Syntax patterns: more more statistics.

[Ruslan Shevchenko]

>
> The "widening operator (syntax)" check is not really helpful for the
> widening
> conversion stuff. The "semantic" one seems to be correct. I suspect the
> "syntax"
> one is also picking up a corresponding "encoding" operation corresponding
> to the
> "decoding" operation which is normally being detected by the "semantics"
> check
> (which is why the syntax value is very close to double the number for the
> semantic check). To save processing time, please drop the "syntax" check.
>

Ok. thanks, will be dropped. (for now to disable possible comment out
lines 214-221 in checkers_coin.def)

Also I forgot to drop old incorrect elvis check. (lines 124-143)

> Oh,
>
> And thanks very much for all your effort on this.
>

 Thanks for good words.  Problems are quite funny to justify scheduled time.

// Interesting that near 50% of my a priory assumptions about
frequency of constructions in Java code was wrong ;)