the files
returned by this call of `GetLoadingPreferences`
will be sorted alphabetically, rather than by
their intended order.
Kubernetes-commit: ffa084f81129ea685b176a282921c4d54906c539
* client-go: Replace depracted ErrWaitTimeout with recommended method
* Fix UT and Integration tests
* IT test
Kubernetes-commit: ffe306d67958297202e9492ea644b42c0e7e694d
Calling os.Exit() skipped the
`defer os.RemoveAll(tmp)`
above and thus dangling files & dirs would appear in their $TMPDIR.
Since go1.15 we can safely remove os.Exit() from TestMain() function.
As this exact issue was common enough so that go1.15 changed[1] the API of
TestMain to no longer require os.Exit to be called.
Reproducer:
$ (cd staging/src/k8s.io/client-go; go test ./tools/clientcmd)
$ ls -d /tmp/testkubeconfig*
/tmp/testkubeconfig1015943687
[1]: https://go-review.googlesource.com/c/go/+/219639
Kubernetes-commit: 457df1cf9817e29381609fbf2d85854a6b6a52ac
This raises awareness so that developers hopefully get it right without having to
learn about it from compile errors. It also explains that creating a PR to
fix the naming is not desired.
Kubernetes-commit: f40892f4813f29fc305d59073d2464a1fd7d6fe5
Normal execution always started to trigger this after adding more logging in
the E2E framework's exec_util.go:
I0506 21:23:56.781188 6341 exec_util.go:201] unexpected error trying to use websockets for pod exec: <nil>
That the "should fall back" implementation gets called when there is no error
and thus no reason to even consider falling back is odd. Now the execute
implementation checks for nil first.
This is the same approach taken also in the portforward code:
b35c5c0a30/staging/src/k8s.io/client-go/tools/portforward/fallback_dialer.go (L52)
Kubernetes-commit: 35870c480d64a8ae35c0e5839cc40b0115b7328a
- Add a new VeryShortWatchError struct for error matching, returned
by `handleAnyWatch`, up through `Reflector.ListAndWatch`.
- Update test expectations to match exact errors.
Kubernetes-commit: 6eff9db0f10db72f2c64390e106a80621d136439
Change:
- refactor Reflector.ListAndWatch and Reflector.watch to always call
watcher.Stop.
- refactor Reflector.handleAnyWatch to always call watcher.Stop,
EXCEPT when exitOnWatchListBookmarkReceived &&
watchListBookmarkReceived.
- Update unit tests with these new expectations.
Effect:
- ensures watcher.Stop is always called at least once
- avoids deadlocks in watcher implementations when watcher.Watch is
called, but watcher.Stop is never called.
Note: It's impossible to guarantee that Stop will only be called once.
So watch.Interface implementations must tollerate Stop being called
multiple times.
Kubernetes-commit: 3e609ecf6e945bf4562bddfc563fde9a4c3d0d90
This fixes variable passed as error reason upon websocker message read failure.
Previously a wrong variable was passed resulting in returning failure with nil error reason.
Kubernetes-commit: e1d3aaf8612c490fae965272a304d74fccf7e294
The event Object is created from the referenced Object name, however,
there is no guarantee that the name from the referenced Object will be a
valid Event Name.
For those Objects with name not valid for an Event, generate a new valid
name using an UUID.
Kubernetes-commit: ee36b817df06f84ce1a48ef4d65ed559c3775404
The ability to automatically stop on context cancellation was new functionality
that adds complexity and wasn't really used in Kubernetes. If someone wants
this, they can add it outside of the function.
A *WithLogger variant avoids the complexity and is consistent with
NewStreamWatcherWithLogger over in apimachinery.
Kubernetes-commit: 1a8d8c9b4a33daf9330434e1ad544ef3571722a3
The Lister and Watcher interfaces only supported methods without context, but
were typically implemented with client-go API calls which need a context. New
interfaces get added using the same approach as in
https://github.com/kubernetes/kubernetes/pull/129109.
Kubernetes-commit: 6688adae142e37114d9dfa8d94cd1d8a91fbcc13
