[plt-scheme] implications of large string literals

Are there adverse performance implications of large strings or large 
string literals?  Say, (just guessing) with the GC or some I/O routines?

I have a syntax transformer that creates code containing large string 
literals.  Typically these literals will be a few hundred or thousand 
characters in size, but they could be a million characters or more.

Sometimes this will be used in very short-lived processes  (fraction of 
a second and then MzScheme exits), and sometimes in long-lived ones 
(months without exit).

If such large string literals hurt performance, I could have the 
transformer break them up into chunks and process each separately.  
However, if all string literal sizes are pretty much the same to PLT, 
I'd rather make one call with a million-character string than a thousand 
calls with thousand-character strings.

-- 
http://www.neilvandyke.org/