fix: use --dry-run=server by default for Helm v4 to fix .Capabilities.APIVersions by yxxhero · Pull Request #928 · databus23/helm-diff

yxxhero · 2026-02-11T06:38:51Z

Summary

Fixes #894

For Helm v4, --dry-run=server should be used by default to get the correct .Capabilities.APIVersions.Has behavior. Without this, conditions like {{ if .Capabilities.APIVersions.Has "monitoring.coreos.com/v1/PrometheusRule" }} are incorrectly evaluated by helm-diff when using Helm v4.0.0.

Changes

Modified cmd/helm.go to use --dry-run=server by default for Helm v4 when the user hasn't explicitly set --dry-run=client
Avoided duplicate --dry-run arguments for Helm v4

Testing

All existing tests pass.

….APIVersions For Helm v4, --dry-run=server should be used by default to get the correct .Capabilities.APIVersions.Has behavior. This is only applied when the user hasn't explicitly set --dry-run=client. Fixes #894 Signed-off-by: yxxhero <aiopsclub@163.com>

Signed-off-by: yxxhero <aiopsclub@163.com>

Add unit tests for PR #928 to verify the dry-run flag behavior for Helm v4, which uses --dry-run=server by default to get correct .Capabilities.APIVersions. Signed-off-by: yxxhero <aiopsclub@163.com>

Copilot

Pull request overview

This pull request fixes issue #894 where .Capabilities.APIVersions.Has is incorrectly evaluated when using Helm v4. The fix changes the default dry-run mode from client to server for Helm v4 to ensure that API version capabilities are correctly populated during template rendering. The change is scoped to Helm v4 only, maintaining backward compatibility with Helm v3.

Changes:

Modified cmd/helm.go to use --dry-run=server by default for Helm v4 when cluster access is allowed and the user hasn't explicitly set --dry-run=client
Added comprehensive test coverage in cmd/helm_test.go to verify the new dry-run flag behavior across different Helm versions and modes
Updated import formatting in cmd/upgrade_test.go to follow codebase conventions

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 6 comments.

File	Description
cmd/helm.go	Implements the core fix by defaulting to `--dry-run=server` for Helm v4 to populate `.Capabilities.APIVersions` correctly, while avoiding duplicate dry-run flags
cmd/helm_test.go	Adds test helper function and comprehensive test cases covering Helm v3/v4 behavior with different dry-run modes
cmd/upgrade_test.go	Minor import formatting change to use multi-line import style consistent with other test files
.github/workflows/ci.yaml	Adds HELM_DEBUG environment variable to one integration test for debugging output

Copilot · 2026-02-14T03:10:45Z

cmd/helm_test.go

+		})
+	}
+}
+


The test suite is missing coverage for dryRunMode="true" and dryRunMode="false" cases. According to cmd/upgrade.go, these are valid values that users can pass (line 25: validDryRunValues includes "true" and "false"). The behavior is documented as: "true" should behave like "client" and "false" should behave like "none". Test cases should be added to verify that these values are handled correctly for both Helm v3 and Helm v4, especially since the current implementation has a bug where dryRunMode="true" results in no dry-run flags being added for Helm v4.

Suggested change

func TestGetTemplateDryRunFlagsBoolModes(t *testing.T) {

cases := []struct {

name string

config dryRunFlagsConfig

expected []string

}{

{

name: "Helm v3 dryRunMode=true behaves like client",

config: dryRunFlagsConfig{

isHelmV4: false,

supportsDryRunLookup: true,

clusterAccessAllowed: true,

disableValidation: false,

dryRunMode: "true",

},

expected: []string{"--validate", "--dry-run=client"},

},

{

name: "Helm v3 dryRunMode=false behaves like none",

config: dryRunFlagsConfig{

isHelmV4: false,

supportsDryRunLookup: true,

clusterAccessAllowed: true,

disableValidation: false,

dryRunMode: "false",

},

expected: []string{"--validate"},

},

{

name: "Helm v4 dryRunMode=true behaves like client",

config: dryRunFlagsConfig{

isHelmV4: true,

supportsDryRunLookup: true,

clusterAccessAllowed: true,

disableValidation: false,

dryRunMode: "true",

},

expected: []string{"--dry-run=client"},

},

{

name: "Helm v4 dryRunMode=false behaves like none",

config: dryRunFlagsConfig{

isHelmV4: true,

supportsDryRunLookup: true,

clusterAccessAllowed: true,

disableValidation: false,

dryRunMode: "false",

},

expected: []string{},

},

}

for _, tc := range cases {

t.Run(tc.name, func(t *testing.T) {

actual := getTemplateDryRunFlags(tc.config)

if !reflect.DeepEqual(actual, tc.expected) {

t.Errorf("Expected %v, got %v", tc.expected, actual)

}

})

}

}

Copilot · 2026-02-14T03:10:46Z

cmd/helm_test.go

+			if cfg.dryRunMode != "client" {
+				flags = append(flags, "--dry-run=server")
+			}
+		} else {
+			flags = append(flags, "--validate")
+		}
+	}
+
+	if cfg.supportsDryRunLookup {
+		if !(cfg.isHelmV4 && cfg.dryRunMode != "client") {
+			if cfg.dryRunMode == "server" {
+				flags = append(flags, "--dry-run=server")
+			} else {
+				flags = append(flags, "--dry-run=client")
+			}
+		}


The test helper function getTemplateDryRunFlags has the same bug as the actual implementation in cmd/helm.go. The conditions on lines 23 and 32 don't handle dryRunMode="true" correctly. This needs to be fixed along with the actual implementation. The conditions should check for both "client" and "true" when deciding whether to skip server mode for Helm v4.

