In #4550, @shuuji proposed the name name `MRB_STR_NO_MULTI_BYTE` for
more precise description. Although I agree that the name name is
correct, but the flag means the string does not contain multi byte UTF-8
characters, i.e. all characters fit in the range of ASCII.

Fix heap buffer overflow; ref #4549

This patch is showed in #4549.

Refine document to mrb_get_args()` [ci skip]

Fix and refine error message in `mrb_obj_to_sym()`

Before this patch:

  $ bin/mruby -e '1.respond_to?(2)'  #=> nil is not a symbol

After this patch (same as Ruby):

  $ bin/mruby -e '1.respond_to?(2)'  #=> 2 is not a symbol nor a string

Remove unused C header file from `src/etc.c`

Fix `include`, `prepend` and `extend` to frozen object

Nested `assert` for mrbtest

Replace `String#[]=` method by C implements

It is integration with part of argument parsing used in `mrb_str_aset_m()`.

The purpose is to eliminate string objects that are temporarily created during processing.

Ref commit 35319bed

When nesting `assert` used in test, it is indented and displayed.
Assertion numbers are concatenated by `"-"` at this time.

The purpose is to match the apparent numbers when failing with `assert_mruby`
which is defined by `mrbgems/mruby-bin-mruby/bintest/mruby.rb` for example.

Child assertions "skip" and "info" are reported as parent assertions "info"
and `$ok_test += 1`. The child assertions "ko" and "crash" are reported as the
parent assertion "ko" and `$ko_test += 1`.

When child assertions are mixed, "ko" takes precedence.

Incompatibility:

- `$mrbtest_assert_idx` takes `nil` or an integer array object.
  So far it was `nil` or an integer.
- `$asserts` points to the top of the internal stack in the `assert`.
- `$mrbtest_assert` points to the top of the internal stack in `assert`.

Use `__ENCODING__` in tests

Remove unnecessary backticks in `src/range.c`; ref #2858

It cannot be used for `String#size` test if judging whether or not `MRB_UTF8_STRING` is defined by result of `String#size`.

Add modification tests for immediate value

Fix broken links for mruby.org

A part of a fix for issue mruby/mruby.github.io#50

Silence unused label warnings from gcc; ref #4524

  mruby/mruby/src/string.c:1722:4: warning: label 'bytes' defined but not used [-Wunused-label]
      bytes:
      ^~~~~

Change to UTF-8 string reversing with in place

Fix `String#byteslice` with `MRB_UTF8_STRING` and some edge cases

Fix rational new on 32 bits mode

Fix complex new on 32 bits mode

Example:

  $ bin/mruby -e '
    p "あa".byteslice(1)
    p "bar".byteslice(3)
    p "bar".byteslice(4..0)
  '

  Before this patch:

    "a"
    ""
    RangeError (4..0 out of range)

  After this patch (same as Ruby):

    "\x81"
    nil
    nil

Fix `struct RRange` overflow on 32-bit CPU with `MRB_NAN_BOXING`

Remove unneeded statement in `SET_OBJ_VALUE` with `boxing_word.h`

Fix argument specs to `Array`

Merge pull request #4530 from shuujii/compare-obj-pointer-directly-instead-of-using-mrb_obj_eq-in-mrb_gc_unregister

Compare obj pointer directly instead of using mrb_obj_eq in mrb_gc_unregister

The #4520 tried to address the issue, but it changes the type of
`mrb_check_frozen` argument; close #4520