Add a description about optional hash objects for GC.stat and
GC.latest_gc_info. [Bug #14408]
The patch is provided by sho-h (Sho Hashimoto).
Thank you so much.
The command prompt on Windows always uses Unicode to take input and print
output but most Reline implementation depends on Encoding.default_external.
This commit introduces an abstracted structure about the encoding of Reline.
https://github.com/ruby/irb/issues/55
If we had put multiple open braces on a line the with no closing brace
spaces_of_nest array keeps getting '0' added to it. This means that when
we pop off of this array we are saying that we should be in position zero
for the next line. This is an issue because we don't always want to be
in position 0 after a closing brace.
Example:
```
[[[
]
]
]
```
In the above example the 'spaces_of_nest' array looks like this after
the first line is entered: [0,0,0]. We really want to be indented 4
spaces for the 1st closing brace 2 for the 2nd and 0 for the 3rd. i.e.
we want it to be: [0,2,4].
We also saw this issue with a heredoc inside of an array.
```
[<<FOO]
hello
FOO
```
https://github.com/ruby/irb/commit/80c69c8272
This commit fixes the check_newline_depth_difference method to multiple
open braces on one line into account. Before this change we were
subtracting from the depth in check_newline_depth_difference on
every open brace. This is the right thing to do if the opening and
closing brace are on the same line. For example in a method definition we
have an opening and closing parentheses we want to add 1 to our depth,
and then remove it.
```
def foo()
end
```
However this isn't the correct behavior when the brace spans multiple
lines. If a brace spans multiple lines we don't want to subtract from
check_newline_depth_difference and we want to treat the braces the same
way as we do `end` and allow check_corresponding_token_depth to pop the
correct depth.
Example of bad behavior:
```
def foo()
[
]
puts 'bar'
end
```
Example of desired behavior:
```
def foo()
[
]
puts 'bar'
end
```
https://github.com/ruby/irb/commit/7dc8af01e0
`shell: bash` runs bash on msys which prefers git on msys too,
then checked out in CRLF mode. Cygwin sed doesn't consider the CR
a part of EOL code, though.
ar_talbe (Hash representation for <=8 size) can use transient heap
and the memory area can move. So we need to restore `pair' ptr after
`func` call (which can run any programs) because of moving.
ISO/IEC 9899:1999 section 6.5.7 states that "If the value of the right
operand is negative or is greater than or equal to the width of the
promoted left operand, the behavior is undefined". So we have to take
care of such situations.
This has not been a problem because contemporary C compilers are
extraordinary smart to compile the series of shifts into a single
ROTLQ/ROTRQ machine instruction. In contrast to what C says those
instructions have fully defined behaviour for all possible inputs.
Hence it has been quite difficult to observe the undefined-ness of such
situations. But undefined is undefined. We should not rely on such
target-specific assumptions.
We are fixing the situation by carefully avoiding shifts with out-of-
range values. At least GCC since 4.6.3 and Clang since 8.0 can issue
the exact same instructions like before the changeset.
Also in case of Intel processors, there supposedly be intrinsics named
_rotr/_rotl that do exactly what we need. They, in practice, are absent
on Clang before 9.x so we cannot blindly use. But we can at least save
MSVC.
See also:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=57157https://bugs.llvm.org/show_bug.cgi?id=17332
Replaces `#ifdef _MSC_VER` with more accurate version checks. Also,
`defined(_WIN64) && defined(__AVX2__)` is redundant because there is no
such tihng like a 32bit AVX2 machine.