Ensure coverage script uses TESTFLAGS for consistent test execution behavior with shuffle and other flags across all test modes. Fixes issue #30892. Signed-off-by: Hsiu-Chi Tsai <84045975+thc1006@users.noreply.github.com>

.golangci.yml

@@ -1,22 +1,8 @@
-formatters:
+run:
-  enable:
+  timeout: 10m
-    - gofmt
-    - goimports
-
-  exclusions:
-    generated: lax
-
-  settings:
-    gofmt:
-      simplify: true
-
-    goimports:
-      local-prefixes:
-        - helm.sh/helm/v4
 
 linters:
-  default: none
+  disable-all: true
-
   enable:
     - depguard
     - dupl
@@ -27,45 +13,28 @@ linters:
     - nakedret
     - revive
     - staticcheck
-    - thelper
     - unused
-    - usestdlibvars
+    - gofmt
-    - usetesting
+    - goimports
-
-  exclusions:
-    generated: lax
-
-    presets:
-      - comments
-      - common-false-positives
-      - legacy
-      - std-error-handling
-
-    rules: []
-
-    warn-unused: true
 
-  settings:
+linters-settings:
+  gofmt:
+    simplify: true
+  goimports:
+    local-prefixes: helm.sh/helm/v4
   depguard:
     rules:
-        Main:
+      main:
         deny:
           - pkg: github.com/hashicorp/go-multierror
             desc: "use errors instead"
           - pkg: github.com/pkg/errors  
             desc: "use errors instead"
-
   dupl:
     threshold: 400
-
   gomodguard:
     blocked:
       modules:
         - github.com/evanphx/json-patch:
             recommendations:
               - github.com/evanphx/json-patch/v5
-
-run:
-  timeout: 10m
-
-version: "2"

CONTRIBUTING.md

@@ -232,25 +232,69 @@ docs](https://helm.sh/docs/developers/).
 
 To run the test suite, use the `make test` command. This will run all the unit tests.
 
-By default, the tests are run in a randomized order and each test is run 3 times. This is to help
+#### Test Configuration Variables
-detect flaky tests. If you want to run the tests in a specific order, you can use the `-shuffle=off`
-flag. If you want to run the tests a different number of times, you can use the `-count=n` flag,
-where `n` is the number of times you want to run the tests.
 
-For example, to run the tests in order and only once, you can use the following command:
+Helm uses two main variables to configure test execution:
 
+- **TESTFLAGS**: Controls test execution flags (default: `-shuffle=on`)
+- **TESTCOUNT**: Controls how many times tests are run (default: 1, CI uses 2)
+
+#### Default Test Behavior
+
+By default, tests are run in a randomized order with shuffle enabled. This helps detect order-dependent and flaky tests that might pass in development but fail in different environments.
+
+#### Common Test Commands
+
+Run tests with default settings (shuffle enabled, run once):
 ```bash
-make test TESTFLAGS="-shuffle=off -count=1"
+make test
 ```
 
-If a test fails, the seed that was used to randomize the tests will be printed. You can use this
+Run tests without shuffle for faster debugging:
-seed to reproduce the failure. For example, if the seed was `12345`, you can use the following
+```bash
-command to reproduce the failure:
+make test TESTFLAGS="-shuffle=off"
+```
 
+Run tests multiple times to catch flaky behavior:
+```bash
+make test TESTCOUNT=3
+```
+
+Run tests with a specific seed to reproduce failures:
 ```bash
 make test TESTFLAGS="-shuffle=on -seed=12345"
 ```
 
+Run tests multiple times with a specific seed:
+```bash
+make test TESTFLAGS="-shuffle=on -seed=12345" TESTCOUNT=3
+```
+
+Run only coverage tests:
+```bash
+make test-coverage
+```
+
+#### CI Behavior
+
+The continuous integration system runs tests with enhanced settings to catch more issues:
+- Tests are run twice (`TESTCOUNT=2`) to detect flaky behavior
+- Shuffle is always enabled with random seeds
+- Both unit tests and coverage tests are executed
+
+#### Reproducing Test Failures
+
+If a test fails, the seed used for randomization will be printed in the output. Use this seed to reproduce the exact failure:
+
+```bash
+make test TESTFLAGS="-shuffle=on -seed=1642529943"
+```
+
+For persistent issues, run the test multiple times with the same seed:
+```bash
+make test TESTFLAGS="-shuffle=on -seed=1642529943" TESTCOUNT=5
+```
+
 ## Pull Requests
 
 Like any good open source project, we use Pull Requests (PRs) to track code changes.

scripts/coverage.sh

@@ -24,7 +24,7 @@ generate_cover_data() {
   for d in $(go list ./...) ; do
     (
       local output="${coverdir}/${d//\//-}.cover"
-      go test -coverprofile="${output}" -covermode="$covermode" "$d"
+      go test -coverprofile="${output}" -covermode="$covermode" ${TESTFLAGS:-} "$d"
     )
   done