Talk:How the Nasal GC works

Todo: discuss GC pressure traits

• number of total objects
• lifetime of objects (short lived vs. long lived)
• functions with long argument lists vs. no parameter list (arg) or global variables
• object complexity/nested objects (hashes, vectors)
• expensive operations (such as repeated string concatenation)
• funcs vs. hash.methods() (with shared variables via me)
• anonymous funcs vs. named functions
• memory pool freeing and reallocations
• memory fragmentation

Todo: Publish GC Stats in property tree

• total memory allocated
• size of each pool
• size of constants table
• total references (active objects)
• references per module/sub module
• pool requests per frame/second/minute
• pool requests per module
• new naAlloc() allocations per frame/second/minute
• new naAlloc() allocations per module
• free() calls per module
• pool size fluctuations per time unit (frame,second,minute)
• time spent in each gc function
• threading overhead (locking)?

Todo: Allow GC parameters to be tweaked via properties

• minimum/maximum pool size (free blocks)
• add property to disable collecting (just allocating)
• add property to trigger GC