Merge pull request #4861 from shuujii/fix-behavior-of-Kernel-Integer-to-numbers-ending-with-_-and-spaces

Fix behavior of `Kernel#Integer` to numbers ending with `_` and spaces

#### Before this patch:

  ```ruby
  Integer("1_ ")  #=> 1
  ```

#### After this patch (same as Ruby):

  ```ruby
  Integer("1_ ")  #=> ArgumentError
  ```

Merge pull request #4860 from shuujii/fix-behavior-of-String-to_i-Kernel-Integer-to-numbers-starting-with-_

Fix behavior of `String#to_i`/`Kernel#Integer` to numbers starting with `_`

#### Before this patch:

  ```ruby
  Integer("_1")  #=> 1
  "_1".to_i      #=> 1
  ```

#### After this patch (same as Ruby):

  ```ruby
  Integer("_1")  #=> ArgumentError
  "_1".to_i      #=> 0
  ```

Merge pull request #4858 from shuujii/fix-that-String-to_f-accepts-consecutive-_-as-a-numeric-expression

Fix that `String#to_f` accepts consecutive `_` as a numeric expression

Consecutive `_` is not allowed as a numeric expression:

  1_2__3             #=> SyntaxError
  Float("1_2__3")    #=> ArgumentError
  Integer("1_2__3")  #=> ArgumentError
  "1_2__3".to_i      #=> 12

But `String#to_f` accept it, so I fixed the issue.

Before this patch:

  "1_2__3".to_f      #=> 123

After this patch:

  "1_2__3".to_f      #=> 12

Now identifiers like `_1abc` are allowed.

Fix keyword arguments not be obtained with `mrb_get_args()`; Fix #4754

Fix the error message of `Kernel#Float`

#### Before this patch:

  ```
  $ bin/mruby -e 'Float("1_a")'
  -e:1: invalid string for float(a) (ArgumentError)
  ```

#### After this patch:

  ```
  $ bin/mruby -e 'Float("1_a")'
  -e:1: invalid string for float("1_a") (ArgumentError)
  ```

Add tests to `Math`

If ":" is after "|" and there is no "?" or "*", the keyword argument
could not be obtained and it was not initialized with `undef`.

For example: "|oo:"

Auto detect `MRB_TIME_T_UINT`

Refine `mrb_alloca()`

* The allocated memory is guaranteed to be aligned for any data type (it was
  not guaranteed when string type is embed).
* Make allocation size exactly specified size (does not allocate space for a
  null byte).

Silence Clang warning with `MRB_INT64` and `MRB_32BIT` in `time.c`

Silence the following warnings:

  ```
  /mruby/mrbgems/mruby-time/src/time.c:871:15: warning: result of comparison of constant 9223372036854775807 with expression of type 'time_t' (aka 'long') is always false [-Wtautological-constant-out-of-range-compare]
    if (tm->sec > MRB_INT_MAX || tm->sec < MRB_INT_MIN) {
        ~~~~~~~ ^ ~~~~~~~~~~~
  /mruby/mrbgems/mruby-time/src/time.c:871:40: warning: result of comparison of constant -9223372036854775808 with expression of type 'time_t' (aka 'long') is always false [-Wtautological-constant-out-of-range-compare]
    if (tm->sec > MRB_INT_MAX || tm->sec < MRB_INT_MIN) {
                                 ~~~~~~~ ^ ~~~~~~~~~~~
  /mruby/mrbgems/mruby-time/src/time.c:887:16: warning: result of comparison of constant 9223372036854775807 with expression of type 'time_t' (aka 'long') is always false [-Wtautological-constant-out-of-range-compare]
    if (tm->usec > MRB_INT_MAX || tm->usec < MRB_INT_MIN) {
        ~~~~~~~~ ^ ~~~~~~~~~~~
  /mruby/mrbgems/mruby-time/src/time.c:887:42: warning: result of comparison of constant -9223372036854775808 with expression of type 'time_t' (aka 'long') is always false [-Wtautological-constant-out-of-range-compare]
    if (tm->usec > MRB_INT_MAX || tm->usec < MRB_INT_MIN) {
                                ~~~~~~~~ ^ ~~~~~~~~~~~
  ```

Silence Clang warning with `MRB_INT32` and `MRB_64BIT` in `time.c`

Silence the following warning:

  ```
  /mruby/mrbgems/mruby-time/src/time.c:258:60: warning: result of comparison of constant -9223372036854775808 with expression of type 'mrb_int' (aka 'int') is always false [-Wtautological-constant-out-of-range-compare]
        if ((mrb_time_int)i > MRB_TIME_MAX || MRB_TIME_MIN > i) {
                                              ~~~~~~~~~~~~ ^ ~
  ```

Merge pull request #4848 from shuujii/fix-MRB_FIXNUM_SHIFT-with-MRB_WORD_BOXING-MRB_INT32-and-MRB_64BIT

Fix `MRB_FIXNUM_SHIFT` with `MRB_WORD_BOXING`, `MRB_INT32` and `MRB_64BIT`

### Example

  ```ruby
  # example.rb
  max32 = 2**30 - 1 + 2**30
  min32 = -max32-1
  [max32, max32+1, min32, min32-1].each{|n| p [n, n.class]}
  ```

#### Before this patch:

  ```
  $ bin/mruby example.rb
  [2147483647, Float]
  [2147483648, Float]
  [-2147483648, Float]
  [-2147483649, Float]
  ```

#### After this patch:

  ```
  $ bin/mruby example.rb
  [2147483647, Fixnum]
  [2147483648, Float]
  [-2147483648, Fixnum]
  [-2147483649, Float]
  ```

Quit `mruby -v` immediately if no program is given for Ruby compatibility

Fix `mruby --verbose` (regression by #4827)

#### Before this patch:

  ```
  $ bin/mruby --verbose -e 'p 1'
  bin/mruby: Cannot open program file: --verbose
  ```

#### After this patch:

  ```
  $ bin/mruby --verbose -e 'p 1'
  00001 NODE_SCOPE:
  (snip)
  irep 0x7fe97041df30 nregs=4 nlocals=1 pools=0 syms=1 reps=0 iseq=11
  file: -e
      1 000 OP_LOADSELF R1
  (snip)

  1
  ```

Change the URL of `mruby.org` [ci skip]

- `http` -> `https`
- Remove `www.`

Support `--` (end of options) to `mruby` command

#### Before this patch:

  ```
  $ bin/mruby -e 'p ARGV' -- -x
  bin/mruby: invalid option -- (-h will show valid options)
  ```

#### After this patch:

  ```
  $ bin/mruby -e 'p ARGV' -- -x
  ["-x"]
  ```

Add assertion to `RVALUE` size

Rename `BITSIZE` to `BIT` and `BIT` to `BIT_POS` for consistency

Fix `ARGV` value in `mruby` command (regression by #4827)

#### Before this patch:

  ```
  $ bin/mruby -e 'p ARGV' a b
  ["bin/mruby", "-e", "p ARGV", "a", "b"]
  ```

#### After this patch:

  ```
  $ bin/mruby -e 'p ARGV' a b
  ["a", "b"]
  ```