fix healthcheck timeouts and ut8 coercion

this commit fixes two bugs and adds regression tests.

when getting healthcheck values from an image, if the image does not
have a timeout defined, this resulted in a 0 value for timeout.  The
default as described in the man pages is 30s.

when inspecting a container with a healthcheck command, a customer
observed that the &, <, and > characters were being converted into a
unicode escape value.  It turns out json marshalling will by default
coerce string values to ut8.

Fixes: bz2028408

Signed-off-by: Brent Baude <bbaude@redhat.com>

1 files changed, 13 insertions, 0 deletions

diff --git a/libpod/define/healthchecks.go b/libpod/define/healthchecks.go
index 4114262b6..bde449d30 100644
--- a/libpod/define/healthchecks.go
+++ b/libpod/define/healthchecks.go
@@ -34,3 +34,16 @@ const (
 	// HealthCheckDefined means the healthcheck was found on the container
 	HealthCheckDefined HealthCheckStatus = iota
 )
+
+// Healthcheck defaults.  These are used both in the cli as well in
+// libpod and were moved from cmd/podman/common
+const (
+	// DefaultHealthCheckInterval default value
+	DefaultHealthCheckInterval = "30s"
+	// DefaultHealthCheckRetries default value
+	DefaultHealthCheckRetries uint = 3
+	// DefaultHealthCheckStartPeriod default value
+	DefaultHealthCheckStartPeriod = "0s"
+	// DefaultHealthCheckTimeout default value
+	DefaultHealthCheckTimeout = "30s"
+)