Added some compiler statistics utilities. The idea is that all you do
is place a call to SAMPLE/1 or SAMPLE/2 at the appropriate place in
the code, and NOT rely on their arguments for a side effect.
Each statistic has a name (a symbol), a type (a symbol) and perhaps
some optional parameters. The calls to SAMPLE/n create the statistic
if it unknown, and then incorporate the data for generating a report.
(COMPILER:RESET-STATISTICS!)
(COMPILER:DISPLAY-STATISTICS)
(SAMPLE/1 '(name type [parameters]) datum)
(SAMPLE/2 '(name type [parameters]) datum1 datum2)
Current unary statistics are
'(name AVERAGE) <x>) - number, sum and average
'(name HISTOGRAM VECTOR) - histogram small non-negative integers
'(name HISTOGRAM HASH-TABLE) - histogram numbers, strings, symbols etc
'(name HISTOGRAM) - same as HASH-TABLE
Currently there are no binary statistics.