Message size increase should be less than negligible for newline (2 bytes) and whitespace characters. It may become considerable, in the range of Giga-sized messages but it should still be negligible in context, since we’re dealing with a Giga message anyway.
I say “should” because I admit I haven’t explicitly compared the sizes. Do check this out first if you haven’t done this already, @pmk. I’ll give this a check in the morning.
Having said that, I’ve used pretty-print feature for JSON on IS, but it’s unavailable for XML at present.
I don’t suppose there’s any Java method to remove pretty-print, but the question is - is the Java invocation overhead on every single message, worthwhile?
I think not - remember, it’s so easy to over-engineer performance aspects and lose the battle against return on investment, just like how you were opposed to using string manipulation (and rightly so).
Can we introduce a pretty-print flag and/or server property for XML as well, @John_Carter4?
KM