Copilot · 2026-02-14T03:10:46Z

.github/workflows/ci.yaml

+        env:
+          HELM_DEBUG: "true"


The HELM_DEBUG environment variable is being set for this specific test. HELM_DEBUG only controls debug logging output and is not required for the test to pass. Unless there's a specific reason to debug this particular test (which should be documented), this appears to be a leftover from development/debugging and should be removed to keep the CI configuration clean and consistent.

cmd/helm.go

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated no new comments.

- Fix conditions to properly handle dryRunMode="true" (behaves like client) - Fix conditions to properly handle dryRunMode="false" (behaves like none) - Add test cases for boolean dry-run modes - Remove leftover HELM_DEBUG from CI workflow Signed-off-by: yxxhero <aiopsclub@163.com>

yxxhero added 2 commits February 11, 2026 14:37

add debug flag

8d83d67

Signed-off-by: yxxhero <aiopsclub@163.com>

yxxhero mentioned this pull request Feb 11, 2026

.Capabilities.APIVersions.Has incorrectly filled when using helm v4.0.0 #894

Closed

test: add tests for Helm v4 --dry-run=server default behavior

b97de2b

Add unit tests for PR #928 to verify the dry-run flag behavior for Helm v4, which uses --dry-run=server by default to get correct .Capabilities.APIVersions. Signed-off-by: yxxhero <aiopsclub@163.com>

yxxhero requested a review from Copilot February 14, 2026 03:05

Copilot started reviewing on behalf of yxxhero February 14, 2026 03:06 View session

Copilot AI reviewed Feb 14, 2026

View reviewed changes

yxxhero force-pushed the fix/issue-894-helm-v4-dry-run-server branch from 8c2bb0d to 1995d0f Compare February 14, 2026 03:39

yxxhero requested a review from Copilot February 14, 2026 03:54

Copilot started reviewing on behalf of yxxhero February 14, 2026 03:54 View session

Copilot AI reviewed Feb 14, 2026

View reviewed changes

yxxhero force-pushed the fix/issue-894-helm-v4-dry-run-server branch from 1995d0f to 5fe2d8d Compare February 14, 2026 04:24

yxxhero merged commit f60c1a7 into master Feb 14, 2026
24 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix: use --dry-run=server by default for Helm v4 to fix .Capabilities.APIVersions#928

fix: use --dry-run=server by default for Helm v4 to fix .Capabilities.APIVersions#928
yxxhero merged 4 commits intomasterfrom
fix/issue-894-helm-v4-dry-run-server

yxxhero commented Feb 11, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Feb 14, 2026

Uh oh!

Copilot AI Feb 14, 2026

Uh oh!

Copilot AI Feb 14, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

+func TestGetTemplateDryRunFlagsBoolModes(t *testing.T) {
+	cases := []struct {
+		name     string
+		config   dryRunFlagsConfig
+		expected []string
+	}{
+		{
+			name: "Helm v3 dryRunMode=true behaves like client",
+			config: dryRunFlagsConfig{
+				isHelmV4:             false,
+				supportsDryRunLookup: true,
+				clusterAccessAllowed: true,
+				disableValidation:    false,
+				dryRunMode:           "true",
+			},
+			expected: []string{"--validate", "--dry-run=client"},
+		},
+		{
+			name: "Helm v3 dryRunMode=false behaves like none",
+			config: dryRunFlagsConfig{
+				isHelmV4:             false,
+				supportsDryRunLookup: true,
+				clusterAccessAllowed: true,
+				disableValidation:    false,
+				dryRunMode:           "false",
+			},
+			expected: []string{"--validate"},
+		},
+		{
+			name: "Helm v4 dryRunMode=true behaves like client",
+			config: dryRunFlagsConfig{
+				isHelmV4:             true,
+				supportsDryRunLookup: true,
+				clusterAccessAllowed: true,
+				disableValidation:    false,
+				dryRunMode:           "true",
+			},
+			expected: []string{"--dry-run=client"},
+		},
+		{
+			name: "Helm v4 dryRunMode=false behaves like none",
+			config: dryRunFlagsConfig{
+				isHelmV4:             true,
+				supportsDryRunLookup: true,
+				clusterAccessAllowed: true,
+				disableValidation:    false,
+				dryRunMode:           "false",
+			},
+			expected: []string{},
+		},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			actual := getTemplateDryRunFlags(tc.config)
+			if !reflect.DeepEqual(actual, tc.expected) {
+				t.Errorf("Expected %v, got %v", tc.expected, actual)
+			}
+		})
+	}
+}

Conversation

yxxhero commented Feb 11, 2026

Summary

Changes

Testing

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Feb 14, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 14, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 14, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant