Don't include number of CPUs in generated docs

Firstly, it causes the "check for uncommitted diffs" CI step to
mysteriously fail in GitHub due to mismatch in the generated docs.

Here's an example of the diff in docs/modules/ROOT/pages/ec_opa_test.adoc:
 --p, --parallel:: the number of tests that can run in parallel, defaulting to the number of CPUs (explicitly set with 0). Benchmarks are always run sequentially. (Default: 16)
 +-p, --parallel:: the number of tests that can run in parallel, defaulting to the number of CPUs (explicitly set with 0). Benchmarks are always run sequentially. (Default: 4)

Secondly, having the number of CPUs hard-coded in static docs does
not make much sense.

Notes:
- Originally I implemented this workaround in the template internal/documentation/asciidoc/cli/cli.tmp
  which worked fine, but made the template quite messy.
- See relevant opa code here:
  https://github.com/open-policy-agent/opa/blob/d0c0ae9730b1ecb06a29c341c707c265138f0494/cmd/test.go#L569C69-L569C76
- This is was tough one to debug!

Ref: https://issues.redhat.com/browse/EC-1130

Author: simonbaird

docs/modules/ROOT/pages/ec_opa_test.adoc

@@ -91,7 +91,7 @@ ec opa test <path> [path [...]] [flags]
 -h, --help:: help for test (Default: false)
 --ignore:: set file and directory names to ignore during loading (e.g., '.*' excludes hidden files) (Default: [])
 -m, --max-errors:: set the number of errors to allow before compilation fails early (Default: 10)
--p, --parallel:: the number of tests that can run in parallel, defaulting to the number of CPUs (explicitly set with 0). Benchmarks are always run sequentially. (Default: 16)
+-p, --parallel:: the number of tests that can run in parallel, defaulting to the number of CPUs (explicitly set with 0). Benchmarks are always run sequentially.
 -r, --run:: run only test cases matching the regular expression
 -s, --schema:: set schema file path or directory path
 -t, --target:: set the runtime to exercise (Default: rego)

internal/documentation/asciidoc/cli/cli.go

@@ -133,14 +133,14 @@ func options(flags *pflag.FlagSet) []option {
 			opt := option{
 				flag.Name,
 				flag.Shorthand,
-				flag.DefValue,
+				flagDefValueMaybe(flag),
 				flag.Usage,
 			}
 			result = append(result, opt)
 		} else {
 			opt := option{
 				Name:         flag.Name,
-				DefaultValue: flag.DefValue,
+				DefaultValue: flagDefValueMaybe(flag),
 				Usage:        flag.Usage,
 			}
 			result = append(result, opt)
@@ -149,3 +149,14 @@ func options(flags *pflag.FlagSet) []option {
 
 	return result
 }
+
+func flagDefValueMaybe(flag *pflag.Flag) string {
+	// For `ec opa test` the default value for the --parallel flag is `runtime.NumCPU()`.
+	// We don't want to show that in the docs since it causes problems in the CI and it
+	// also it makes no sense in static documentation.
+	if flag.Name == "parallel" && strings.Contains(flag.Usage, "defaulting to the number of CPUs") {
+		// Don't show the default value in the docs
+		return ""
+	}
+	return flag.DefValue
+}