Instead of `irep -> proc` chain, we use `irep -> irep` chain to
avoid GC bugs like #3804. We added `target_class` reference to
`mrb_irep` struct. That means one more word consumption per `irep`.

Trying to make the code more readable

Update print.c

Strings not containing a newline are not printed synchronously
ex. bin/mruby -e '["a", "b", "c", "\n", "d", "e", "f", "\n"].each {|e| print e ; usleep 200000}'

Trying to make the source code more understandable by removing mixing of ternary if statements with attributions.

Tested local jump errors caused by break in a proc

It may be better to raise exceptions, but CRuby doesn't.

Unless `MRB_NAN_BOXING` is specified.

Allowed to pass multiple exceptions to assert_raise

Refactored #assert_raise and #assert_nothing_raised

This reverts commit 15d48efa.

`__ENCODING__' returns the current encoding name (string), unlike
CRuby that returns the encoding object.

Array refactorings and speed improvements

Example benchmark:
$ time build/bench/bin/mruby -e "Array.new(2_000_000){ |i| i }.index{ |i| i == 1_999_999 }"

Before:
real    0m0.934s
user    0m0.922s
sys     0m0.003s

After:
real    0m0.590s
user    0m0.583s
sys     0m0.007s

Fix building with MRB_USE_FLOAT=1

For performance reason. Segmented list consumes slightly less memory
but takes a lot of time especially when there are many slots in the
segment lists (e.g. class constants).