Fixes #4633.

Although not used by default, the Reificator class at
https://github.com/jruby/jruby/blob/9.1.10.0/core/src/main/java/org/jruby/RubyClass.java#L1583
checks which version of callMethod it should use, and needs all three
versions: no argument, one argument, and more than one argument.

One of these overrides did not exist, and thus would result on a

```
java.lang.NoSuchMethodError: rubyobj.BrokenReify.callMethod(
Ljava/lang/String;Lorg/jruby/runtime/builtin/IRubyObject;)
Lorg/jruby/runtime/builtin/IRubyObject;
```

when running an example with -Xreify.classes=true

```ruby
class BrokenReify
  def initialize_copy(other)
    super
  end
end

puts BrokenReify.new.clone
```

With this fix, the above example starts instead giving a
java.lang.StackOverflowError due to another bug/bad interaction of the
reify.classes option, which I'll report separately.

Issue #4444

… FormatBag with BigInteger

Fix missing callMethod override

[WIP] Implement RubyDateParser in Java

Also allow access control to trigger quiet bail out of jar index.

Fix wrong FFI types on Windows x64

[jrubyc] respect Ruby visibility rules (do not generate private methods)

…eLexer.java

…rsion

in Java vs Ruby.  Improvement (BEFORE):

```text
system ~/work/jruby master 1423% jruby ../snippets/date_bench.rb 
Warming up --------------------------------------
           _strptime    13.890k i/100ms
Calculating -------------------------------------
_strptime    168.652k (± 5.7%) i/s -      1.681M in  10.002137s
```

AFTER:
```text
jruby ../snippets/date_bench.rb 
Warming up --------------------------------------
           _strptime    24.108k i/100ms
Calculating -------------------------------------
_strptime    291.972k (± 6.1%) i/s -      2.917M in  10.031620s
```

…ge and it

also gave a modest improvement (BEFORE):

```text
jruby ../snippets/date_bench.rb 
Warming up --------------------------------------
           _strptime    24.108k i/100ms
Calculating -------------------------------------
_strptime    291.972k (± 6.1%) i/s -      2.917M in  10.031620s
```

AFTER

```text
jruby ../snippets/date_bench.rb 
Warming up --------------------------------------
           _strptime    26.267k i/100ms
Calculating -------------------------------------
           _strptime    319.069k (± 5.8%) i/s -      3.178M in   9.998810s
           
```

…e year.

This improves my bench a bit more (BEFORE):

```text
jruby ../snippets/date_bench.rb 
Warming up --------------------------------------
           _strptime    26.267k i/100ms
Calculating -------------------------------------
           _strptime    319.069k (± 5.8%) i/s -      3.178M in   9.998810s
```

AFTER:

```text
jruby  ../snippets/date_bench.rb 
Warming up --------------------------------------
           _strptime    29.320k i/100ms
Calculating -------------------------------------
           _strptime    353.076k (± 2.9%) i/s -      3.548M in  10.057641s
```

I also examined MRIs implementation and they basically do all dyncalls for all
the math which this rewrite in Java does not do.  I think this is ok but it
underlies how much more overhead may be lurking here (since we are doing direct
Java math).

I did not change seconds and smaller units of time since they may exceed long
boundary.  In that case a followup optimization would be to check digits of text
and flip to BigInteger only in those cases otherwise assume long or even int
when it fits.

…ment_how_to_generate_java

Fix StrptimeLexer.flex to add a comment about how to generate the lexer

Handle nil buffer in Zlib::GzipReader#readpartial

upper bound check should not halt the loop since we're going pos--
(and there's a chance were able to find a match at a lower position)

… MRI)

86c3bee Print the output of --verbose/-V and --marker/-m on STDERR
b4236a6 Show more details when unexpected output is produced on STDOUT with mspec -j
b4eefac Remove extra spacing
a44aa8d Handle a closed STDIN in sub-processes
7293fa1 Let the sub-processes die silently if the main process died
91ea445 Inspect extra output sent by a worker
72426f3 Rename the variable ok to success
4bf8137 Use Process.wait2 to avoid relying on global variables
84a8121 Close the child IO when it terminates
d900a49 Highlight the important messages
38c58f3 Grep for a commit starting with the last merge message
f653f9b Simplify, avoid extra exceptions and add test for concurrent mspec mkdir_p
46ae38c fix race condition
9a80e01 Open files in binary modes for copying in MSpec's #cp
165d46d Use RUBY_ENGINE directly and remove RUBY_NAME
9a3fc4b Remove the --name option
2f7b200 Make all MSpec guards private
8a9182c Make all MSpec helpers private

git-subtree-dir: spec/mspec
git-subtree-split: 86c3bee671c1668fe5ac67902cbce450e5f590af

