As these loops will be executed only once for every #hash invokation,
it would make sense to defer the decision to unroll the loops to the
runtime.

In principle, this should allow the JIT compiler to remove all range
checks within the loop. I haven't had time to verify this though.

While one could wish that JIT compilation optimised the eight sequential
byte reads into a single long read, it in fact does not.

This implementation should fallback to the slow implementation in a
context where Unsafe fails to load, but I haven't figured out how to
test that properly.

The ScriptingContainer class contains two methods named `setProfile`,
which differ only by their single argument type and do very different
things from one another.  One of the two signatures accepts
an argument of type `ProfilingMode`, which, in other places in the
code, is exposed via more explicitly named accessor methods such as
`getProfilingMode` and `setProfilingMode`.  I'm wondering if maybe
the method in `ScriptingContainer` was simply accidentally mis-named.

This commit deprecates the `setProfile(ProfilingMode)` method, and
introduces a replacement method `setProfilingMode(ProfilingMode)`,
which is more consistent with the rest of the code.

'system' method of multi args was not able to support
Windows internal commands.
(ex) echo, cd, etc

updates variadic invoker to match MRI
variadic functions may now take enum arguments

also adds ffi spec for failure

* make sure changes in src/templates will be used with out mvn clean

…porting single Strings.

See ruby/ruby@9cdd297

Made Exception#set_backtrace compliant with Ruby documentation

…_multiplication

Fix for issue 1695 : BigDecimal and Rational multiplication

Implemented multi args 'system' method on Windows fix #1074

fix enums to return nil on unknown symbol

Fixes #2006 and replaces #2007.

tweak include order

Optimize SipHash using sun.misc.Unsafe

Thanks to @byteit101 for the patch!

…ing-mode

Deprecate overloaded `setProfile`, prefer `setProfilingMode`

86a75dc incorrectly updated the method names the profiler needs to
ignore, resulting in incorrect profiler output.

This resulted in graph output including JRuby::Profiler calls and
incorrectly reported a total time of "0.00" with "100%" time spent in
every method.

Revert to the correct names, fixing the output.

The markdown.rb from rdoc appeared to have some newlines in the
middle of a string and a comment. I removed those after copying
from the gem, but I'm not sure how they got there.

…ences to be regexp values

…racter_after_backslash

improve handling of non ascii character after backslash

Fix profile output

…e - should fix #2042

…n once per suffix.

This change reduces the number of repeated failed classpath lookups, each of which generates a ClassNotFoundException.

Limit searches for service libraries to once per baseName, rather than once per suffix.