This reverts commit b692e49b1e.
The json-patch bump in k/k was reverted, so the corresponding bump in
kustomize should be reverted too.
Signed-off-by: Stephen Kitt <skitt@redhat.com>
* Add labels, annotations, namespaces, and names tasks
* Remove redundant information from ref/labels ref/annotations
* Update labels and annotations example names for consistency
* Remove name, prefix and suffix api examples
* Add link to tasks in reference
* Add link to tasks section for ref/configMapGenerator and ref/secretGenerator
* Add Labels/Annotatations headers
* Add Labels
* Add Template Labels
* Cleanup Add Template Labels
* Consolidate commonLabels and labels.includeSelectors
* Improve commonAnnotations example
* Add labels and annotations ref links
* Add generated ConfigMap to namespace example
* Add name headers
* Change header weights so they appear in sidebar
* Add namespace/name links
* Add generated ConfigMap to namePrefix example
* Add name propagation example
* Add more description of name propagation
* template labels
* Address feedback for labels
* Address names feedback
* Update for consistency
* Address feedback
* Remove API
Rename files that deal with configmaps and secrets to include the name of the
package as a prefix, as those were not following the pattern from the remaining
files in the package before.
* Fix a mistake in the comparison between elements in the ConfigMap and Secret args
test that causes it to become flaky.
* Rename the package in configmapSecretFlagsAndArgs_test.go back to util since the
testpackage linter has been disabled.
* fix patch.Target is nil in writePatchTargets
* add test case
* lint
* lint err not check
* remove new lin in imports
* rollback changes to `cmd :=`
* remove extra lines
* feat: add new command 'edit set configmap'
* Add a new command 'edit set configmap' to allow editing the values of an
already-existing configmap in a kustomization file.
* Add tests to validate the new feature.
* fix: add tests, minor refactoring to use constants
* Include tests to validade the new function ValidateSet, included to do
necessary validations when running the 'kustomize edit set configmap' command.
* Minor refactorings to use the existing constants in the 'edit set configmap'
command.
* Add dashes before each item in the comment explaining how ExpandFileSource()
works so IDEs don't try to reformat the list and remove the indentation in it.
* Because this change mutates the list of literal sources, ensure that both add
and set save the resulting list in a predictable order to make it easier to
check when new items are added/removed and aid in testing.
* Since literal sources are the only bit that's important in this test, verify
that the literal sources in the actual result is equal to what we expected it
to be.
* fix: change format to print resource name
Use '%q' formatter instead of '%s' to print resource name
Co-authored-by: Varsha <varshaprasad96@gmail.com>
* fix: add changes from code review
* Unexport constant that is used only in the scope of a single function.
* Add extra validation to ensure format is correct with one single '=' per key-value
pair.
* Add extra set of tests to validate format.
* Update test case to match new printed format in the error message.
* fix: rollback sort for edit add/set configmap
* chore: rename test package and unexport functions
Rename the test package from set_test back to set and unexport functions that do
not need to be exported anymore for testing purposes.
* feat: handle empty and default namespace as equal
Handle the empty and the default namespaces as equal. Add tests to validate this
scenario.
---------
Co-authored-by: Varsha <varshaprasad96@gmail.com>
Update the namespace handling in the edit add configmap and secret commands to
handle the empty namespace and the default namespace in the same way. Before
this change, if a configmap/secret was created using kustomize edit add where
one command was issued with default as the namespace and the other without a
namespace specified it would create two separate configmap generators, and then
kustomize build would fail if merge was not the strategy for either.
