Consider the online state when attempting to reconnect
There's actually a comprehensive error condition handling in the method above which properly distinguishing between transient and persistent problems, but we just ignore that decision here and continuously reconnect.
Together with kimap!9 (merged) this fixes the infinite SSL error dialog loop when rejecting to ignore an SSL error to a large extend. You still get the dialog twice now, and then after a few minutes again as this is considered to be a transient error (e.g. caused by capture portals). This at least gives you the opportunity now to actually fix the configuration or remove the resource.
(Bug 423424 remains open for SMTP)