fix: template helmDefaults.postRendererArgs with release data (#2583)

PR #1839 introduced template rendering for postRendererArgs, but PR #2510
reverted it while fixing a separate regression. This left helmDefaults-level
postRendererArgs containing template expressions (e.g. {{ .Release.Name }})
passed to helm as literal strings instead of being resolved per-release.

Add renderPostRendererArgs() that templates helmDefaults.postRendererArgs
at flag-generation time using the release's template data, reusing the
existing createReleaseTemplateData() helper. Release-level args are already
templated by ExecuteTemplateExpressions and CLI args are static, so only
the helmDefaults path needs rendering.

Fixes #2580

Signed-off-by: opencode <opencode@users.noreply.github.com>
Signed-off-by: yxxhero <aiopsclub@163.com>

test/integration/test-cases/postrender-defaults-args/input/echo-args.bash

@@ -0,0 +1,12 @@
+#!/usr/bin/env bash
+arg=$1
+cat
+echo "---"
+cat <<EOS
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: rendered-arg-${arg}
+data:
+  arg: ${arg}
+EOS