[luci] LuCId daemon explicit garbage collection

[Bryan Mayland]

Oh man, I wasn't looking for a patch I just wanted to start a discussion 
about what was the right thing to do.

With this change, my instrumentation is showing that each child process 
inherits on average 300KB less (380KB vs 680KB) LUA state.  Overall 
response times seem unchanged overall.  Serving a call() page that just 
returns a short string still takes ~310ms +/-20ms from accept() to data 
coming out.

I grabbed a screen capture of a LUA state memory usage graph (shown in 
the bright green).  The oscillating pattern is before the change, going 
from 470KB to 890KB and the flat part a the end is after the change 
where it hovers around 380KB-390KB.
https://picasaweb.google.com/lh/photo/w3cJryIu6IwdcHzdGhWSsmagQEnRdXMcytaEIVOlUv8?feat=directlink

Is this being considered for inclusion in trunk?