Range#overlap? assumes that there is no minimum value.
This assumption makes
+(...-Float::INFINITY).overlap?((...-Float::INFINITY))+ returns true
while +(...-Float::INFINITY)+ is empty.
Caused by 04a92a6764. This treats
unbounded ranges of arbitrary objects the same as how unbounded
string ranges are treated:
(..x) === y # (y <=> x) <= 0
(...x) === y # (y <=> x) < 0
(x..) === y # (x <=> y) <= 0
Fixes [Bug #19864]
On Range#bsearch for endless ranges, we try positions at `begin + 2**i` (i = 0, 1, 2, ...)
to find a point that satisfies a given condition.
Subsequently, we perform binary searching with the interval `[begin, begin + 2**n]`.
However, the interval `[begin + 2**(n-1), begin + 2**n]` is sufficient for binary search
because `begin + 2**(n-1)` does not satisfy the condition.
The same applies to beginless ranges.
Prior to this commit the `OPTIMIZED_CMP` macro relied on a method lookup
to determine whether `<=>` was overridden. The result of the lookup was
cached, but only for the duration of the specific method that
initialized the cmp_opt_data cache structure.
With this method lookup, `[x,y].max` is slower than doing `x > y ?
x : y` even though there's an optimized instruction for "new array max".
(John noticed somebody a proposed micro-optimization based on this fact
in https://github.com/mastodon/mastodon/pull/19903.)
```rb
a, b = 1, 2
Benchmark.ips do |bm|
bm.report('conditional') { a > b ? a : b }
bm.report('method') { [a, b].max }
bm.compare!
end
```
Before:
```
Comparison:
conditional: 22603733.2 i/s
method: 19820412.7 i/s - 1.14x (± 0.00) slower
```
This commit replaces the method lookup with a new CMP basic op, which
gives the examples above equivalent performance.
After:
```
Comparison:
method: 24022466.5 i/s
conditional: 23851094.2 i/s - same-ish: difference falls within
error
```
Relevant benchmarks show an improvement to Array#max and Array#min when
not using the optimized newarray_max instruction as well. They are
noticeably faster for small arrays with the relevant types, and the same
or maybe a touch faster on larger arrays.
```
$ make benchmark COMPARE_RUBY=<master@5958c305> ITEM=array_min
$ make benchmark COMPARE_RUBY=<master@5958c305> ITEM=array_max
```
The benchmarks added in this commit also look generally improved.
Co-authored-by: John Hawthorn <jhawthorn@github.com>
Beginless ranges previously raised TypeError for this case,
except for string ranges, which had unexpected behavior:
('a'..'z').include?('ww') # false
(..'z').include?('ww') # previously true, now TypeError
Use of include? with endless ranges could previously result
in an infinite loop.
This splits off a range_string_cover_internal function from
range_include_internal.
Fixes [Bug #18580]
Previously `(2..).cover?("2"..)` was false, but
`(..2).cover?(.."2")` was true. This changes it so both are false,
treating beginless ranges the same as endless ranges in regards to
type checks.
This also adds documentation to #cover? to describe behavior with
beginless and endless ranges, testing each documentation example,
which is how this bug was found.
Fixes [Bug #18155]
Previously, include? would return true for the end of the range,
when it should return false because the range is exclusive.
Research and Analysis by Victor Shepelev.
Fixes [Bug #18577]
I used this regex:
([A-Za-z]+)\.html#(?:class|module)-[A-Za-z]+-label-([A-Za-z0-9\-\+]+)
And performed a global find & replace for this:
rdoc-ref:$1@$2
In commit:7817a438eb1803e7b3358f43bd1f38479badfbdc, the implementation
of `Time#succ`, which had been deprecated for 10 years, was finally
removed.
During that time, there was an explicit `instance_of?` check in
source:range.c#L350 with a comment that the check should be removed
once `Time#succ` is removed.
Since `Time#succ` is now gone, this check should be removed.
Note: this should be coordinated with adding a version guard to the
corresponding check in ruby/spec as well.
NARUSE, Yui
Kouhei Yanagita
Victor Shepelev
Burdette Lamar
Tanaka Akira
Nobuyoshi Nakada
Jeremy Evans
Shouichi Kamiya
Marco Costa
Daniel Colson
S-H-GAMELINKS
Yusuke Endoh
Jean Boussier
Takashi Kokubun
Peter Zhu
Jörg W Mittag