874f42f Set the external encoding of the GzipReader
c2285d8 Improve Time#{gmtime,utc} spec
20ce06e Improve Time#gmtime spec regarding frozen Time instances
ed3075b Correct usage of ruby_bug in Enumerable#uniq
5cecc96 Add version guard for SecureRandom.random_number with a Range
62cbf6e Fix spec: Time#localtime only raises if it needs to change the time zone on a frozen Time
e4a7f8a Put in quarantine specs which are waiting for a resolution in MRI
45de261 Zlib::GzipReader#pos underflows after calling #ungetbyte or #ungetc at start of file
b2d03bf BigDecimal with Float::INFINITY is buggy on MRI before 2.4
d109058 Add spec for nil buffer passed to Zlib::GzipReader#readpartial
090c87b Add failing specs showing that frozen Time objects can be mutated
18eece4 [spec] for SecureRandom.random_number (esp. for range argument)
f1a4344 Add spec for Zlib::GzipReader#gets("")
02eeed6 Add specs for Zlib::GzipReader#ungetbyte and #ungetc
00ad47e Fix Float::INFINITY BigDecimal comparison
e797cc6 Enumerable#uniq has been fixed in 2.5
582b9f4 Add an example Kernel#clone with `freeze` option
52b8515 Add MatchData#values_at specs for the new named captures
44a0731 Write MatchData#named_captures specs
370d2ad Fix ruby_bug use
12ca459 Write Enumerator::Lazy#uniq specs
2a2343c Follow ruby-lang `Bug #13669` in Enumerable#uniq
39d313d Write Enumerable#uniq specs (#447)
d313f22 Merge pull request #446 from kachick/enum-sum
e8a37ce Suppress rubocop
7f2961d Write Enumerable#sum specs
ec1c161 Add RuboCop on CI to be notified of new failures
1f29d20 Disable Lint/AssignmentInCondition as it warns for assignment in while
5101462 Use while + assignment instead of loop + break in DummyFTP
d334823 Disable a few cops which do not make sense for ruby/spec
2a0bdb2 Allow EmptyExpression in language specs
2d9590a Fix one useless empty expression
fe66187 Disable a few cops which do not make sense for ruby/spec
bc928b4 Fix alignment issues discovered by Lint/BlockAlignment
3f45fea Add `.rubocop_todo.yml`
47358a5 Add `.rubocop.yml`
358ca15 Close the server in Socket#connect_nonblock spec
b82d719 Add spec that a failing -S script sets the exit status correctly
c730f07 Use just 'r' for the return value variable
8654896 Close the write side to ensure the read side of the pipe see the data
cd1b911 Remove unused arguments to DummyFTP server
b2a6032 Remove SocketSpecs.port, and let the servers find an available port
736411f Let TCPServer find an available port
74cf1d7 Let UDPSocket find an available port
7d8eeb3 Cleanup and reset state in NetHTTPSpecs.stop_server
edc6bb5 Make sure the server is started to give the port
a1c2d66 Run the Kernel#yield_self spec on TruffleRuby
3639bb0 Implement Kernel#yield_self
83485ce Use before/after in UNIXServer#accept spec
a90afa1 Make {UNIX,TCP}Server#accept specs faster
b344842 Detect whether readline is available, no matter if the feature is enabled or not
5fb6c56 Fix DRb.start_service to use any available port
fac8624 Remove bad spec which cannot behave nicely in a multi-process environment
7f264f8 Remove empty files in drb specs
0fe2eb4 Create the file in the File::TMPFILE spec in its own directory
666458c Solaris 10 x86 raises SEGV
f4f9b02 Clean up a bit the Process.setpriority specs
a7bbcc4 Enable Process.setpriority for a single process on more platforms.
f28c663 Run in separate process/pgrp [Bug#13609]
6325c59 rubyspec: Fix method redefinition warning
bdddf9c Explain why we do not capture stderr
9b6201b Run the C-API specs on Windows!
d6642a0 Relax rb_str_hash spec as conversion from st_index_t is very complex
e8a430a Negative timestamp seems disallowed on x64-mingw32
327370e Silence make warnings when ruby/spec is run under make
fdbc0d4 Use nmake to build C APi specs on mswin
97fe8c7 Use RUBY_ENGINE instead of RUBY_NAME
33529f2 Remove duplicated `KernelBlockGiven` module
71097bc Improve spec for String#split with capture and limit
d8dac70 Simplify condition for run_mkmf_in_process
419e82b add spec for String#split with capture and limit
a34d2a5 MRI needs $extout = nil to build C extensions normally
b2e6a3a Fix compile_extension() for MRI
ca32ae2 Replace variable by its only usage
23261e2 Remove implementation-specific headers
ce7fbc0 Simplify as we only need to copy a single library file
2a047c8 * remove trailing spaces.
84ea66e Use mkmf to build C API specs
c278889 Only compile once the module_under_autoload extension

git-subtree-dir: spec/ruby
git-subtree-split: 874f42fc50c378d12d2fb72a7045a446a2ecc354